Aren
/
NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
NetBSD/sys/dev/audio.c

4486 lines
112 KiB
C
Raw Normal View History

add a helper function to get the hw device_t from an audio_softc
2014-11-18 04:50:12 +03:00
/* $NetBSD: audio.c,v 1.266 2014/11/18 01:50:12 jmcneill Exp $ */
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
/*-
* Copyright (c) 2008 The NetBSD Foundation, Inc.
* All rights reserved.
*
* This code is derived from software contributed to The NetBSD Foundation
* by Andrew Doran.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
* ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
* TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
* BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
New audio subsystem
1995-02-21 04:35:58 +03:00
/*
* Copyright (c) 1991-1993 Regents of the University of California.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the Computer Systems
* Engineering Group at Lawrence Berkeley Laboratory.
* 4. Neither the name of the University nor of the Laboratory may be used
* to endorse or promote products derived from this software without
* specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
/*
* This is a (partially) SunOS-compatible /dev/audio driver for NetBSD.
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
*
New audio subsystem
1995-02-21 04:35:58 +03:00
* This code tries to do something half-way sensible with
* half-duplex hardware, such as with the SoundBlaster hardware. With
* half-duplex hardware allowing O_RDWR access doesn't really make
* sense. However, closing and opening the device to "turn around the
* line" is relatively expensive and costs a card reset (which can
* take some time, at least for the SoundBlaster hardware). Instead
* we allow O_RDWR access, and provide an ioctl to set the "mode",
* i.e. playing or recording.
*
* If you write to a half-duplex device in record mode, the data is
* tossed. If you read from the device in play mode, you get silence
* filled buffers at the rate at which samples are naturally
* generated.
*
* If you try to set both play and record mode on a half-duplex
* device, playing takes precedence.
*/
/*
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
* Locking: there are three locks.
*
* - sc_lock, provided by the underlying driver. This is an adaptive lock,
* returned in the second parameter to hw_if->get_locks(). It is known
* as the "thread lock".
*
* It serializes access to state in all places except the
* driver's interrupt service routine. This lock is taken from process
* context (example: access to /dev/audio). It is also taken from soft
* interrupt handlers in this module, primarily to serialize delivery of
* wakeups. This lock may be used/provided by modules external to the
* audio subsystem, so take care not to introduce a lock order problem.
* LONG TERM SLEEPS MUST NOT OCCUR WITH THIS LOCK HELD.
*
* - sc_intr_lock, provided by the underlying driver. This may be either a
* spinlock (at IPL_SCHED or IPL_VM) or an adaptive lock (IPL_NONE),
* returned in the first parameter to hw_if->get_locks(). It is known as
* the "interrupt lock".
*
* It provides atomic access to the device's hardware state, and to audio
* channel data that may be accessed by the hardware driver's ISR.
* In all places outside the ISR, sc_lock must be held before taking
* sc_intr_lock. This is to ensure that groups of hardware operations are
* made atomically. SLEEPS CANNOT OCCUR WITH THIS LOCK HELD.
*
* - sc_dvlock, private to this module. This is a custom reader/writer lock
* built on sc_lock and a condition variable. Some operations release
* sc_lock in order to allocate memory, to wait for in-flight I/O to
* complete, to copy to/from user context, etc. sc_dvlock serializes
* changes to filters and audio device settings while a read/write to the
* hardware is in progress. A write lock is taken only under exceptional
* circumstances, for example when opening /dev/audio or changing audio
* parameters. Long term sleeps and copy to/from user space may be done
* with this lock held.
*
* List of hardware interface methods, and which locks are held when each
* is called by this module:
*
* METHOD INTR THREAD NOTES
* ----------------------- ------- ------- -------------------------
* open x x
* close x x
* drain x x
* query_encoding - x
* set_params - x
* round_blocksize - x
* commit_settings - x
* init_output x x
* init_input x x
* start_output x x
* start_input x x
* halt_output x x
* halt_input x x
* speaker_ctl x x
* getdev - x
* setfd - x
* set_port - x
* get_port - x
* query_devinfo - x
* allocm - - Called at attach time
* freem - - Called at attach time
* round_buffersize - x
* mappage - - Mem. unchanged after attach
* get_props - x
* trigger_output x x
* trigger_input x x
* dev_ioctl - x
* get_locks - - Called at attach time
New audio subsystem
1995-02-21 04:35:58 +03:00
*/
add RCSIDs
2001-11-13 08:32:49 +03:00
#include <sys/cdefs.h>
add a helper function to get the hw device_t from an audio_softc
2014-11-18 04:50:12 +03:00
__KERNEL_RCSID(0, "$NetBSD: audio.c,v 1.266 2014/11/18 01:50:12 jmcneill Exp $");
add RCSIDs
2001-11-13 08:32:49 +03:00
New audio subsystem
1995-02-21 04:35:58 +03:00
#include "audio.h"
#if NAUDIO > 0
#include <sys/param.h>
#include <sys/ioctl.h>
#include <sys/fcntl.h>
#include <sys/vnode.h>
#include <sys/select.h>
Implement poll(2).
1996-09-07 16:40:22 +04:00
#include <sys/poll.h>
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
#include <sys/kmem.h>
New audio subsystem
1995-02-21 04:35:58 +03:00
#include <sys/malloc.h>
#include <sys/proc.h>
#include <sys/systm.h>
#include <sys/syslog.h>
Make mode a flag set. Full-dpulex fixes from John Kohl.
1995-07-07 05:52:30 +04:00
#include <sys/kernel.h>
Bring prototypes into scope.
1996-03-14 22:05:07 +03:00
#include <sys/signalvar.h>
- Don't use dev_conf.h - Remove unnecessary prototypes - Fix ioctl arguments - Don't use kprintf(...TO_LOG...); directly. This does not set the log priority, and uses priority of the previous message. Use log() with %:
1996-03-31 01:51:23 +03:00
#include <sys/conf.h>
New audio subsystem
1995-02-21 04:35:58 +03:00
#include <sys/audioio.h>
Audio changes: - Change the way attach and open works to allow multiple audio devices. - Split the mulaw.c file into two to avoid dragging in mulaw convertsion when they are not needed. Add 16 bit alaw/mulaw tables. - Change the way audio properties are gotten. - Recognize more versions os SoundBlaster.
1997-08-01 02:33:08 +04:00
#include <sys/device.h>
Use the softint API.
2007-10-08 20:18:02 +04:00
#include <sys/intr.h>
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
#include <sys/cpu.h>
Use intr.h on alpha and i386.
1996-05-13 05:01:55 +04:00
New audio subsystem
1995-02-21 04:35:58 +03:00
#include <dev/audio_if.h>
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
#include <dev/audiovar.h>
New audio subsystem
1995-02-21 04:35:58 +03:00
Convert to new orthogonal audio encoding scheme and implement some of the new encodings. The change to ioctl AUDIO_GETENC is NOT backwards compatible.
1997-05-07 22:51:31 +04:00
#include <machine/endian.h>
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
/* #define AUDIO_DEBUG 1 */
Make mode a flag set. Full-dpulex fixes from John Kohl.
1995-07-07 05:52:30 +04:00
#ifdef AUDIO_DEBUG
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
#define DPRINTF(x) if (audiodebug) printf x
Add a few more debug messages and clean up debug printing a little.
1998-04-28 13:07:12 +04:00
#define DPRINTFN(n,x) if (audiodebug>(n)) printf x
Sampling rate and mono-stereo conversion described in http://mail-index.netbsd.org/tech-kern/2002/03/04/0005.html auconv.c: Add conversion functions audio.c: Sample alignment, calling conversion functions, etc. audio_if.h: Add four hw_* members to "struct audio_params" audiovar.h: Add conversion buffers, etc. auich and uaudio: Add conversion request code to *_set_params().
2002-03-07 17:37:02 +03:00
int audiodebug = AUDIO_DEBUG;
New audio subsystem
1995-02-21 04:35:58 +03:00
#else
#define DPRINTF(x)
Add a few more debug messages and clean up debug printing a little.
1998-04-28 13:07:12 +04:00
#define DPRINTFN(n,x)
New audio subsystem
1995-02-21 04:35:58 +03:00
#endif
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
#define ROUNDSIZE(x) x &= -16 /* round to nice boundary */
#define SPECIFIED(x) (x != ~0)
#define SPECIFIED_CH(x) (x != (u_char)~0)
New audio subsystem
1995-02-21 04:35:58 +03:00
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
/* #define AUDIO_PM_IDLE */
#ifdef AUDIO_PM_IDLE
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
int audio_idle_timeout = 30;
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#endif
Allow more than one hardware driver to attach to the audio upper layer.
1997-05-28 03:24:56 +04:00
int audio_blk_ms = AUDIO_BLK_MS;
New audio subsystem
1995-02-21 04:35:58 +03:00
ANSIfy.
2001-10-03 03:31:54 +04:00
int audiosetinfo(struct audio_softc *, struct audio_info *);
Add a new ioctl AUDIO_GETBUFINFO. It works like AUDIO_GETINFO, but doesn't obtain the ports, gain and balance related parameters. Those generally require reading from the hardware and therefore are much more expensive to obtain. Modify OSS emulation to use the new ioctl where possible. This reduces CPU usage of mplayer during mp3 playback with my Thinkpad from 20% to < 1% and from 50% to 20% during Xvid playback. Review and comments from jmcneill@
2007-06-11 17:05:46 +04:00
int audiogetinfo(struct audio_softc *, struct audio_info *, int);
ANSIfy.
2001-10-03 03:31:54 +04:00
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
int audio_open(dev_t, struct audio_softc *, int, int, struct lwp *);
int audio_close(struct audio_softc *, int, int, struct lwp *);
ANSIfy.
2001-10-03 03:31:54 +04:00
int audio_read(struct audio_softc *, struct uio *, int);
int audio_write(struct audio_softc *, struct uio *, int);
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.
2007-03-04 08:59:00 +03:00
int audio_ioctl(struct audio_softc *, u_long, void *, int, struct lwp *);
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
int audio_poll(struct audio_softc *, int, struct lwp *);
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
int audio_kqfilter(struct audio_softc *, struct knote *);
ANSIfy.
2001-10-03 03:31:54 +04:00
paddr_t audio_mmap(struct audio_softc *, off_t, int);
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
int mixer_open(dev_t, struct audio_softc *, int, int, struct lwp *);
int mixer_close(struct audio_softc *, int, int, struct lwp *);
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.
2007-03-04 08:59:00 +03:00
int mixer_ioctl(struct audio_softc *, u_long, void *, int, struct lwp *);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
static void mixer_remove(struct audio_softc *);
ANSIfy.
2001-10-03 03:31:54 +04:00
static void mixer_signal(struct audio_softc *);
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
ANSIfy.
2001-10-03 03:31:54 +04:00
void audio_init_record(struct audio_softc *);
void audio_init_play(struct audio_softc *);
int audiostartr(struct audio_softc *);
int audiostartp(struct audio_softc *);
void audio_rint(void *);
void audio_pint(void *);
int audio_check_params(struct audio_params *);
void audio_calc_blksize(struct audio_softc *, int);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
void audio_fill_silence(struct audio_params *, uint8_t *, int);
ANSIfy.
2001-10-03 03:31:54 +04:00
int audio_silence_copyout(struct audio_softc *, int, struct uio *);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
void audio_init_ringbuffer(struct audio_softc *,
struct audio_ringbuffer *, int);
ANSIfy.
2001-10-03 03:31:54 +04:00
int audio_initbufs(struct audio_softc *);
void audio_calcwater(struct audio_softc *);
int audio_drain(struct audio_softc *);
void audio_clear(struct audio_softc *);
deal with PR#46232: ensure locking around audio_clear() is handled consistently and avoid locking against self.
2012-03-26 22:26:10 +04:00
void audio_clear_intr_unlocked(struct audio_softc *sc);
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void audio_pint_silence
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
(struct audio_softc *, struct audio_ringbuffer *, uint8_t *, int);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
int audio_alloc_ring
ANSIfy.
2001-10-03 03:31:54 +04:00
(struct audio_softc *, struct audio_ringbuffer *, int, size_t);
void audio_free_ring(struct audio_softc *, struct audio_ringbuffer *);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
static int audio_setup_pfilters(struct audio_softc *, const audio_params_t *,
stream_filter_list_t *);
static int audio_setup_rfilters(struct audio_softc *, const audio_params_t *,
stream_filter_list_t *);
static void audio_stream_dtor(audio_stream_t *);
static int audio_stream_ctor(audio_stream_t *, const audio_params_t *, int);
static void stream_filter_list_append
Protect audio_write's filter graph from being modified in the middle of a write. Fixes the 'audioctl of death' problem (PR#32563).
2006-04-18 23:15:27 +04:00
(stream_filter_list_t *, stream_filter_factory_t,
const audio_params_t *);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
static void stream_filter_list_prepend
Protect audio_write's filter graph from being modified in the middle of a write. Fixes the 'audioctl of death' problem (PR#32563).
2006-04-18 23:15:27 +04:00
(stream_filter_list_t *, stream_filter_factory_t,
const audio_params_t *);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
static void stream_filter_list_set
(stream_filter_list_t *, int, stream_filter_factory_t,
const audio_params_t *);
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
int audio_set_defaults(struct audio_softc *, u_int);
prototypes for ccd audio and ncr5380
1996-03-07 18:00:07 +03:00
Split device_t and softc for audio(4), and other related cosmetic changes.
2008-03-04 21:23:44 +03:00
int audioprobe(device_t, cfdata_t, void *);
void audioattach(device_t, device_t, void *);
int audiodetach(device_t, int);
int audioactivate(device_t, enum devact);
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#ifdef AUDIO_PM_IDLE
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
static void audio_idle(void *);
static void audio_activity(device_t, devactive_t);
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#endif
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
A pointer typedef entails trading too much flexibility to declare const and non-const types, and the kernel uses both const and non-const PMF qualifiers and device suspensors, so change the pmf_qual_t and device_suspensor_t typedefs from "pointers to const" to non-pointer, non-const types.
2010-02-25 01:37:54 +03:00
static bool audio_suspend(device_t dv, const pmf_qual_t *);
static bool audio_resume(device_t dv, const pmf_qual_t *);
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
static void audio_volume_down(device_t);
static void audio_volume_up(device_t);
static void audio_volume_toggle(device_t);
static void audio_mixer_capture(struct audio_softc *);
static void audio_mixer_restore(struct audio_softc *);
Halt input/output on soft suspend if currently running, and restore them on soft resume.
2006-06-19 14:19:08 +04:00
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
static int audio_get_props(struct audio_softc *);
static bool audio_can_playback(struct audio_softc *);
static bool audio_can_capture(struct audio_softc *);
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
static void audio_softintr_rd(void *);
static void audio_softintr_wr(void *);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
static int audio_enter(dev_t, krw_t, struct audio_softc **);
static void audio_exit(struct audio_softc *);
static int audio_waitio(struct audio_softc *, kcondvar_t *);
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
struct portname {
minor KNF
2005-01-15 19:23:03 +03:00
const char *name;
int mask;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
};
sprinkle const qualifiers to const stuff
2003-03-31 22:47:58 +04:00
static const struct portname itable[] = {
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
{ AudioNmicrophone, AUDIO_MICROPHONE },
{ AudioNline, AUDIO_LINE_IN },
{ AudioNcd, AUDIO_CD },
add missing initializers.
2006-08-28 03:54:50 +04:00
{ 0, 0 }
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
};
sprinkle const qualifiers to const stuff
2003-03-31 22:47:58 +04:00
static const struct portname otable[] = {
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
{ AudioNspeaker, AUDIO_SPEAKER },
{ AudioNheadphone, AUDIO_HEADPHONE },
{ AudioNline, AUDIO_LINE_OUT },
add missing initializers.
2006-08-28 03:54:50 +04:00
{ 0, 0 }
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
};
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
void au_setup_ports(struct audio_softc *, struct au_mixer_ports *,
mixer_devinfo_t *, const struct portname *);
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
int au_set_gain(struct audio_softc *, struct au_mixer_ports *,
minor KNF
2005-01-15 19:23:03 +03:00
int, int);
ANSIfy.
2001-10-03 03:31:54 +04:00
void au_get_gain(struct audio_softc *, struct au_mixer_ports *,
minor KNF
2005-01-15 19:23:03 +03:00
u_int *, u_char *);
ANSIfy.
2001-10-03 03:31:54 +04:00
int au_set_port(struct audio_softc *, struct au_mixer_ports *,
minor KNF
2005-01-15 19:23:03 +03:00
u_int);
ANSIfy.
2001-10-03 03:31:54 +04:00
int au_get_port(struct audio_softc *, struct au_mixer_ports *);
int au_get_lr_value(struct audio_softc *, mixer_ctrl_t *,
minor KNF
2005-01-15 19:23:03 +03:00
int *, int *);
ANSIfy.
2001-10-03 03:31:54 +04:00
int au_set_lr_value(struct audio_softc *, mixer_ctrl_t *,
minor KNF
2005-01-15 19:23:03 +03:00
int, int);
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
int au_portof(struct audio_softc *, char *, int);
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
typedef struct uio_fetcher {
stream_fetcher_t base;
struct uio *uio;
int usedhigh;
int last_used;
} uio_fetcher_t;
static void uio_fetcher_ctor(uio_fetcher_t *, struct uio *, int);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
static int uio_fetcher_fetch_to(struct audio_softc *, stream_fetcher_t *,
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
audio_stream_t *, int);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
static int null_fetcher_fetch_to(struct audio_softc *, stream_fetcher_t *,
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
audio_stream_t *, int);
Merge the gehenna-devsw branch into the trunk. This merge changes the device switch tables from static array to dynamically generated by config(8). - All device switches is defined as a constant structure in device drivers. - The new grammer ``device-major'' is introduced to ``files''. device-major <prefix> char <num> [block <num>] [<rules>] - All device major numbers must be listed up in port dependent majors.<arch> by using this grammer. - Added the new naming convention. The name of the device switch must be <prefix>_[bc]devsw for auto-generation of device switch tables. - The backward compatibility of loading block/character device switch by LKM framework is broken. This is necessary to convert from block/character device major to device name in runtime and vice versa. - The restriction to assign device major by LKM is completely removed. We don't need to reserve LKM entries for dynamic loading of device switch. - In compile time, device major numbers list is packed into the kernel and the LKM framework will refer it to assign device major number dynamically.
2002-09-06 17:18:43 +04:00
dev_type_open(audioopen);
dev_type_close(audioclose);
dev_type_read(audioread);
dev_type_write(audiowrite);
dev_type_ioctl(audioioctl);
dev_type_poll(audiopoll);
dev_type_mmap(audiommap);
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
dev_type_kqfilter(audiokqfilter);
Merge the gehenna-devsw branch into the trunk. This merge changes the device switch tables from static array to dynamically generated by config(8). - All device switches is defined as a constant structure in device drivers. - The new grammer ``device-major'' is introduced to ``files''. device-major <prefix> char <num> [block <num>] [<rules>] - All device major numbers must be listed up in port dependent majors.<arch> by using this grammer. - Added the new naming convention. The name of the device switch must be <prefix>_[bc]devsw for auto-generation of device switch tables. - The backward compatibility of loading block/character device switch by LKM framework is broken. This is necessary to convert from block/character device major to device name in runtime and vice versa. - The restriction to assign device major by LKM is completely removed. We don't need to reserve LKM entries for dynamic loading of device switch. - In compile time, device major numbers list is packed into the kernel and the LKM framework will refer it to assign device major number dynamically.
2002-09-06 17:18:43 +04:00
const struct cdevsw audio_cdevsw = {
Change (mostly mechanically) every cdevsw/bdevsw I can find to use designated initializers. I have not built every extant kernel so I have probably broken at least one build; however I've also found and fixed some wrong cdevsw/bdevsw entries so even if so I think we come out ahead.
2014-03-16 09:20:22 +04:00
.d_open = audioopen,
.d_close = audioclose,
.d_read = audioread,
.d_write = audiowrite,
.d_ioctl = audioioctl,
.d_stop = nostop,
.d_tty = notty,
.d_poll = audiopoll,
.d_mmap = audiommap,
.d_kqfilter = audiokqfilter,
Add d_discard to all struct cdevsw instances I could find. All have been set to "nodiscard"; some should get a real implementation.
2014-07-25 12:10:31 +04:00
.d_discard = nodiscard,
Change (mostly mechanically) every cdevsw/bdevsw I can find to use designated initializers. I have not built every extant kernel so I have probably broken at least one build; however I've also found and fixed some wrong cdevsw/bdevsw entries so even if so I think we come out ahead.
2014-03-16 09:20:22 +04:00
.d_flag = D_OTHER | D_MPSAFE
Merge the gehenna-devsw branch into the trunk. This merge changes the device switch tables from static array to dynamically generated by config(8). - All device switches is defined as a constant structure in device drivers. - The new grammer ``device-major'' is introduced to ``files''. device-major <prefix> char <num> [block <num>] [<rules>] - All device major numbers must be listed up in port dependent majors.<arch> by using this grammer. - Added the new naming convention. The name of the device switch must be <prefix>_[bc]devsw for auto-generation of device switch tables. - The backward compatibility of loading block/character device switch by LKM framework is broken. This is necessary to convert from block/character device major to device name in runtime and vice versa. - The restriction to assign device major by LKM is completely removed. We don't need to reserve LKM entries for dynamic loading of device switch. - In compile time, device major numbers list is packed into the kernel and the LKM framework will refer it to assign device major number dynamically.
2002-09-06 17:18:43 +04:00
};
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
Respect the law: Use A-law and mu-law as spellings as far as easily possible. Inspired by Igor Sobrado in PR 19680.
2003-04-06 22:20:07 +04:00
/* The default audio mode: 8 kHz mono mu-law */
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
const struct audio_params audio_default = {
.sample_rate = 8000,
.encoding = AUDIO_ENCODING_ULAW,
.precision = 8,
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
.validbits = 8,
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
.channels = 1,
};
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
Detach audio(4) during shutdown.
2009-04-18 00:04:35 +04:00
CFATTACH_DECL3_NEW(audio, sizeof(struct audio_softc),
audioprobe, audioattach, audiodetach, audioactivate, NULL, NULL,
DVF_DETACH_SHUTDOWN);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
Adjust for config changes.
1998-01-12 11:44:08 +03:00
extern struct cfdriver audio_cd;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
int
Split device_t and softc for audio(4), and other related cosmetic changes.
2008-03-04 21:23:44 +03:00
audioprobe(device_t parent, cfdata_t match, void *aux)
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
{
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_attach_args *sa;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
minor KNF
2005-01-15 19:23:03 +03:00
sa = aux;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
DPRINTF(("audioprobe: type=%d sa=%p hw=%p\n",
sa->type, sa, sa->hwif));
Do audio attachment the right way. From Chris Demetriou. (Maybe I'll get the hang of the configuration stuff one of these days. :-)
1997-10-29 05:00:20 +03:00
return (sa->type == AUDIODEV_TYPE_AUDIO) ? 1 : 0;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
void
Split device_t and softc for audio(4), and other related cosmetic changes.
2008-03-04 21:23:44 +03:00
audioattach(device_t parent, device_t self, void *aux)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_softc *sc;
struct audio_attach_args *sa;
const struct audio_hw_if *hwp;
void *hdlp;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
int error;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
mixer_devinfo_t mi;
For the "record.port", choose "record.source" over "inputs.source" if the former exists. Partial fix for PR kern/29178.
2005-02-14 02:53:20 +03:00
int iclass, mclass, oclass, rclass, props;
int record_master_found, record_source_found;
don't hold the thread lock while calling allocm() or freem(). fixes PR#46121
2012-04-06 10:15:13 +04:00
bool can_capture, can_playback;
New audio subsystem
1995-02-21 04:35:58 +03:00
Split device_t and softc for audio(4), and other related cosmetic changes.
2008-03-04 21:23:44 +03:00
sc = device_private(self);
sc->dev = self;
minor KNF
2005-01-15 19:23:03 +03:00
sa = aux;
hwp = sa->hwif;
hdlp = sa->hdl;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
cv_init(&sc->sc_rchan, "audiord");
cv_init(&sc->sc_wchan, "audiowr");
cv_init(&sc->sc_lchan, "audiolk");
if (hwp == 0 || hwp->get_locks == 0) {
printf(": missing method\n");
panic("audioattach");
}
hwp->get_locks(hdlp, &sc->sc_intr_lock, &sc->sc_lock);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
#ifdef DIAGNOSTIC
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if (hwp->query_encoding == 0 ||
Change the interface between high and lowlevel audio drivers again: Set the encoding parameters slightly differently. Remove the SW encoding/decodinf functions from this interface and move them to the audio_parameter struct; this is both more efficient and flexible.
1997-05-10 02:16:27 +04:00
hwp->set_params == 0 ||
Add trigger_output and trigger_input entry points, which replace init_{in,out}put and start_{in,out}put, and are only called once.
1998-08-10 00:28:07 +04:00
(hwp->start_output == 0 && hwp->trigger_output == 0) ||
(hwp->start_input == 0 && hwp->trigger_input == 0) ||
New audio subsystem
1995-02-21 04:35:58 +03:00
hwp->halt_output == 0 ||
hwp->halt_input == 0 ||
hwp->getdev == 0 ||
hwp->set_port == 0 ||
hwp->get_port == 0 ||
Audio changes: - Change the way attach and open works to allow multiple audio devices. - Split the mulaw.c file into two to avoid dragging in mulaw convertsion when they are not needed. Add 16 bit alaw/mulaw tables. - Change the way audio properties are gotten. - Recognize more versions os SoundBlaster.
1997-08-01 02:33:08 +04:00
hwp->query_devinfo == 0 ||
hwp->get_props == 0) {
Minor output formatting buglet.
1999-09-23 15:53:13 +04:00
printf(": missing method\n");
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
sc->hw_if = 0;
return;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
}
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
#endif
New audio subsystem
1995-02-21 04:35:58 +03:00
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
sc->hw_if = hwp;
sc->hw_hdl = hdlp;
sc->sc_dev = parent;
sc->sc_lastinfovalid = false;
Report all AUDIO_PROP_* properties when attaching an audio device.
1999-12-15 15:09:34 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
props = audio_get_props(sc);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
aprintify
2007-12-01 07:50:50 +03:00
Report all AUDIO_PROP_* properties when attaching an audio device.
1999-12-15 15:09:34 +03:00
if (props & AUDIO_PROP_FULLDUPLEX)
aprintify
2007-12-01 07:50:50 +03:00
aprint_normal(": full duplex");
Indicate whether the decive is full or half duplex at attach time.
1999-02-19 20:09:16 +03:00
else
aprintify
2007-12-01 07:50:50 +03:00
aprint_normal(": half duplex");
Report all AUDIO_PROP_* properties when attaching an audio device.
1999-12-15 15:09:34 +03:00
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
if (props & AUDIO_PROP_PLAYBACK)
aprint_normal(", playback");
if (props & AUDIO_PROP_CAPTURE)
aprint_normal(", capture");
Report all AUDIO_PROP_* properties when attaching an audio device.
1999-12-15 15:09:34 +03:00
if (props & AUDIO_PROP_MMAP)
aprintify
2007-12-01 07:50:50 +03:00
aprint_normal(", mmap");
Report all AUDIO_PROP_* properties when attaching an audio device.
1999-12-15 15:09:34 +03:00
if (props & AUDIO_PROP_INDEPENDENT)
aprintify
2007-12-01 07:50:50 +03:00
aprint_normal(", independent");
Report all AUDIO_PROP_* properties when attaching an audio device.
1999-12-15 15:09:34 +03:00
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
aprint_naive("\n");
aprintify
2007-12-01 07:50:50 +03:00
aprint_normal("\n");
Report all AUDIO_PROP_* properties when attaching an audio device.
1999-12-15 15:09:34 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
don't hold the thread lock while calling allocm() or freem(). fixes PR#46121
2012-04-06 10:15:13 +04:00
can_playback = audio_can_playback(sc);
can_capture = audio_can_capture(sc);
mutex_exit(sc->sc_lock);
if (can_playback) {
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
error = audio_alloc_ring(sc, &sc->sc_pr,
AUMODE_PLAY, AU_RING_SIZE);
if (error) {
sc->hw_if = NULL;
aprint_error("audio: could not allocate play buffer\n");
return;
}
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
}
don't hold the thread lock while calling allocm() or freem(). fixes PR#46121
2012-04-06 10:15:13 +04:00
if (can_capture) {
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
error = audio_alloc_ring(sc, &sc->sc_rr,
AUMODE_RECORD, AU_RING_SIZE);
if (error) {
if (sc->sc_pr.s.start != 0)
audio_free_ring(sc, &sc->sc_pr);
sc->hw_if = NULL;
aprint_error("audio: could not allocate record buffer\n");
return;
}
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
}
Sampling rate and mono-stereo conversion described in http://mail-index.netbsd.org/tech-kern/2002/03/04/0005.html auconv.c: Add conversion functions audio.c: Sample alignment, calling conversion functions, etc. audio_if.h: Add four hw_* members to "struct audio_params" audiovar.h: Add conversion buffers, etc. auich and uaudio: Add conversion request code to *_set_params().
2002-03-07 17:37:02 +03:00
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
sc->sc_lastgain = 128;
don't hold the thread lock while calling allocm() or freem(). fixes PR#46121
2012-04-06 10:15:13 +04:00
mutex_enter(sc->sc_lock);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
error = audio_set_defaults(sc, 0);
mutex_exit(sc->sc_lock);
if (error != 0) {
aprintify
2007-12-01 07:50:50 +03:00
aprint_error("audioattach: audio_set_defaults() failed\n");
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
sc->hw_if = NULL;
return;
}
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
- convert the audio softintr's to SOFTINT_SERIAL. right now they're on the same priority level as USB and if an audio softintr is blocked waiting for a usb event, they'll dead lock. also, audio being higher priority is a fine idea. - revert the previous audio.c changes as they were unsafe and this change obviates the need for them. XXX: convert the audio softint's to use a high priority thread or a XXX: workqueue would be better again.
2011-12-13 16:26:32 +04:00
sc->sc_sih_rd = softint_establish(SOFTINT_SERIAL | SOFTINT_MPSAFE,
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
audio_softintr_rd, sc);
- convert the audio softintr's to SOFTINT_SERIAL. right now they're on the same priority level as USB and if an audio softintr is blocked waiting for a usb event, they'll dead lock. also, audio being higher priority is a fine idea. - revert the previous audio.c changes as they were unsafe and this change obviates the need for them. XXX: convert the audio softint's to use a high priority thread or a XXX: workqueue would be better again.
2011-12-13 16:26:32 +04:00
sc->sc_sih_wr = softint_establish(SOFTINT_SERIAL | SOFTINT_MPSAFE,
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
audio_softintr_wr, sc);
For the "record.port", choose "record.source" over "inputs.source" if the former exists. Partial fix for PR kern/29178.
2005-02-14 02:53:20 +03:00
iclass = mclass = oclass = rclass = -1;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
sc->sc_inports.index = -1;
Back out revision 1.132. This revision broke binary compatibility with binaries which were working fine on NetBSD 1.5. We defer this change until a longer-term fix is found as explain in the commit message in revision 1.132. Further details can be found in PR17159.
2002-06-23 05:36:07 +04:00
sc->sc_inports.master = -1;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
sc->sc_inports.nports = 0;
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
sc->sc_inports.isenum = false;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
sc->sc_inports.allports = 0;
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
sc->sc_inports.isdual = false;
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
sc->sc_inports.mixerout = -1;
sc->sc_inports.cur_port = -1;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
sc->sc_outports.index = -1;
Back out revision 1.132. This revision broke binary compatibility with binaries which were working fine on NetBSD 1.5. We defer this change until a longer-term fix is found as explain in the commit message in revision 1.132. Further details can be found in PR17159.
2002-06-23 05:36:07 +04:00
sc->sc_outports.master = -1;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
sc->sc_outports.nports = 0;
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
sc->sc_outports.isenum = false;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
sc->sc_outports.allports = 0;
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
sc->sc_outports.isdual = false;
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
sc->sc_outports.mixerout = -1;
sc->sc_outports.cur_port = -1;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
sc->sc_monitor_port = -1;
Comment as to why we run through the devinfo list twice, in audioattach().
2005-02-14 02:50:22 +03:00
/*
* Read through the underlying driver's list, picking out the class
* names from the mixer descriptions. We'll need them to decode the
* mixer descriptions on the next pass through the loop.
*/
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
for(mi.index = 0; ; mi.index++) {
if (hwp->query_devinfo(hdlp, &mi) != 0)
break;
For the "record.port", choose "record.source" over "inputs.source" if the former exists. Partial fix for PR kern/29178.
2005-02-14 02:53:20 +03:00
/*
* The type of AUDIO_MIXER_CLASS merely introduces a class.
* All the other types describe an actual mixer.
*/
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
if (mi.type == AUDIO_MIXER_CLASS) {
For the "record.port", choose "record.source" over "inputs.source" if the former exists. Partial fix for PR kern/29178.
2005-02-14 02:53:20 +03:00
if (strcmp(mi.label.name, AudioCinputs) == 0)
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
iclass = mi.mixer_class;
if (strcmp(mi.label.name, AudioCmonitor) == 0)
mclass = mi.mixer_class;
if (strcmp(mi.label.name, AudioCoutputs) == 0)
oclass = mi.mixer_class;
For the "record.port", choose "record.source" over "inputs.source" if the former exists. Partial fix for PR kern/29178.
2005-02-14 02:53:20 +03:00
if (strcmp(mi.label.name, AudioCrecord) == 0)
rclass = mi.mixer_class;
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
}
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
/* Allocate save area. Ensure non-zero allocation. */
sc->sc_nmixer_states = mi.index;
sc->sc_mixer_state = kmem_alloc(sizeof(mixer_ctrl_t) *
sc->sc_nmixer_states + 1, KM_SLEEP);
Comment as to why we run through the devinfo list twice, in audioattach().
2005-02-14 02:50:22 +03:00
/*
* This is where we assign each control in the "audio" model, to the
* underlying "mixer" control. We walk through the whole list once,
* assigning likely candidates as we come across them.
*/
For the "record.port", choose "record.source" over "inputs.source" if the former exists. Partial fix for PR kern/29178.
2005-02-14 02:53:20 +03:00
record_master_found = 0;
record_source_found = 0;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
for(mi.index = 0; ; mi.index++) {
if (hwp->query_devinfo(hdlp, &mi) != 0)
break;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mi.index < sc->sc_nmixer_states);
Do not permit a class to be recorded as a master or monitor port.
1998-11-25 16:44:13 +03:00
if (mi.type == AUDIO_MIXER_CLASS)
continue;
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
if (mi.mixer_class == iclass) {
For the "record.port", choose "record.source" over "inputs.source" if the former exists. Partial fix for PR kern/29178.
2005-02-14 02:53:20 +03:00
/*
* AudioCinputs is only a fallback, when we don't
* find what we're looking for in AudioCrecord, so
* check the flags before accepting one of these.
*/
if (strcmp(mi.label.name, AudioNmaster) == 0
&& record_master_found == 0)
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
sc->sc_inports.master = mi.index;
For the "record.port", choose "record.source" over "inputs.source" if the former exists. Partial fix for PR kern/29178.
2005-02-14 02:53:20 +03:00
if (strcmp(mi.label.name, AudioNsource) == 0
&& record_source_found == 0) {
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
if (mi.type == AUDIO_MIXER_ENUM) {
int i;
for(i = 0; i < mi.un.e.num_mem; i++)
if (strcmp(mi.un.e.member[i].label.name,
AudioNmixerout) == 0)
sc->sc_inports.mixerout =
mi.un.e.member[i].ord;
}
au_setup_ports(sc, &sc->sc_inports, &mi,
itable);
}
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
if (strcmp(mi.label.name, AudioNdac) == 0 &&
sc->sc_outports.master == -1)
sc->sc_outports.master = mi.index;
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
} else if (mi.mixer_class == mclass) {
if (strcmp(mi.label.name, AudioNmonitor) == 0)
sc->sc_monitor_port = mi.index;
} else if (mi.mixer_class == oclass) {
if (strcmp(mi.label.name, AudioNmaster) == 0)
sc->sc_outports.master = mi.index;
if (strcmp(mi.label.name, AudioNselect) == 0)
au_setup_ports(sc, &sc->sc_outports, &mi,
otable);
For the "record.port", choose "record.source" over "inputs.source" if the former exists. Partial fix for PR kern/29178.
2005-02-14 02:53:20 +03:00
} else if (mi.mixer_class == rclass) {
/*
* These are the preferred mixers for the audio record
* controls, so set the flags here, but don't check.
*/
if (strcmp(mi.label.name, AudioNmaster) == 0) {
sc->sc_inports.master = mi.index;
record_master_found = 1;
}
#if 1 /* Deprecated. Use AudioNmaster. */
if (strcmp(mi.label.name, AudioNrecord) == 0) {
sc->sc_inports.master = mi.index;
record_master_found = 1;
}
if (strcmp(mi.label.name, AudioNvolume) == 0) {
sc->sc_inports.master = mi.index;
record_master_found = 1;
}
#endif
if (strcmp(mi.label.name, AudioNsource) == 0) {
if (mi.type == AUDIO_MIXER_ENUM) {
int i;
for(i = 0; i < mi.un.e.num_mem; i++)
if (strcmp(mi.un.e.member[i].label.name,
AudioNmixerout) == 0)
sc->sc_inports.mixerout =
mi.un.e.member[i].ord;
}
au_setup_ports(sc, &sc->sc_inports, &mi,
itable);
record_source_found = 1;
}
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
}
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
Add some more debug output.
2000-05-01 21:10:40 +04:00
DPRINTF(("audio_attach: inputs ports=0x%x, input master=%d, "
"output ports=0x%x, output master=%d\n",
sc->sc_inports.allports, sc->sc_inports.master,
sc->sc_outports.allports, sc->sc_outports.master));
Halt input/output on soft suspend if currently running, and restore them on soft resume.
2006-06-19 14:19:08 +04:00
Welcome to 4.99.55: - Add a lot of missing selinit() and seldestroy() calls. - Merge selwakeup() and selnotify() calls into a single selnotify(). - Add an additional 'events' argument to selnotify() call. It will indicate which event (POLL_IN, POLL_OUT, etc) happen. If unknown, zero may be used. Note: please pass appropriate value of 'events' where possible. Proposed on: <tech-kern>
2008-03-01 17:16:49 +03:00
selinit(&sc->sc_rsel);
selinit(&sc->sc_wsel);
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#ifdef AUDIO_PM_IDLE
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
callout_init(&sc->sc_idle_counter, 0);
callout_setfunc(&sc->sc_idle_counter, audio_idle, self);
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#endif
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
if (!pmf_device_register(self, audio_suspend, audio_resume))
aprint_error_dev(self, "couldn't establish power handler\n");
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#ifdef AUDIO_PM_IDLE
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
if (!device_active_register(self, audio_activity))
aprint_error_dev(self, "couldn't register activity handler\n");
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#endif
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
if (!pmf_event_register(self, PMFE_AUDIO_VOLUME_DOWN,
audio_volume_down, true))
aprint_error_dev(self, "couldn't add volume down handler\n");
if (!pmf_event_register(self, PMFE_AUDIO_VOLUME_UP,
audio_volume_up, true))
aprint_error_dev(self, "couldn't add volume up handler\n");
if (!pmf_event_register(self, PMFE_AUDIO_VOLUME_TOGGLE,
audio_volume_toggle, true))
aprint_error_dev(self, "couldn't add volume toggle handler\n");
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#ifdef AUDIO_PM_IDLE
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
callout_schedule(&sc->sc_idle_counter, audio_idle_timeout * hz);
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#endif
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
}
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
int
Split device_t and softc for audio(4), and other related cosmetic changes.
2008-03-04 21:23:44 +03:00
audioactivate(device_t self, enum devact act)
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
{
Simplify these device-activation hooks using the following semantic patch. XXX sc_dying must die. @@ type device_t; identifier act, midi_softc, midiactivate, sc, self; @@ int midiactivate(device_t self, enum devact act) { ( struct midi_softc *sc = device_private(self); | - struct midi_softc *sc; + struct midi_softc *sc = device_private(self); ... - sc = device_private(self); ) ... switch (act) { - case DVACT_ACTIVATE: - return (EOPNOTSUPP); - case DVACT_DEACTIVATE: ( sc->dying | sc->sc_dying ) = ( 1 | true ) ; - break; + return 0; + default: + return EOPNOTSUPP; } - return (0); }
2009-12-07 01:42:48 +03:00
struct audio_softc *sc = device_private(self);
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
switch (act) {
case DVACT_DEACTIVATE:
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
sc->sc_dying = true;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
Simplify these device-activation hooks using the following semantic patch. XXX sc_dying must die. @@ type device_t; identifier act, midi_softc, midiactivate, sc, self; @@ int midiactivate(device_t self, enum devact act) { ( struct midi_softc *sc = device_private(self); | - struct midi_softc *sc; + struct midi_softc *sc = device_private(self); ... - sc = device_private(self); ) ... switch (act) { - case DVACT_ACTIVATE: - return (EOPNOTSUPP); - case DVACT_DEACTIVATE: ( sc->dying | sc->sc_dying ) = ( 1 | true ) ; - break; + return 0; + default: + return EOPNOTSUPP; } - return (0); }
2009-12-07 01:42:48 +03:00
return 0;
default:
return EOPNOTSUPP;
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
}
}
int
Split device_t and softc for audio(4), and other related cosmetic changes.
2008-03-04 21:23:44 +03:00
audiodetach(device_t self, int flags)
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
{
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_softc *sc;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
int maj, mn, i;
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
Split device_t and softc for audio(4), and other related cosmetic changes.
2008-03-04 21:23:44 +03:00
sc = device_private(self);
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
DPRINTF(("audio_detach: sc=%p flags=%d\n", sc, flags));
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
/* Start draining existing accessors of the device. */
mutex_enter(sc->sc_lock);
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
sc->sc_dying = true;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
cv_broadcast(&sc->sc_wchan);
cv_broadcast(&sc->sc_rchan);
mutex_exit(sc->sc_lock);
/* locate the major number */
maj = cdevsw_lookup_major(&audio_cdevsw);
/*
* Nuke the vnodes for any open instances (calls close).
* Will wait until any activity on the device nodes has ceased.
*
* XXXAD NOT YET.
*
* XXXAD NEED TO PREVENT NEW REFERENCES THROUGH AUDIO_ENTER().
*/
mn = device_unit(self);
vdevgone(maj, mn | SOUND_DEVICE, mn | SOUND_DEVICE, VCHR);
vdevgone(maj, mn | AUDIO_DEVICE, mn | AUDIO_DEVICE, VCHR);
vdevgone(maj, mn | AUDIOCTL_DEVICE, mn | AUDIOCTL_DEVICE, VCHR);
vdevgone(maj, mn | MIXER_DEVICE, mn | MIXER_DEVICE, VCHR);
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
pmf_event_deregister(self, PMFE_AUDIO_VOLUME_DOWN,
audio_volume_down, true);
pmf_event_deregister(self, PMFE_AUDIO_VOLUME_UP,
audio_volume_up, true);
pmf_event_deregister(self, PMFE_AUDIO_VOLUME_TOGGLE,
audio_volume_toggle, true);
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#ifdef AUDIO_PM_IDLE
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
callout_halt(&sc->sc_idle_counter, sc->sc_lock);
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
Deregister activity handler on detach.
2007-12-23 22:14:03 +03:00
device_active_deregister(self, audio_activity);
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#endif
Deregister activity handler on detach.
2007-12-23 22:14:03 +03:00
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
pmf_device_deregister(self);
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
/* free resources */
audio_free_ring(sc, &sc->sc_pr);
audio_free_ring(sc, &sc->sc_rr);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
for (i = 0; i < sc->sc_nrfilters; i++) {
sc->sc_rfilters[i]->dtor(sc->sc_rfilters[i]);
sc->sc_rfilters[i] = NULL;
audio_stream_dtor(&sc->sc_rstreams[i]);
}
sc->sc_nrfilters = 0;
for (i = 0; i < sc->sc_npfilters; i++) {
sc->sc_pfilters[i]->dtor(sc->sc_pfilters[i]);
sc->sc_pfilters[i] = NULL;
audio_stream_dtor(&sc->sc_pstreams[i]);
}
sc->sc_npfilters = 0;
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
if (sc->sc_sih_rd) {
Use the softint API.
2007-10-08 20:18:02 +04:00
softint_disestablish(sc->sc_sih_rd);
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
sc->sc_sih_rd = NULL;
}
if (sc->sc_sih_wr) {
Use the softint API.
2007-10-08 20:18:02 +04:00
softint_disestablish(sc->sc_sih_wr);
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
sc->sc_sih_wr = NULL;
}
Halt input/output on soft suspend if currently running, and restore them on soft resume.
2006-06-19 14:19:08 +04:00
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#ifdef AUDIO_PM_IDLE
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
callout_destroy(&sc->sc_idle_counter);
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#endif
Welcome to 4.99.55: - Add a lot of missing selinit() and seldestroy() calls. - Merge selwakeup() and selnotify() calls into a single selnotify(). - Add an additional 'events' argument to selnotify() call. It will indicate which event (POLL_IN, POLL_OUT, etc) happen. If unknown, zero may be used. Note: please pass appropriate value of 'events' where possible. Proposed on: <tech-kern>
2008-03-01 17:16:49 +03:00
seldestroy(&sc->sc_rsel);
seldestroy(&sc->sc_wsel);
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
cv_destroy(&sc->sc_rchan);
cv_destroy(&sc->sc_wchan);
cv_destroy(&sc->sc_lchan);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return 0;
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
}
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
int
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
au_portof(struct audio_softc *sc, char *name, int class)
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
{
mixer_devinfo_t mi;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
for(mi.index = 0;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
sc->hw_if->query_devinfo(sc->hw_hdl, &mi) == 0;
mi.index++)
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
if (mi.mixer_class == class && strcmp(mi.label.name, name) == 0)
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
return mi.index;
return -1;
}
void
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
au_setup_ports(struct audio_softc *sc, struct au_mixer_ports *ports,
mixer_devinfo_t *mi, const struct portname *tbl)
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
{
int i, j;
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
ports->index = mi->index;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
if (mi->type == AUDIO_MIXER_ENUM) {
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
ports->isenum = true;
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
for(i = 0; tbl[i].name; i++)
for(j = 0; j < mi->un.e.num_mem; j++)
if (strcmp(mi->un.e.member[j].label.name,
tbl[i].name) == 0) {
ports->allports |= tbl[i].mask;
ports->aumask[ports->nports] = tbl[i].mask;
ports->misel[ports->nports] =
mi->un.e.member[j].ord;
ports->miport[ports->nports] =
au_portof(sc, mi->un.e.member[j].label.name,
mi->mixer_class);
if (ports->mixerout != -1 &&
au_setup_ports - don't hide ports->nports++ inside another expression, especially if that other expression is the right hand side operand of && which is short-circuiting. The effect was that audiorecord(1) would always fail trying to set recording port with AUDIO_SETINFO. Tested on wss@pnpbios. It looks like this problem has been reported as PR/35239, filed against port-hp700.
2007-08-15 01:34:03 +04:00
ports->miport[ports->nports] != -1)
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
ports->isdual = true;
au_setup_ports - don't hide ports->nports++ inside another expression, especially if that other expression is the right hand side operand of && which is short-circuiting. The effect was that audiorecord(1) would always fail trying to set recording port with AUDIO_SETINFO. Tested on wss@pnpbios. It looks like this problem has been reported as PR/35239, filed against port-hp700.
2007-08-15 01:34:03 +04:00
++ports->nports;
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
}
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
} else if (mi->type == AUDIO_MIXER_SET) {
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
for(i = 0; tbl[i].name; i++)
for(j = 0; j < mi->un.s.num_mem; j++)
if (strcmp(mi->un.s.member[j].label.name,
tbl[i].name) == 0) {
ports->allports |= tbl[i].mask;
ports->aumask[ports->nports] = tbl[i].mask;
ports->misel[ports->nports] =
mi->un.s.member[j].mask;
au_setup_ports - don't hide ports->nports++ inside another expression, especially if that other expression is the right hand side operand of && which is short-circuiting. The effect was that audiorecord(1) would always fail trying to set recording port with AUDIO_SETINFO. Tested on wss@pnpbios. It looks like this problem has been reported as PR/35239, filed against port-hp700.
2007-08-15 01:34:03 +04:00
ports->miport[ports->nports] =
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
au_portof(sc, mi->un.s.member[j].label.name,
mi->mixer_class);
au_setup_ports - don't hide ports->nports++ inside another expression, especially if that other expression is the right hand side operand of && which is short-circuiting. The effect was that audiorecord(1) would always fail trying to set recording port with AUDIO_SETINFO. Tested on wss@pnpbios. It looks like this problem has been reported as PR/35239, filed against port-hp700.
2007-08-15 01:34:03 +04:00
++ports->nports;
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
}
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
}
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Revert previous. Not only audio_attach_mi() but also audioprint() have to be separated. midi.c has the same problem, and a little more complicated. These will be revisited later.
2014-11-01 10:54:18 +03:00
/*
* Called from hardware driver. This is where the MI audio driver gets
* probed/attached to the hardware driver.
*/
device_t
audio_attach_mi(const struct audio_hw_if *ahwp, void *hdlp, device_t dev)
{
struct audio_attach_args arg;
#ifdef DIAGNOSTIC
if (ahwp == NULL) {
aprint_error("audio_attach_mi: NULL\n");
return 0;
}
#endif
arg.type = AUDIODEV_TYPE_AUDIO;
arg.hwif = ahwp;
arg.hdl = hdlp;
return config_found(dev, &arg, audioprint);
}
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
#ifdef AUDIO_DEBUG
ANSIfy.
2001-10-03 03:31:54 +04:00
void audio_printsc(struct audio_softc *);
minor KNF
2005-01-15 19:23:03 +03:00
void audio_print_params(const char *, struct audio_params *);
New audio subsystem
1995-02-21 04:35:58 +03:00
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
void
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_printsc(struct audio_softc *sc)
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
{
printf("hwhandle %p hw_if %p ", sc->hw_hdl, sc->hw_if);
printf("open 0x%x mode 0x%x\n", sc->sc_open, sc->sc_mode);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
printf("rchan 0x%x wchan 0x%x ", cv_has_waiters(&sc->sc_rchan),
cv_has_waiters(&sc->sc_wchan));
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
printf("rring used 0x%x pring used=%d\n",
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
audio_stream_get_used(&sc->sc_rr.s),
audio_stream_get_used(&sc->sc_pr.s));
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
printf("rbus 0x%x pbus 0x%x ", sc->sc_rbus, sc->sc_pbus);
printf("blksize %d", sc->sc_pr.blksize);
printf("hiwat %d lowat %d\n", sc->sc_pr.usedhigh, sc->sc_pr.usedlow);
}
New audio subsystem
1995-02-21 04:35:58 +03:00
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
void
minor KNF
2005-01-15 19:23:03 +03:00
audio_print_params(const char *s, struct audio_params *p)
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
{
minor KNF
2005-01-15 19:23:03 +03:00
printf("%s enc=%u %uch %u/%ubit %uHz\n", s, p->encoding, p->channels,
p->validbits, p->precision, p->sample_rate);
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
}
#endif
int
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
audio_alloc_ring(struct audio_softc *sc, struct audio_ringbuffer *r,
ANSIfy.
2001-10-03 03:31:54 +04:00
int direction, size_t bufsize)
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
{
minor KNF
2005-01-15 19:23:03 +03:00
const struct audio_hw_if *hw;
void *hdl;
hw = sc->hw_if;
hdl = sc->hw_hdl;
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
/*
* Alloc DMA play and record buffers
*/
if (bufsize < AUMINBUF)
bufsize = AUMINBUF;
Pass the direction to the allocm and round_buffersize methods. Some drivers need this to properly allocate DMAable memory.
1999-02-17 05:37:38 +03:00
ROUNDSIZE(bufsize);
don't hold the thread lock while calling allocm() or freem(). fixes PR#46121
2012-04-06 10:15:13 +04:00
if (hw->round_buffersize) {
mutex_enter(sc->sc_lock);
Pass the direction to the allocm and round_buffersize methods. Some drivers need this to properly allocate DMAable memory.
1999-02-17 05:37:38 +03:00
bufsize = hw->round_buffersize(hdl, direction, bufsize);
don't hold the thread lock while calling allocm() or freem(). fixes PR#46121
2012-04-06 10:15:13 +04:00
mutex_exit(sc->sc_lock);
}
Avoid the name free so the MALLOCLOG option works again. Fix some typos.
1998-03-03 12:16:15 +03:00
if (hw->allocm)
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
r->s.start = hw->allocm(hdl, direction, bufsize);
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
else
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
r->s.start = kmem_alloc(bufsize, KM_SLEEP);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (r->s.start == 0)
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
return ENOMEM;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
r->s.bufsize = bufsize;
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
return 0;
}
void
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_free_ring(struct audio_softc *sc, struct audio_ringbuffer *r)
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
{
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
if (r->s.start == 0)
return;
Pass the direction to the allocm and round_buffersize methods. Some drivers need this to properly allocate DMAable memory.
1999-02-17 05:37:38 +03:00
if (sc->hw_if->freem)
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
sc->hw_if->freem(sc->hw_hdl, r->s.start, r->s.bufsize);
Pass the direction to the allocm and round_buffersize methods. Some drivers need this to properly allocate DMAable memory.
1999-02-17 05:37:38 +03:00
else
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
kmem_free(r->s.start, r->s.bufsize);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
r->s.start = 0;
}
static int
audio_setup_pfilters(struct audio_softc *sc, const audio_params_t *pp,
stream_filter_list_t *pfilters)
{
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
stream_filter_t *pf[AUDIO_MAX_FILTERS], *of[AUDIO_MAX_FILTERS];
audio_stream_t ps[AUDIO_MAX_FILTERS], os[AUDIO_MAX_FILTERS];
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
const audio_params_t *from_param;
audio_params_t *to_param;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
int i, n, onfilters;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
Protect audio_write's filter graph from being modified in the middle of a write. Fixes the 'audioctl of death' problem (PR#32563).
2006-04-18 23:15:27 +04:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
/* Construct new filters. */
mutex_exit(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
memset(pf, 0, sizeof(pf));
memset(ps, 0, sizeof(ps));
from_param = pp;
for (i = 0; i < pfilters->req_size; i++) {
n = pfilters->req_size - i - 1;
to_param = &pfilters->filters[n].param;
audio_check_params(to_param);
pf[i] = pfilters->filters[n].factory(sc, from_param, to_param);
if (pf[i] == NULL)
break;
if (audio_stream_ctor(&ps[i], from_param, AU_RING_SIZE))
break;
if (i > 0)
pf[i]->set_fetcher(pf[i], &pf[i - 1]->base);
from_param = to_param;
}
if (i < pfilters->req_size) { /* failure */
DPRINTF(("%s: pfilters failure\n", __func__));
for (; i >= 0; i--) {
if (pf[i] != NULL)
pf[i]->dtor(pf[i]);
audio_stream_dtor(&ps[i]);
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return EINVAL;
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
/* Swap in new filters. */
mutex_enter(sc->sc_intr_lock);
memcpy(of, sc->sc_pfilters, sizeof(of));
memcpy(os, sc->sc_pstreams, sizeof(os));
onfilters = sc->sc_npfilters;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
memcpy(sc->sc_pfilters, pf, sizeof(pf));
memcpy(sc->sc_pstreams, ps, sizeof(ps));
sc->sc_npfilters = pfilters->req_size;
for (i = 0; i < pfilters->req_size; i++) {
pf[i]->set_inputbuffer(pf[i], &sc->sc_pstreams[i]);
}
/* hardware format and the buffer near to userland */
if (pfilters->req_size <= 0) {
sc->sc_pr.s.param = *pp;
sc->sc_pustream = &sc->sc_pr.s;
} else {
sc->sc_pr.s.param = pfilters->filters[0].param;
sc->sc_pustream = &sc->sc_pstreams[0];
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
/* Destroy old filters. */
mutex_exit(sc->sc_lock);
for (i = 0; i < onfilters; i++) {
of[i]->dtor(of[i]);
audio_stream_dtor(&os[i]);
}
mutex_enter(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
#ifdef AUDIO_DEBUG
printf("%s: HW-buffer=%p pustream=%p\n",
__func__, &sc->sc_pr.s, sc->sc_pustream);
for (i = 0; i < pfilters->req_size; i++) {
char num[100];
snprintf(num, 100, "[%d]", i);
audio_print_params(num, &sc->sc_pstreams[i].param);
}
audio_print_params("[HW]", &sc->sc_pr.s.param);
#endif /* AUDIO_DEBUG */
Protect audio_write's filter graph from being modified in the middle of a write. Fixes the 'audioctl of death' problem (PR#32563).
2006-04-18 23:15:27 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return 0;
}
static int
audio_setup_rfilters(struct audio_softc *sc, const audio_params_t *rp,
stream_filter_list_t *rfilters)
{
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
stream_filter_t *rf[AUDIO_MAX_FILTERS], *of[AUDIO_MAX_FILTERS];
audio_stream_t rs[AUDIO_MAX_FILTERS], os[AUDIO_MAX_FILTERS];
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
const audio_params_t *to_param;
audio_params_t *from_param;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
int i, onfilters;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
/* Construct new filters. */
mutex_exit(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
memset(rf, 0, sizeof(rf));
memset(rs, 0, sizeof(rs));
for (i = 0; i < rfilters->req_size; i++) {
from_param = &rfilters->filters[i].param;
audio_check_params(from_param);
to_param = i + 1 < rfilters->req_size
? &rfilters->filters[i + 1].param : rp;
rf[i] = rfilters->filters[i].factory(sc, from_param, to_param);
if (rf[i] == NULL)
break;
if (audio_stream_ctor(&rs[i], to_param, AU_RING_SIZE))
break;
if (i > 0) {
rf[i]->set_fetcher(rf[i], &rf[i - 1]->base);
} else {
/* rf[0] has no previous fetcher because
* the audio hardware fills data to the
* input buffer. */
rf[0]->set_inputbuffer(rf[0], &sc->sc_rr.s);
}
}
if (i < rfilters->req_size) { /* failure */
DPRINTF(("%s: rfilters failure\n", __func__));
for (; i >= 0; i--) {
if (rf[i] != NULL)
rf[i]->dtor(rf[i]);
audio_stream_dtor(&rs[i]);
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return EINVAL;
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
/* Swap in new filters. */
mutex_enter(sc->sc_intr_lock);
memcpy(of, sc->sc_rfilters, sizeof(of));
memcpy(os, sc->sc_rstreams, sizeof(os));
onfilters = sc->sc_nrfilters;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
memcpy(sc->sc_rfilters, rf, sizeof(rf));
memcpy(sc->sc_rstreams, rs, sizeof(rs));
sc->sc_nrfilters = rfilters->req_size;
for (i = 1; i < rfilters->req_size; i++) {
rf[i]->set_inputbuffer(rf[i], &sc->sc_rstreams[i - 1]);
}
/* hardware format and the buffer near to userland */
if (rfilters->req_size <= 0) {
sc->sc_rr.s.param = *rp;
sc->sc_rustream = &sc->sc_rr.s;
} else {
sc->sc_rr.s.param = rfilters->filters[0].param;
sc->sc_rustream = &sc->sc_rstreams[rfilters->req_size - 1];
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
#ifdef AUDIO_DEBUG
printf("%s: HW-buffer=%p pustream=%p\n",
__func__, &sc->sc_rr.s, sc->sc_rustream);
audio_print_params("[HW]", &sc->sc_rr.s.param);
for (i = 0; i < rfilters->req_size; i++) {
char num[100];
snprintf(num, 100, "[%d]", i);
audio_print_params(num, &sc->sc_rstreams[i].param);
}
#endif /* AUDIO_DEBUG */
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
/* Destroy old filters. */
mutex_exit(sc->sc_lock);
for (i = 0; i < onfilters; i++) {
of[i]->dtor(of[i]);
audio_stream_dtor(&os[i]);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
return 0;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
}
static void
audio_stream_dtor(audio_stream_t *stream)
{
minor KNF
2005-01-15 19:23:03 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (stream->start != NULL)
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
kmem_free(stream->start, stream->bufsize);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
memset(stream, 0, sizeof(audio_stream_t));
}
static int
audio_stream_ctor(audio_stream_t *stream, const audio_params_t *param, int size)
{
int frame_size;
size = min(size, AU_RING_SIZE);
stream->bufsize = size;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
stream->start = kmem_alloc(size, KM_SLEEP);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (stream->start == NULL)
return ENOMEM;
Fix panic when precision < 8. From kent@.
2005-01-17 17:13:19 +03:00
frame_size = (param->precision + 7) / 8 * param->channels;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
size = (size / frame_size) * frame_size;
stream->end = stream->start + size;
stream->inp = stream->start;
stream->outp = stream->start;
stream->used = 0;
stream->param = *param;
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
stream->loop = false;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return 0;
}
static void
stream_filter_list_append(stream_filter_list_t *list,
stream_filter_factory_t factory,
const audio_params_t *param)
{
minor KNF
2005-01-15 19:23:03 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (list->req_size >= AUDIO_MAX_FILTERS) {
printf("%s: increase AUDIO_MAX_FILTERS in sys/dev/audio_if.h\n",
__func__);
return;
}
list->filters[list->req_size].factory = factory;
list->filters[list->req_size].param = *param;
list->req_size++;
}
static void
stream_filter_list_set(stream_filter_list_t *list, int i,
stream_filter_factory_t factory,
const audio_params_t *param)
{
Protect audio_write's filter graph from being modified in the middle of a write. Fixes the 'audioctl of death' problem (PR#32563).
2006-04-18 23:15:27 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (i < 0 || i >= AUDIO_MAX_FILTERS) {
printf("%s: invalid index: %d\n", __func__, i);
return;
}
Protect audio_write's filter graph from being modified in the middle of a write. Fixes the 'audioctl of death' problem (PR#32563).
2006-04-18 23:15:27 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
list->filters[i].factory = factory;
list->filters[i].param = *param;
if (list->req_size <= i)
list->req_size = i + 1;
}
static void
stream_filter_list_prepend(stream_filter_list_t *list,
stream_filter_factory_t factory,
const audio_params_t *param)
{
minor KNF
2005-01-15 19:23:03 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (list->req_size >= AUDIO_MAX_FILTERS) {
printf("%s: increase AUDIO_MAX_FILTERS in sys/dev/audio_if.h\n",
__func__);
return;
}
memmove(&list->filters[1], &list->filters[0],
sizeof(struct stream_filter_req) * list->req_size);
list->filters[0].factory = factory;
list->filters[0].param = *param;
list->req_size++;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
/*
* Look up audio device and acquire locks for device access.
*/
static int
audio_enter(dev_t dev, krw_t rw, struct audio_softc **scp)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
struct audio_softc *sc;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
/* First, find the device and take sc_lock. */
device_private(device_lookup()) -> device_lookup_private() ok cube@
2008-06-11 02:53:08 +04:00
sc = device_lookup_private(&audio_cd, AUDIOUNIT(dev));
Use device_lookup().
2000-07-06 04:43:04 +04:00
if (sc == NULL)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return ENXIO;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
if (sc->sc_dying) {
mutex_exit(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return EIO;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
}
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
/* Acquire device access lock. */
switch (rw) {
case RW_WRITER:
while (__predict_false(sc->sc_dvlock != 0)) {
cv_wait(&sc->sc_lchan, sc->sc_lock);
}
sc->sc_dvlock = -1;
break;
case RW_READER:
while (__predict_false(sc->sc_dvlock < 0)) {
cv_wait(&sc->sc_lchan, sc->sc_lock);
}
sc->sc_dvlock++;
break;
default:
panic("audio_enter");
}
*scp = sc;
return 0;
}
/*
* Release reference to device acquired with audio_enter().
*/
static void
audio_exit(struct audio_softc *sc)
{
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
KASSERT(sc->sc_dvlock != 0);
Add a 'powerstate' callback to audio_hw_if, to allow hw drivers to power themselves down when not in use.
2006-03-07 18:17:47 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
/* Release device level lock. */
if (__predict_false(sc->sc_dvlock < 0)) {
sc->sc_dvlock = 0;
} else {
sc->sc_dvlock--;
}
cv_broadcast(&sc->sc_lchan);
mutex_exit(sc->sc_lock);
}
/*
* Wait for I/O to complete, releasing device lock.
*/
static int
audio_waitio(struct audio_softc *sc, kcondvar_t *chan)
{
int error;
krw_t rw;
KASSERT(mutex_owned(sc->sc_lock));
/* Release device level lock while sleeping. */
if (__predict_false(sc->sc_dvlock < 0)) {
sc->sc_dvlock = 0;
rw = RW_WRITER;
} else {
KASSERT(sc->sc_dvlock > 0);
sc->sc_dvlock--;
rw = RW_READER;
}
cv_broadcast(&sc->sc_lchan);
/* Wait for pending I/O to complete. */
error = cv_wait_sig(chan, sc->sc_lock);
/* Re-acquire device level lock. */
if (__predict_false(rw == RW_WRITER)) {
while (__predict_false(sc->sc_dvlock != 0)) {
cv_wait(&sc->sc_lchan, sc->sc_lock);
}
sc->sc_dvlock = -1;
} else {
while (__predict_false(sc->sc_dvlock < 0)) {
cv_wait(&sc->sc_lchan, sc->sc_lock);
}
sc->sc_dvlock++;
}
return error;
}
int
audioopen(dev_t dev, int flags, int ifmt, struct lwp *l)
{
struct audio_softc *sc;
int error;
if ((error = audio_enter(dev, RW_WRITER, &sc)) != 0)
return error;
device_active(sc->dev, DVA_SYSTEM);
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
switch (AUDIODEV(dev)) {
New audio subsystem
1995-02-21 04:35:58 +03:00
case SOUND_DEVICE:
case AUDIO_DEVICE:
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
error = audio_open(dev, sc, flags, ifmt, l);
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
break;
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
case AUDIOCTL_DEVICE:
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
error = 0;
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
case MIXER_DEVICE:
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
error = mixer_open(dev, sc, flags, ifmt, l);
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
default:
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
error = ENXIO;
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
audio_exit(sc);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
int
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
audioclose(dev_t dev, int flags, int ifmt, struct lwp *l)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_softc *sc;
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
int error;
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if ((error = audio_enter(dev, RW_WRITER, &sc)) != 0)
return error;
Split device_t and softc for audio(4), and other related cosmetic changes.
2008-03-04 21:23:44 +03:00
device_active(sc->dev, DVA_SYSTEM);
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
switch (AUDIODEV(dev)) {
New audio subsystem
1995-02-21 04:35:58 +03:00
case SOUND_DEVICE:
case AUDIO_DEVICE:
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
error = audio_close(sc, flags, ifmt, l);
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
case MIXER_DEVICE:
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
error = mixer_close(sc, flags, ifmt, l);
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
break;
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
case AUDIOCTL_DEVICE:
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
error = 0;
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
default:
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
error = ENXIO;
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
audio_exit(sc);
Add a 'powerstate' callback to audio_hw_if, to allow hw drivers to power themselves down when not in use.
2006-03-07 18:17:47 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
int
ANSIfy.
2001-10-03 03:31:54 +04:00
audioread(dev_t dev, struct uio *uio, int ioflag)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
struct audio_softc *sc;
int error;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if ((error = audio_enter(dev, RW_READER, &sc)) != 0)
return error;
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
switch (AUDIODEV(dev)) {
New audio subsystem
1995-02-21 04:35:58 +03:00
case SOUND_DEVICE:
case AUDIO_DEVICE:
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
error = audio_read(sc, uio, ioflag);
break;
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
case AUDIOCTL_DEVICE:
New audio subsystem
1995-02-21 04:35:58 +03:00
case MIXER_DEVICE:
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
error = ENODEV;
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
default:
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
error = ENXIO;
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
audio_exit(sc);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
int
ANSIfy.
2001-10-03 03:31:54 +04:00
audiowrite(dev_t dev, struct uio *uio, int ioflag)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
struct audio_softc *sc;
int error;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if ((error = audio_enter(dev, RW_READER, &sc)) != 0)
return error;
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
switch (AUDIODEV(dev)) {
New audio subsystem
1995-02-21 04:35:58 +03:00
case SOUND_DEVICE:
case AUDIO_DEVICE:
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
error = audio_write(sc, uio, ioflag);
break;
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
case AUDIOCTL_DEVICE:
New audio subsystem
1995-02-21 04:35:58 +03:00
case MIXER_DEVICE:
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
error = ENODEV;
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
default:
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
error = ENXIO;
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
audio_exit(sc);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
int
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.
2007-03-04 08:59:00 +03:00
audioioctl(dev_t dev, u_long cmd, void *addr, int flag, struct lwp *l)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_softc *sc;
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
int error;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
krw_t rw;
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
/* Figure out which lock type we need. */
switch (cmd) {
case AUDIO_FLUSH:
case AUDIO_SETINFO:
case AUDIO_DRAIN:
case AUDIO_SETFD:
rw = RW_WRITER;
break;
default:
rw = RW_READER;
break;
}
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if ((error = audio_enter(dev, rw, &sc)) != 0)
return error;
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
switch (AUDIODEV(dev)) {
New audio subsystem
1995-02-21 04:35:58 +03:00
case SOUND_DEVICE:
case AUDIO_DEVICE:
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
case AUDIOCTL_DEVICE:
Split device_t and softc for audio(4), and other related cosmetic changes.
2008-03-04 21:23:44 +03:00
device_active(sc->dev, DVA_SYSTEM);
Revert previous and instead select whether to call mixer_ioctl() or audio_ioctl() based on whether the command smells like a mixer ioctl or not.
2009-09-24 20:03:11 +04:00
if (IOCGROUP(cmd) == IOCGROUP(AUDIO_MIXER_READ))
OSS audio allows mixer operations on the dsp device. NetBSD would return EINVAL in these circumstances. This can break audio in apps running under Linux emulation (e.g. Citrix ICA client will mute all audio when volume control used). Therefore, pass unrecognised ioctls attempted on dsp devices to mixer_ioctl.
2009-09-24 15:13:38 +04:00
error = mixer_ioctl(sc, cmd, addr, flag, l);
Revert previous and instead select whether to call mixer_ioctl() or audio_ioctl() based on whether the command smells like a mixer ioctl or not.
2009-09-24 20:03:11 +04:00
else
error = audio_ioctl(sc, cmd, addr, flag, l);
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
case MIXER_DEVICE:
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
error = mixer_ioctl(sc, cmd, addr, flag, l);
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
default:
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
error = ENXIO;
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
audio_exit(sc);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
int
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
audiopoll(dev_t dev, int events, struct lwp *l)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_softc *sc;
PR-30566: Poll must not return <sys/errno.h> values. Start with those places I can easily test.
2005-06-21 18:01:11 +04:00
int revents;
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
/* Don't bother with device level lock here. */
sc = device_lookup_private(&audio_cd, AUDIOUNIT(dev));
if (sc == NULL)
return ENXIO;
mutex_enter(sc->sc_lock);
if (sc->sc_dying) {
mutex_exit(sc->sc_lock);
return EIO;
}
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
switch (AUDIODEV(dev)) {
New audio subsystem
1995-02-21 04:35:58 +03:00
case SOUND_DEVICE:
case AUDIO_DEVICE:
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
revents = audio_poll(sc, events, l);
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
break;
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
case AUDIOCTL_DEVICE:
New audio subsystem
1995-02-21 04:35:58 +03:00
case MIXER_DEVICE:
PR-30566: Poll must not return <sys/errno.h> values. Start with those places I can easily test.
2005-06-21 18:01:11 +04:00
revents = 0;
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
default:
PR-30566: Poll must not return <sys/errno.h> values. Start with those places I can easily test.
2005-06-21 18:01:11 +04:00
revents = POLLERR;
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
break;
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
PR-30566: Poll must not return <sys/errno.h> values. Start with those places I can easily test.
2005-06-21 18:01:11 +04:00
return revents;
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
}
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
int
audiokqfilter(dev_t dev, struct knote *kn)
{
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_softc *sc;
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
int rv;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
/* Don't bother with device level lock here. */
Use device_lookup_private() rather than using cd_devs[] directly to get softc. XXX maybe we should change a type of cd_devs[] in struct cfdriver from (void *) to device_t.
2008-06-08 22:18:33 +04:00
sc = device_lookup_private(&audio_cd, AUDIOUNIT(dev));
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if (sc == NULL)
return ENXIO;
mutex_enter(sc->sc_lock);
if (sc->sc_dying) {
mutex_exit(sc->sc_lock);
return EIO;
}
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
switch (AUDIODEV(dev)) {
case SOUND_DEVICE:
case AUDIO_DEVICE:
rv = audio_kqfilter(sc, kn);
break;
case AUDIOCTL_DEVICE:
case MIXER_DEVICE:
rv = 1;
break;
default:
rv = 1;
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return rv;
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
}
Change the kernel mmap interface so that the offset to map is an "off_t" and the return value is a "paddr_t" to allow mappings at offsets past 2^31 bytes. Somewhat inspired by FreeBSD, which only changed the offset to a "vm_offset_t". Includes updates for the i386, pc532 and sh3 mmmmap from Jason Thorpe.
2000-06-26 08:55:19 +04:00
paddr_t
ANSIfy.
2001-10-03 03:31:54 +04:00
audiommap(dev_t dev, off_t off, int prot)
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
{
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_softc *sc;
Change the kernel mmap interface so that the offset to map is an "off_t" and the return value is a "paddr_t" to allow mappings at offsets past 2^31 bytes. Somewhat inspired by FreeBSD, which only changed the offset to a "vm_offset_t". Includes updates for the i386, pc532 and sh3 mmmmap from Jason Thorpe.
2000-06-26 08:55:19 +04:00
paddr_t error;
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
/*
* Acquire a reader lock. audio_mmap() will drop sc_lock
* in order to allow the device's mmap routine to sleep.
* Although not yet possible, we want to prevent memory
* from being allocated or freed out from under us.
*/
if ((error = audio_enter(dev, RW_READER, &sc)) != 0)
return 1;
Split device_t and softc for audio(4), and other related cosmetic changes.
2008-03-04 21:23:44 +03:00
device_active(sc->dev, DVA_SYSTEM); /* XXXJDM */
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
switch (AUDIODEV(dev)) {
case SOUND_DEVICE:
case AUDIO_DEVICE:
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
error = audio_mmap(sc, off, prot);
break;
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
case AUDIOCTL_DEVICE:
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
case MIXER_DEVICE:
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
error = -1;
break;
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
default:
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
error = -1;
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
audio_exit(sc);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
/*
* Audio driver
*/
Make some functions non-static.
1995-04-18 03:04:31 +04:00
void
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
audio_init_ringbuffer(struct audio_softc *sc, struct audio_ringbuffer *rp,
int mode)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
int nblks;
minor KNF
2005-01-15 19:23:03 +03:00
int blksize;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
minor KNF
2005-01-15 19:23:03 +03:00
blksize = rp->blksize;
Fix divide by 0 bug in audio driver.
1997-07-28 03:06:04 +04:00
if (blksize < AUMINBLK)
blksize = AUMINBLK;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (blksize > rp->s.bufsize / AUMINNOBLK)
blksize = rp->s.bufsize / AUMINNOBLK;
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
ROUNDSIZE(blksize);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
DPRINTF(("audio_init_ringbuffer: MI blksize=%d\n", blksize));
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
if (sc->hw_if->round_blocksize)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
blksize = sc->hw_if->round_blocksize(sc->hw_hdl, blksize,
mode, &rp->s.param);
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
if (blksize <= 0)
panic("audio_init_ringbuffer: blksize");
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
nblks = rp->s.bufsize / blksize;
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
DPRINTF(("audio_init_ringbuffer: final blksize=%d\n", blksize));
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
rp->blksize = blksize;
rp->maxblks = nblks;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
rp->s.end = rp->s.start + nblks * blksize;
rp->s.outp = rp->s.inp = rp->s.start;
rp->s.used = 0;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
rp->stamp = 0;
Fixes a bug that audio_offset_t::samples for AUDIO_GETxOFFS has incorrect value. It also affects SNDCTL_DSP_GETOPTR of ossaudio. PR: kern/32405 by Stephane Witzman
2006-05-10 05:56:21 +04:00
rp->stamp_last = 0;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
rp->fstamp = 0;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
rp->drops = 0;
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
rp->copying = false;
rp->needfill = false;
rp->mmapped = false;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
int
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_initbufs(struct audio_softc *sc)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
minor KNF
2005-01-15 19:23:03 +03:00
const struct audio_hw_if *hw;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
int error;
New audio subsystem
1995-02-21 04:35:58 +03:00
%x -> 0x%x
1997-08-06 11:39:59 +04:00
DPRINTF(("audio_initbufs: mode=0x%x\n", sc->sc_mode));
minor KNF
2005-01-15 19:23:03 +03:00
hw = sc->hw_if;
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
if (audio_can_capture(sc)) {
audio_init_ringbuffer(sc, &sc->sc_rr, AUMODE_RECORD);
if (hw->init_input && (sc->sc_mode & AUMODE_RECORD)) {
error = hw->init_input(sc->hw_hdl, sc->sc_rr.s.start,
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
sc->sc_rr.s.end - sc->sc_rr.s.start);
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
if (error)
return error;
}
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
}
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
if (audio_can_playback(sc)) {
audio_init_ringbuffer(sc, &sc->sc_pr, AUMODE_PLAY);
sc->sc_sil_count = 0;
if (hw->init_output && (sc->sc_mode & AUMODE_PLAY)) {
error = hw->init_output(sc->hw_hdl, sc->sc_pr.s.start,
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
sc->sc_pr.s.end - sc->sc_pr.s.start);
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
if (error)
return error;
}
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
}
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
Restructure a little to make sure AUDIO_FLUSH restarts the looping DMA.
1997-08-07 03:08:26 +04:00
#ifdef AUDIO_INTR_TIME
Allow option AUDIO_INTR_TIME even if the platform does not support FP in the kernel.
1998-08-04 15:26:14 +04:00
#define double u_long
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
if (audio_can_playback(sc)) {
sc->sc_pnintr = 0;
sc->sc_pblktime = (u_long)(
(double)sc->sc_pr.blksize * 100000 /
(double)(sc->sc_pparams.precision / NBBY *
sc->sc_pparams.channels *
sc->sc_pparams.sample_rate)) * 10;
DPRINTF(("audio: play blktime = %lu for %d\n",
sc->sc_pblktime, sc->sc_pr.blksize));
}
if (audio_can_capture(sc)) {
sc->sc_rnintr = 0;
sc->sc_rblktime = (u_long)(
(double)sc->sc_rr.blksize * 100000 /
(double)(sc->sc_rparams.precision / NBBY *
sc->sc_rparams.channels *
sc->sc_rparams.sample_rate)) * 10;
DPRINTF(("audio: record blktime = %lu for %d\n",
sc->sc_rblktime, sc->sc_rr.blksize));
}
Allow option AUDIO_INTR_TIME even if the platform does not support FP in the kernel.
1998-08-04 15:26:14 +04:00
#undef double
Restructure a little to make sure AUDIO_FLUSH restarts the looping DMA.
1997-08-07 03:08:26 +04:00
#endif
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
return 0;
Restructure a little to make sure AUDIO_FLUSH restarts the looping DMA.
1997-08-07 03:08:26 +04:00
}
void
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_calcwater(struct audio_softc *sc)
Restructure a little to make sure AUDIO_FLUSH restarts the looping DMA.
1997-08-07 03:08:26 +04:00
{
minor KNF
2005-01-15 19:23:03 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
/* set high at 100% */
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
if (audio_can_playback(sc)) {
sc->sc_pr.usedhigh =
sc->sc_pustream->end - sc->sc_pustream->start;
/* set low at 75% of usedhigh */
sc->sc_pr.usedlow = sc->sc_pr.usedhigh * 3 / 4;
if (sc->sc_pr.usedlow == sc->sc_pr.usedhigh)
sc->sc_pr.usedlow -= sc->sc_pr.blksize;
}
if (audio_can_capture(sc)) {
sc->sc_rr.usedhigh =
sc->sc_rustream->end - sc->sc_rustream->start -
sc->sc_rr.blksize;
sc->sc_rr.usedlow = 0;
DPRINTF(("%s: plow=%d phigh=%d rlow=%d rhigh=%d\n", __func__,
sc->sc_pr.usedlow, sc->sc_pr.usedhigh,
sc->sc_rr.usedlow, sc->sc_rr.usedhigh));
}
New audio subsystem
1995-02-21 04:35:58 +03:00
}
int
__unused removal on arguments; approved by core.
2006-11-16 04:32:37 +03:00
audio_open(dev_t dev, struct audio_softc *sc, int flags, int ifmt,
struct lwp *l)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
int error;
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
u_int mode;
constify audio_hw_if, midi_hw_if, and radio_hw_if.
2004-10-29 16:57:15 +04:00
const struct audio_hw_if *hw;
New audio subsystem
1995-02-21 04:35:58 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
New audio subsystem
1995-02-21 04:35:58 +03:00
hw = sc->hw_if;
minor KNF
2005-01-15 19:23:03 +03:00
if (hw == NULL)
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
return ENXIO;
New audio subsystem
1995-02-21 04:35:58 +03:00
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
DPRINTF(("audio_open: flags=0x%x sc=%p hdl=%p\n",
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
flags, sc, sc->hw_hdl));
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
Allow the audio device to be opened for read resp. write by separate calls to open(). Previously only one call to open() was allowed. This change should allow some Linux derived programs to work unaltered. XXX This isn't really possible to implement correctly in a simple way. So this implementation is incorrect: It is not until device has been closed by both the reader and writer that it is possible to open it at all again.
2005-05-19 00:10:25 +04:00
if (((flags & FREAD) && (sc->sc_open & AUOPEN_READ)) ||
((flags & FWRITE) && (sc->sc_open & AUOPEN_WRITE)))
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return EBUSY;
If one or more playback filters are applied, minimize loading size of the hardware buffer to improve accuracy of AUDIO_GETIOFFS.
2006-05-16 17:46:19 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (hw->open != NULL) {
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
error = hw->open(sc->hw_hdl, flags);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (error)
return error;
}
New audio subsystem
1995-02-21 04:35:58 +03:00
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
sc->sc_async_audio = 0;
sc->sc_sil_count = 0;
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
sc->sc_rbus = false;
sc->sc_pbus = false;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
sc->sc_eof = 0;
sc->sc_playdrop = 0;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
Enable full-duplex mode by default on drivers that support it if the underlying audio device was opened read/write. This is consistent with Linux OSS behaviour and fixes a bug with certain applications (including Skype) that assume this behaviour. Fixes PR# 30044.
2005-04-25 17:19:46 +04:00
sc->sc_full_duplex =
Make sure audio device starts in half-duplex mode and document the fact.
1997-08-26 23:03:55 +04:00
(flags & (FWRITE|FREAD)) == (FWRITE|FREAD) &&
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
(audio_get_props(sc) & AUDIO_PROP_FULLDUPLEX);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
mode = 0;
if (flags & FREAD) {
New audio subsystem
1995-02-21 04:35:58 +03:00
sc->sc_open |= AUOPEN_READ;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
mode |= AUMODE_RECORD;
}
if (flags & FWRITE) {
New audio subsystem
1995-02-21 04:35:58 +03:00
sc->sc_open |= AUOPEN_WRITE;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
mode |= AUMODE_PLAY | AUMODE_PLAY_ALL;
}
New audio subsystem
1995-02-21 04:35:58 +03:00
/*
* Multiplex device: /dev/audio (MU-Law) and /dev/sound (linear)
* The /dev/audio is always (re)set to 8-bit MU-Law mono
* For the other devices, you get what they were last set to.
*/
if (ISDEVAUDIO(dev)) {
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
error = audio_set_defaults(sc, mode);
} else {
struct audio_info ai;
AUDIO_INITINFO(&ai);
ai.mode = mode;
error = audiosetinfo(sc, &ai);
New audio subsystem
1995-02-21 04:35:58 +03:00
}
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
if (error)
goto bad;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
#ifdef DIAGNOSTIC
New audio subsystem
1995-02-21 04:35:58 +03:00
/*
* Sample rate and precision are supposed to be set to proper
* default values by the hardware driver, so that it may give
* us these values.
*/
Fix typo in last commit; from Thorsten Frueauf <frueauf@ira.uka.de> in PR/3566.
1997-05-01 16:27:10 +04:00
if (sc->sc_rparams.precision == 0 || sc->sc_pparams.precision == 0) {
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
printf("audio_open: 0 precision\n");
return EINVAL;
}
New audio subsystem
1995-02-21 04:35:58 +03:00
#endif
Make sure to recalculate the block size after changing the sampling rate. Also, nuke the global variable audio_blocksize.
1997-03-20 19:13:55 +03:00
The previous revision was my fault. audio_calcwater() needs to be called at every open. Thanks Witold J. Wnuk for pointing out this problem.
2001-01-25 18:25:34 +03:00
/* audio_close() decreases sc_pr.usedlow, recalculate here */
audio_calcwater(sc);
Make sure to recalculate the block size after changing the sampling rate. Also, nuke the global variable audio_blocksize.
1997-03-20 19:13:55 +03:00
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
DPRINTF(("audio_open: done sc_mode = 0x%x\n", sc->sc_mode));
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
return 0;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
bad:
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (hw->close != NULL)
hw->close(sc->hw_hdl);
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
sc->sc_open = 0;
sc->sc_mode = 0;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
sc->sc_full_duplex = 0;
return error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
/*
* Must be called from task context.
*/
void
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_init_record(struct audio_softc *sc)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
mutex_enter(sc->sc_intr_lock);
Make mode a flag set. Full-dpulex fixes from John Kohl.
1995-07-07 05:52:30 +04:00
if (sc->hw_if->speaker_ctl &&
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
(!sc->sc_full_duplex || (sc->sc_mode & AUMODE_PLAY) == 0))
New audio subsystem
1995-02-21 04:35:58 +03:00
sc->hw_if->speaker_ctl(sc->hw_hdl, SPKR_OFF);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
New audio subsystem
1995-02-21 04:35:58 +03:00
}
/*
* Must be called from task context.
*/
void
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_init_play(struct audio_softc *sc)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
mutex_enter(sc->sc_intr_lock);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
sc->sc_wstamp = sc->sc_pr.stamp;
New audio subsystem
1995-02-21 04:35:58 +03:00
if (sc->hw_if->speaker_ctl)
sc->hw_if->speaker_ctl(sc->hw_hdl, SPKR_ON);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Make some functions non-static.
1995-04-18 03:04:31 +04:00
int
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_drain(struct audio_softc *sc)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_ringbuffer *cb;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
int error, drops;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
int i, used;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
KASSERT(mutex_owned(sc->sc_intr_lock));
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
DPRINTF(("audio_drain: enter busy=%d\n", sc->sc_pbus));
minor KNF
2005-01-15 19:23:03 +03:00
cb = &sc->sc_pr;
if (cb->mmapped)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return 0;
used = audio_stream_get_used(&sc->sc_pr.s);
for (i = 0; i < sc->sc_npfilters; i++)
used += audio_stream_get_used(&sc->sc_pstreams[i]);
if (used <= 0)
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
return 0;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
if (!sc->sc_pbus) {
/* We've never started playing, probably because the
* block was too short. Pad it and start now.
*/
Move around splaudio() to (hopefully) get rid of some rare race conditions. Add a missing call to round_blocksize. Add some MIDI stuff that will be used in the future.
1997-09-06 05:14:48 +04:00
int cc;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
uint8_t *inp = cb->s.inp;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
cc = cb->blksize - (inp - cb->s.start) % cb->blksize;
audio_fill_silence(&cb->s.param, inp, cc);
cb->s.inp = audio_stream_add_inp(&cb->s, inp, cc);
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
error = audiostartp(sc);
if (error)
return error;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
}
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
/*
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
* Play until a silence block has been played, then we
* know all has been drained.
* XXX This should be done some other way to avoid
* playing silence.
*/
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
#ifdef DIAGNOSTIC
if (cb->copying) {
printf("audio_drain: copying in progress!?!\n");
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
cb->copying = false;
Change the MI audio driver so it attaches to the MD driver in the normal way. This requires adding a line to the config files to get audio to work again.
1997-08-20 03:49:33 +04:00
}
#endif
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
drops = cb->drops;
Move around splaudio() to (hopefully) get rid of some rare race conditions. Add a missing call to round_blocksize. Add some MIDI stuff that will be used in the future.
1997-09-06 05:14:48 +04:00
error = 0;
while (cb->drops == drops && !error) {
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
DPRINTF(("audio_drain: used=%d, drops=%ld\n",
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
audio_stream_get_used(&sc->sc_pr.s), cb->drops));
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
error = audio_waitio(sc, &sc->sc_wchan);
mutex_enter(sc->sc_intr_lock);
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
if (sc->sc_dying)
error = EIO;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Move around splaudio() to (hopefully) get rid of some rare race conditions. Add a missing call to round_blocksize. Add some MIDI stuff that will be used in the future.
1997-09-06 05:14:48 +04:00
return error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
/*
* Close an audio chip.
*/
/* ARGSUSED */
int
__unused removal on arguments; approved by core.
2006-11-16 04:32:37 +03:00
audio_close(struct audio_softc *sc, int flags, int ifmt,
struct lwp *l)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
minor KNF
2005-01-15 19:23:03 +03:00
const struct audio_hw_if *hw;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
New audio subsystem
1995-02-21 04:35:58 +03:00
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
DPRINTF(("audio_close: sc=%p\n", sc));
minor KNF
2005-01-15 19:23:03 +03:00
hw = sc->hw_if;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
/* Stop recording. */
Make sure output is thoroughly halted on close, and remember that we stopped it.
1998-08-09 11:25:58 +04:00
if ((flags & FREAD) && sc->sc_rbus) {
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
/*
Fix three problems spotted by mycroft: - Closing the audio device should stop recording. - Manipulating the pause function could accidentally start playing or recording. - AUDIO_FLUSH could accidentally start playing or recording.
1997-12-03 04:01:19 +03:00
* XXX Some drivers (e.g. SB) use the same routine
* to halt input and output so don't halt input if
* in full duplex mode. These drivers should be fixed.
*/
minor KNF
2005-01-15 19:23:03 +03:00
if (!sc->sc_full_duplex || hw->halt_input != hw->halt_output)
hw->halt_input(sc->hw_hdl);
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
sc->sc_rbus = false;
Fix three problems spotted by mycroft: - Closing the audio device should stop recording. - Manipulating the pause function could accidentally start playing or recording. - AUDIO_FLUSH could accidentally start playing or recording.
1997-12-03 04:01:19 +03:00
}
New audio subsystem
1995-02-21 04:35:58 +03:00
/*
* Block until output drains, but allow ^C interrupt.
*/
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
sc->sc_pr.usedlow = sc->sc_pr.blksize; /* avoid excessive wakeups */
New audio subsystem
1995-02-21 04:35:58 +03:00
/*
* If there is pending output, let it drain (unless
* the output is paused).
*/
Make sure output is thoroughly halted on close, and remember that we stopped it.
1998-08-09 11:25:58 +04:00
if ((flags & FWRITE) && sc->sc_pbus) {
if (!sc->sc_pr.pause && !audio_drain(sc) && hw->drain)
New audio subsystem
1995-02-21 04:35:58 +03:00
(void)hw->drain(sc->hw_hdl);
minor KNF
2005-01-15 19:23:03 +03:00
hw->halt_output(sc->hw_hdl);
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
sc->sc_pbus = false;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (hw->close != NULL)
hw->close(sc->hw_hdl);
When audioclose is called it's the last close of the device, so there is no need to fiddle with resetting individual bits. Just clear them.
2005-05-04 05:57:23 +04:00
sc->sc_open = 0;
sc->sc_mode = 0;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
sc->sc_full_duplex = 0;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
sc->sc_async_audio = 0;
New audio subsystem
1995-02-21 04:35:58 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return 0;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
int
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_read(struct audio_softc *sc, struct uio *uio, int ioflag)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
struct audio_ringbuffer *cb;
const uint8_t *outp;
uint8_t *inp;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
int error, used, cc, n;
KASSERT(mutex_owned(sc->sc_lock));
New audio subsystem
1995-02-21 04:35:58 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
cb = &sc->sc_rr;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
if (cb->mmapped)
return EINVAL;
New audio subsystem
1995-02-21 04:35:58 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
DPRINTFN(1,("audio_read: cc=%zu mode=%d\n",
uio->uio_resid, sc->sc_mode));
New audio subsystem
1995-02-21 04:35:58 +03:00
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#ifdef AUDIO_PM_IDLE
Don't count poll(2) and select(2) as device activity. Don't activate the device on write(2) if it is suspended and was not idle.
2007-12-13 17:40:36 +03:00
if (device_is_active(&sc->dev) || sc->sc_idle)
device_active(&sc->dev, DVA_SYSTEM);
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#endif
Don't count poll(2) and select(2) as device activity. Don't activate the device on write(2) if it is suspended and was not idle.
2007-12-13 17:40:36 +03:00
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
error = 0;
New audio subsystem
1995-02-21 04:35:58 +03:00
/*
* If hardware is half-duplex and currently playing, return
* silence blocks based on the number of blocks we have output.
*/
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (!sc->sc_full_duplex && (sc->sc_mode & AUMODE_PLAY)) {
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
while (uio->uio_resid > 0 && !error) {
for(;;) {
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
/*
* No need to lock, as any wakeup will be
* held for us while holding sc_lock.
*/
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
cc = sc->sc_pr.stamp - sc->sc_wstamp;
if (cc > 0)
break;
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
DPRINTF(("audio_read: stamp=%lu, wstamp=%lu\n",
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
sc->sc_pr.stamp, sc->sc_wstamp));
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if (ioflag & IO_NDELAY)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return EWOULDBLOCK;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
error = audio_waitio(sc, &sc->sc_rchan);
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
if (sc->sc_dying)
error = EIO;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if (error)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
if (uio->uio_resid < cc)
cc = uio->uio_resid;
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
DPRINTFN(1,("audio_read: reading in write mode, "
"cc=%d\n", cc));
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
error = audio_silence_copyout(sc, cc, uio);
sc->sc_wstamp += cc;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
while (uio->uio_resid > 0 && !error) {
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
while ((used = audio_stream_get_used(sc->sc_rustream)) <= 0) {
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if (!sc->sc_rbus && !sc->sc_rr.pause)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
error = audiostartr(sc);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
if (error)
return error;
if (ioflag & IO_NDELAY)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return EWOULDBLOCK;
DPRINTFN(2, ("audio_read: sleep used=%d\n", used));
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
error = audio_waitio(sc, &sc->sc_rchan);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (sc->sc_dying)
error = EIO;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if (error)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return error;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Sampling rate and mono-stereo conversion described in http://mail-index.netbsd.org/tech-kern/2002/03/04/0005.html auconv.c: Add conversion functions audio.c: Sample alignment, calling conversion functions, etc. audio_if.h: Add four hw_* members to "struct audio_params" audiovar.h: Add conversion buffers, etc. auich and uaudio: Add conversion request code to *_set_params().
2002-03-07 17:37:02 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
outp = sc->sc_rustream->outp;
inp = sc->sc_rustream->inp;
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
cb->copying = true;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
/*
* cc is the amount of data in the sc_rustream excluding
Don't enter an infinite loop in audio_read() when the sc->sc_rustream buffer is full. Fixes kern/34659.
2008-01-23 17:14:55 +03:00
* wrapped data. Note the tricky case of inp == outp, which
* must mean the buffer is full, not empty, because used > 0.
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
*/
Don't enter an infinite loop in audio_read() when the sc->sc_rustream buffer is full. Fixes kern/34659.
2008-01-23 17:14:55 +03:00
cc = outp < inp ? inp - outp :sc->sc_rustream->end - outp;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
DPRINTFN(1,("audio_read: outp=%p, cc=%d\n", outp, cc));
New audio subsystem
1995-02-21 04:35:58 +03:00
Support recording conversions that change the sample size. PR 10029, from Gregory McGarry.
2000-05-02 04:00:00 +04:00
n = uio->uio_resid;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
mutex_exit(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
error = uiomove(__UNCONST(outp), cc, uio);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
mutex_enter(sc->sc_intr_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
n -= uio->uio_resid; /* number of bytes actually moved */
Support recording conversions that change the sample size. PR 10029, from Gregory McGarry.
2000-05-02 04:00:00 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
sc->sc_rustream->outp = audio_stream_add_outp
(sc->sc_rustream, outp, n);
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
cb->copying = false;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
void
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_clear(struct audio_softc *sc)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
ensure that the intr lock is held for all audio_clear() callers.
2012-04-30 06:16:46 +04:00
KASSERT(mutex_owned(sc->sc_intr_lock));
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
Restructure a little to make sure AUDIO_FLUSH restarts the looping DMA.
1997-08-07 03:08:26 +04:00
if (sc->sc_rbus) {
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
cv_broadcast(&sc->sc_rchan);
New audio subsystem
1995-02-21 04:35:58 +03:00
sc->hw_if->halt_input(sc->hw_hdl);
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
sc->sc_rbus = false;
- Fix the conditions for POLLOUT to match those elsewhere - Refactor pause attribute handling so that it it's preserved appropriately and reset in audio_clear(). Now audio/jack-devel works with the native driver.
2008-06-03 10:21:17 +04:00
sc->sc_rr.pause = false;
Restructure a little to make sure AUDIO_FLUSH restarts the looping DMA.
1997-08-07 03:08:26 +04:00
}
if (sc->sc_pbus) {
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
cv_broadcast(&sc->sc_wchan);
Restructure a little to make sure AUDIO_FLUSH restarts the looping DMA.
1997-08-07 03:08:26 +04:00
sc->hw_if->halt_output(sc->hw_hdl);
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
sc->sc_pbus = false;
- Fix the conditions for POLLOUT to match those elsewhere - Refactor pause attribute handling so that it it's preserved appropriately and reset in audio_clear(). Now audio/jack-devel works with the native driver.
2008-06-03 10:21:17 +04:00
sc->sc_pr.pause = false;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
deal with PR#46232: ensure locking around audio_clear() is handled consistently and avoid locking against self.
2012-03-26 22:26:10 +04:00
}
void
audio_clear_intr_unlocked(struct audio_softc *sc)
{
mutex_enter(sc->sc_intr_lock);
audio_clear(sc);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
void
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_calc_blksize(struct audio_softc *sc, int mode)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
const audio_params_t *parm;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
struct audio_ringbuffer *rb;
Make setting the audio block size sticky. Change the calculated block size to correspond to 50 ms of sound (was 20 ms, it seems needlessly small).
1997-08-08 04:03:26 +04:00
if (sc->sc_blkset)
return;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
if (mode == AUMODE_PLAY) {
rb = &sc->sc_pr;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
parm = &rb->s.param;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
} else {
rb = &sc->sc_rr;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
parm = &rb->s.param;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
}
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
rb->blksize = parm->sample_rate * audio_blk_ms / 1000 *
parm->channels * parm->precision / NBBY;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
DPRINTF(("audio_calc_blksize: %s blksize=%d\n",
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
mode == AUMODE_PLAY ? "play" : "record", rb->blksize));
New audio subsystem
1995-02-21 04:35:58 +03:00
}
void
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
audio_fill_silence(struct audio_params *params, uint8_t *p, int n)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
minor KNF
2005-01-15 19:23:03 +03:00
uint8_t auzero0, auzero1;
int nfill;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
minor KNF
2005-01-15 19:23:03 +03:00
auzero1 = 0; /* initialize to please gcc */
nfill = 1;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
switch (params->encoding) {
Don't share the silence block between devices. Make silence filling work for more encodings, and make it device-independent. From Lennart Augustsson, in PR kern/3305.
1997-03-13 05:19:32 +03:00
case AUDIO_ENCODING_ULAW:
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
auzero0 = 0x7f;
Don't share the silence block between devices. Make silence filling work for more encodings, and make it device-independent. From Lennart Augustsson, in PR kern/3305.
1997-03-13 05:19:32 +03:00
break;
case AUDIO_ENCODING_ALAW:
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
auzero0 = 0x55;
Don't share the silence block between devices. Make silence filling work for more encodings, and make it device-independent. From Lennart Augustsson, in PR kern/3305.
1997-03-13 05:19:32 +03:00
break;
TMS320AV110 (mpeg layer 2 decoder chip) and ZBUS Melody board, which uses it. No optionboards yet for Melody, only audio driver. MPEG data types for audio.c
1997-10-17 03:57:56 +04:00
case AUDIO_ENCODING_MPEG_L1_STREAM:
case AUDIO_ENCODING_MPEG_L1_PACKETS:
case AUDIO_ENCODING_MPEG_L1_SYSTEM:
case AUDIO_ENCODING_MPEG_L2_STREAM:
case AUDIO_ENCODING_MPEG_L2_PACKETS:
case AUDIO_ENCODING_MPEG_L2_SYSTEM:
Add AUDIO_ENCODING_AC3 encoding type.
2011-09-06 05:16:43 +04:00
case AUDIO_ENCODING_AC3:
Don't share the silence block between devices. Make silence filling work for more encodings, and make it device-independent. From Lennart Augustsson, in PR kern/3305.
1997-03-13 05:19:32 +03:00
case AUDIO_ENCODING_ADPCM: /* is this right XXX */
In the name of backwards compatibility AUDIO_ENCODING_LINEAR has been renamed AUDIO_ENCODING_SLINEAR and AUDIO_ENCODING_LINEAR reverts to the NetBSD 1.2 sematics. A kernel with COMPAT_12 defined will accept AUDIO_ENCODING_LINEAR and treat it as before, without COMPAT_12 it will be rejected.
1997-07-15 11:46:04 +04:00
case AUDIO_ENCODING_SLINEAR_LE:
case AUDIO_ENCODING_SLINEAR_BE:
Allow 24 and 32bit linear audio. Actually, the silence filler can do any multiple of 8 bits now, but I didn't allow the parameter check to accept more than 32 bit to avoid confusion of drivers that fail to check the parameters themselves thoroughly. This should be changed later.
2001-09-03 22:51:43 +04:00
auzero0 = 0;/* fortunately this works for any number of bits */
Convert to new orthogonal audio encoding scheme and implement some of the new encodings. The change to ioctl AUDIO_GETENC is NOT backwards compatible.
1997-05-07 22:51:31 +04:00
break;
case AUDIO_ENCODING_ULINEAR_LE:
case AUDIO_ENCODING_ULINEAR_BE:
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (params->precision > 8) {
nfill = (params->precision + NBBY - 1)/ NBBY;
Allow 24 and 32bit linear audio. Actually, the silence filler can do any multiple of 8 bits now, but I didn't allow the parameter check to accept more than 32 bit to avoid confusion of drivers that fail to check the parameters themselves thoroughly. This should be changed later.
2001-09-03 22:51:43 +04:00
auzero0 = 0x80;
auzero1 = 0;
Convert to new orthogonal audio encoding scheme and implement some of the new encodings. The change to ioctl AUDIO_GETENC is NOT backwards compatible.
1997-05-07 22:51:31 +04:00
} else
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
auzero0 = 0x80;
Convert to new orthogonal audio encoding scheme and implement some of the new encodings. The change to ioctl AUDIO_GETENC is NOT backwards compatible.
1997-05-07 22:51:31 +04:00
break;
Don't share the silence block between devices. Make silence filling work for more encodings, and make it device-independent. From Lennart Augustsson, in PR kern/3305.
1997-03-13 05:19:32 +03:00
default:
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
DPRINTF(("audio: bad encoding %d\n", params->encoding));
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
auzero0 = 0;
Don't share the silence block between devices. Make silence filling work for more encodings, and make it device-independent. From Lennart Augustsson, in PR kern/3305.
1997-03-13 05:19:32 +03:00
break;
}
Convert to new orthogonal audio encoding scheme and implement some of the new encodings. The change to ioctl AUDIO_GETENC is NOT backwards compatible.
1997-05-07 22:51:31 +04:00
if (nfill == 1) {
while (--n >= 0)
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
*p++ = auzero0; /* XXX memset */
Allow 24 and 32bit linear audio. Actually, the silence filler can do any multiple of 8 bits now, but I didn't allow the parameter check to accept more than 32 bit to avoid confusion of drivers that fail to check the parameters themselves thoroughly. This should be changed later.
2001-09-03 22:51:43 +04:00
} else /* nfill must no longer be 2 */ {
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (params->encoding == AUDIO_ENCODING_ULINEAR_LE) {
Allow 24 and 32bit linear audio. Actually, the silence filler can do any multiple of 8 bits now, but I didn't allow the parameter check to accept more than 32 bit to avoid confusion of drivers that fail to check the parameters themselves thoroughly. This should be changed later.
2001-09-03 22:51:43 +04:00
int k = nfill;
while (--k > 0)
*p++ = auzero1;
n -= nfill - 1;
}
while (n >= nfill) {
fix unitialized variable. from simonb. OI
2002-09-23 07:44:56 +04:00
int k = nfill;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
*p++ = auzero0;
Allow 24 and 32bit linear audio. Actually, the silence filler can do any multiple of 8 bits now, but I didn't allow the parameter check to accept more than 32 bit to avoid confusion of drivers that fail to check the parameters themselves thoroughly. This should be changed later.
2001-09-03 22:51:43 +04:00
while (--k > 0)
*p++ = auzero1;
n -= nfill;
Convert to new orthogonal audio encoding scheme and implement some of the new encodings. The change to ioctl AUDIO_GETENC is NOT backwards compatible.
1997-05-07 22:51:31 +04:00
}
Allow 24 and 32bit linear audio. Actually, the silence filler can do any multiple of 8 bits now, but I didn't allow the parameter check to accept more than 32 bit to avoid confusion of drivers that fail to check the parameters themselves thoroughly. This should be changed later.
2001-09-03 22:51:43 +04:00
if (n-- > 0) /* XXX must be 1 - DIAGNOSTIC check? */
*p++ = auzero0;
Convert to new orthogonal audio encoding scheme and implement some of the new encodings. The change to ioctl AUDIO_GETENC is NOT backwards compatible.
1997-05-07 22:51:31 +04:00
}
New audio subsystem
1995-02-21 04:35:58 +03:00
}
prototypes for ccd audio and ncr5380
1996-03-07 18:00:07 +03:00
int
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_silence_copyout(struct audio_softc *sc, int n, struct uio *uio)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
minor KNF
2005-01-15 19:23:03 +03:00
uint8_t zerobuf[128];
Clean up audio_silence_copyout().
1997-05-11 04:41:12 +04:00
int error;
Don't share the silence block between devices. Make silence filling work for more encodings, and make it device-independent. From Lennart Augustsson, in PR kern/3305.
1997-03-13 05:19:32 +03:00
int k;
New audio subsystem
1995-02-21 04:35:58 +03:00
Clean up audio_silence_copyout().
1997-05-11 04:41:12 +04:00
audio_fill_silence(&sc->sc_rparams, zerobuf, sizeof zerobuf);
New audio subsystem
1995-02-21 04:35:58 +03:00
Clean up audio_silence_copyout().
1997-05-11 04:41:12 +04:00
error = 0;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
while (n > 0 && uio->uio_resid > 0 && !error) {
Clean up audio_silence_copyout().
1997-05-11 04:41:12 +04:00
k = min(n, min(uio->uio_resid, sizeof zerobuf));
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
Clean up audio_silence_copyout().
1997-05-11 04:41:12 +04:00
error = uiomove(zerobuf, k, uio);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
Clean up audio_silence_copyout().
1997-05-11 04:41:12 +04:00
n -= k;
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return error;
}
static int
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
uio_fetcher_fetch_to(struct audio_softc *sc, stream_fetcher_t *self,
audio_stream_t *p, int max_used)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
{
uio_fetcher_t *this;
int size;
int stream_space;
int error;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
KASSERT(!cpu_intr_p());
KASSERT(!cpu_softintr_p());
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
this = (uio_fetcher_t *)self;
Fixes a bug that audio_offset_t::samples for AUDIO_GETxOFFS has incorrect value. It also affects SNDCTL_DSP_GETOPTR of ossaudio. PR: kern/32405 by Stephane Witzman
2006-05-10 05:56:21 +04:00
this->last_used = audio_stream_get_used(p);
if (this->last_used >= this->usedhigh)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return 0;
/*
* uio_fetcher ignores max_used and move the data as
* much as possible in order to return the correct value
* for audio_prinfo::seek and kfilters.
*/
stream_space = audio_stream_get_space(p);
size = min(this->uio->uio_resid, stream_space);
/* the first fragment of the space */
stream_space = p->end - p->inp;
if (stream_space >= size) {
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
error = uiomove(p->inp, size, this->uio);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (error)
return error;
p->inp = audio_stream_add_inp(p, p->inp, size);
} else {
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
error = uiomove(p->inp, stream_space, this->uio);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (error)
return error;
p->inp = audio_stream_add_inp(p, p->inp, stream_space);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
error = uiomove(p->start, size - stream_space, this->uio);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (error)
return error;
p->inp = audio_stream_add_inp(p, p->inp, size - stream_space);
}
this->last_used = audio_stream_get_used(p);
return 0;
}
static int
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
null_fetcher_fetch_to(struct audio_softc *sc, stream_fetcher_t *self,
__unused removal on arguments; approved by core.
2006-11-16 04:32:37 +03:00
audio_stream_t *p, int max_used)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
{
minor KNF
2005-01-15 19:23:03 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return 0;
}
static void
uio_fetcher_ctor(uio_fetcher_t *this, struct uio *u, int h)
{
minor KNF
2005-01-15 19:23:03 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
this->base.fetch_to = uio_fetcher_fetch_to;
this->uio = u;
this->usedhigh = h;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
int
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_write(struct audio_softc *sc, struct uio *uio, int ioflag)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
uio_fetcher_t ufetcher;
audio_stream_t stream;
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_ringbuffer *cb;
stream_fetcher_t *fetcher;
stream_filter_t *filter;
uint8_t *inp, *einp;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
int saveerror, error, n, cc, used;
KASSERT(mutex_owned(sc->sc_lock));
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
DPRINTFN(2,("audio_write: sc=%p count=%zu used=%d(hi=%d)\n",
sc, uio->uio_resid, audio_stream_get_used(sc->sc_pustream),
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
sc->sc_pr.usedhigh));
minor KNF
2005-01-15 19:23:03 +03:00
cb = &sc->sc_pr;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
if (cb->mmapped)
return EINVAL;
if (uio->uio_resid == 0) {
sc->sc_eof++;
return 0;
}
New audio subsystem
1995-02-21 04:35:58 +03:00
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#ifdef AUDIO_PM_IDLE
Don't count poll(2) and select(2) as device activity. Don't activate the device on write(2) if it is suspended and was not idle.
2007-12-13 17:40:36 +03:00
if (device_is_active(&sc->dev) || sc->sc_idle)
device_active(&sc->dev, DVA_SYSTEM);
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#endif
Don't count poll(2) and select(2) as device activity. Don't activate the device on write(2) if it is suspended and was not idle.
2007-12-13 17:40:36 +03:00
New audio subsystem
1995-02-21 04:35:58 +03:00
/*
* If half-duplex and currently recording, throw away data.
*/
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
if (!sc->sc_full_duplex &&
Make mode a flag set. Full-dpulex fixes from John Kohl.
1995-07-07 05:52:30 +04:00
(sc->sc_mode & AUMODE_RECORD)) {
New audio subsystem
1995-02-21 04:35:58 +03:00
uio->uio_offset += uio->uio_resid;
uio->uio_resid = 0;
Make mode a flag set. Full-dpulex fixes from John Kohl.
1995-07-07 05:52:30 +04:00
DPRINTF(("audio_write: half-dpx read busy\n"));
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return 0;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
if (!(sc->sc_mode & AUMODE_PLAY_ALL) && sc->sc_playdrop > 0) {
n = min(sc->sc_playdrop, uio->uio_resid);
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
DPRINTF(("audio_write: playdrop %d\n", n));
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
uio->uio_offset += n;
uio->uio_resid -= n;
sc->sc_playdrop -= n;
if (uio->uio_resid == 0)
return 0;
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
/**
* setup filter pipeline
Sampling rate and mono-stereo conversion described in http://mail-index.netbsd.org/tech-kern/2002/03/04/0005.html auconv.c: Add conversion functions audio.c: Sample alignment, calling conversion functions, etc. audio_if.h: Add four hw_* members to "struct audio_params" audiovar.h: Add conversion buffers, etc. auich and uaudio: Add conversion request code to *_set_params().
2002-03-07 17:37:02 +03:00
*/
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
uio_fetcher_ctor(&ufetcher, uio, cb->usedhigh);
if (sc->sc_npfilters > 0) {
fetcher = &sc->sc_pfilters[sc->sc_npfilters - 1]->base;
} else {
fetcher = &ufetcher.base;
Sampling rate and mono-stereo conversion described in http://mail-index.netbsd.org/tech-kern/2002/03/04/0005.html auconv.c: Add conversion functions audio.c: Sample alignment, calling conversion functions, etc. audio_if.h: Add four hw_* members to "struct audio_params" audiovar.h: Add conversion buffers, etc. auich and uaudio: Add conversion request code to *_set_params().
2002-03-07 17:37:02 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
error = 0;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
while (uio->uio_resid > 0 && !error) {
/* wait if the first buffer is occupied */
minor KNF
2005-01-15 19:23:03 +03:00
while ((used = audio_stream_get_used(sc->sc_pustream))
>= cb->usedhigh) {
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
DPRINTFN(2, ("audio_write: sleep used=%d lowat=%d "
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
"hiwat=%d\n", used,
cb->usedlow, cb->usedhigh));
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
if (ioflag & IO_NDELAY)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return EWOULDBLOCK;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
error = audio_waitio(sc, &sc->sc_wchan);
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
if (sc->sc_dying)
error = EIO;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if (error)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return error;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
New audio subsystem
1995-02-21 04:35:58 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
inp = cb->s.inp;
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
cb->copying = true;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
stream = cb->s;
used = stream.used;
Don't use simple_locks; we can't hold them while calling malloc and friends.
2006-04-19 18:10:58 +04:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
/* Write to the sc_pustream as much as possible. */
mutex_exit(sc->sc_intr_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (sc->sc_npfilters > 0) {
filter = sc->sc_pfilters[0];
filter->set_fetcher(filter, &ufetcher.base);
Protect audio_write's filter graph from being modified in the middle of a write. Fixes the 'audioctl of death' problem (PR#32563).
2006-04-18 23:15:27 +04:00
fetcher = &sc->sc_pfilters[sc->sc_npfilters - 1]->base;
If one or more playback filters are applied, minimize loading size of the hardware buffer to improve accuracy of AUDIO_GETIOFFS.
2006-05-16 17:46:19 +04:00
cc = cb->blksize * 2;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
error = fetcher->fetch_to(sc, fetcher, &stream, cc);
If one or more playback filters are applied, minimize loading size of the hardware buffer to improve accuracy of AUDIO_GETIOFFS.
2006-05-16 17:46:19 +04:00
if (error != 0) {
fetcher = &ufetcher.base;
cc = sc->sc_pustream->end - sc->sc_pustream->start;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
error = fetcher->fetch_to(sc, fetcher,
sc->sc_pustream, cc);
If one or more playback filters are applied, minimize loading size of the hardware buffer to improve accuracy of AUDIO_GETIOFFS.
2006-05-16 17:46:19 +04:00
}
Protect audio_write's filter graph from being modified in the middle of a write. Fixes the 'audioctl of death' problem (PR#32563).
2006-04-18 23:15:27 +04:00
} else {
fetcher = &ufetcher.base;
If one or more playback filters are applied, minimize loading size of the hardware buffer to improve accuracy of AUDIO_GETIOFFS.
2006-05-16 17:46:19 +04:00
cc = stream.end - stream.start;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
error = fetcher->fetch_to(sc, fetcher, &stream, cc);
Protect audio_write's filter graph from being modified in the middle of a write. Fixes the 'audioctl of death' problem (PR#32563).
2006-04-18 23:15:27 +04:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (sc->sc_npfilters > 0) {
Fixes a bug that audio_offset_t::samples for AUDIO_GETxOFFS has incorrect value. It also affects SNDCTL_DSP_GETOPTR of ossaudio. PR: kern/32405 by Stephane Witzman
2006-05-10 05:56:21 +04:00
cb->fstamp += ufetcher.last_used
- audio_stream_get_used(sc->sc_pustream);
Move sampling rate conversion functions to aurateconv.c. Introduce "aurateconv" attribute for audio devices. Add aurateconv to uaudio and auich. (due to kern/15845 and kern/15848)
2002-03-09 23:30:42 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
cb->s.used += stream.used - used;
cb->s.inp = stream.inp;
einp = cb->s.inp;
New audio subsystem
1995-02-21 04:35:58 +03:00
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
/*
* This is a very suboptimal way of keeping track of
* silence in the buffer, but it is simple.
*/
sc->sc_sil_count = 0;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
/*
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
* If the interrupt routine wants the last block filled AND
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
* the copy did not fill the last block completely it needs to
* be padded.
*/
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (cb->needfill && inp < einp &&
(inp - cb->s.start) / cb->blksize ==
(einp - cb->s.start) / cb->blksize) {
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
/* Figure out how many bytes to a block boundary. */
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
cc = cb->blksize - (einp - cb->s.start) % cb->blksize;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
DPRINTF(("audio_write: partial fill %d\n", cc));
} else
cc = 0;
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
cb->needfill = false;
cb->copying = false;
Don't clobber the error returned by uiomove() when starting the audio output (in case we got a partial block). Fixes PR #7722, from Antti Kantee <pooka@iki.fi>.
1999-06-07 23:24:38 +04:00
if (!sc->sc_pbus && !cb->pause) {
saveerror = error;
error = audiostartp(sc);
if (saveerror != 0) {
Spell 'occurred' with two 'r's.
2001-09-16 20:34:23 +04:00
/* Report the first error that occurred. */
Don't clobber the error returned by uiomove() when starting the audio output (in case we got a partial block). Fixes PR #7722, from Antti Kantee <pooka@iki.fi>.
1999-06-07 23:24:38 +04:00
error = saveerror;
}
}
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
if (cc != 0) {
Add a few more debug messages and clean up debug printing a little.
1998-04-28 13:07:12 +04:00
DPRINTFN(1, ("audio_write: fill %d\n", cc));
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
audio_fill_silence(&cb->s.param, einp, cc);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
}
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
Protect audio_write's filter graph from being modified in the middle of a write. Fixes the 'audioctl of death' problem (PR#32563).
2006-04-18 23:15:27 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
int
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.
2007-03-04 08:59:00 +03:00
audio_ioctl(struct audio_softc *sc, u_long cmd, void *addr, int flag,
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
struct lwp *l)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
minor KNF
2005-01-15 19:23:03 +03:00
const struct audio_hw_if *hw;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
struct audio_offset *ao;
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
u_long stamp;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
int error, offs, fd;
Replace the Mach-derived boolean_t type with the C99 bool type. A future commit will replace use of TRUE and FALSE with true and false.
2007-02-22 01:59:35 +03:00
bool rbus, pbus;
New audio subsystem
1995-02-21 04:35:58 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
Change the cmd argument in {audio,mixer}_ioctl() to u_long.
2000-03-01 03:44:35 +03:00
DPRINTF(("audio_ioctl(%lu,'%c',%lu)\n",
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
IOCPARM_LEN(cmd), (char)IOCGROUP(cmd), cmd&0xff));
minor KNF
2005-01-15 19:23:03 +03:00
hw = sc->hw_if;
error = 0;
New audio subsystem
1995-02-21 04:35:58 +03:00
switch (cmd) {
Implement FIONBIO in audio_ioctl.
1997-08-01 21:04:00 +04:00
case FIONBIO:
/* All handled in the upper FS layer. */
break;
Implement delivery of SIGIO on changes to /dev/mixer.
1997-08-11 05:38:12 +04:00
Add FIONREAD ioctl().
2003-05-26 16:43:35 +04:00
case FIONREAD:
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
*(int *)addr = audio_stream_get_used(sc->sc_rustream);
Add FIONREAD ioctl().
2003-05-26 16:43:35 +04:00
break;
Trivial (and probably not very good) implementation of SIGIO.
1996-02-20 13:00:31 +03:00
case FIOASYNC:
if (*(int *)addr) {
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if (sc->sc_async_audio != 0)
error = EBUSY;
else
sc->sc_async_audio = curproc->p_pid;
DPRINTF(("audio_ioctl: FIOASYNC pid %d\n",
curproc->p_pid));
Trivial (and probably not very good) implementation of SIGIO.
1996-02-20 13:00:31 +03:00
} else
Implement delivery of SIGIO on changes to /dev/mixer.
1997-08-11 05:38:12 +04:00
sc->sc_async_audio = 0;
Trivial (and probably not very good) implementation of SIGIO.
1996-02-20 13:00:31 +03:00
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
case AUDIO_FLUSH:
DPRINTF(("AUDIO_FLUSH\n"));
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
rbus = sc->sc_rbus;
pbus = sc->sc_pbus;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
ensure that the intr lock is held for all audio_clear() callers.
2012-04-30 06:16:46 +04:00
audio_clear(sc);
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
error = audio_initbufs(sc);
if (error) {
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
return error;
}
Fix three problems spotted by mycroft: - Closing the audio device should stop recording. - Manipulating the pause function could accidentally start playing or recording. - AUDIO_FLUSH could accidentally start playing or recording.
1997-12-03 04:01:19 +03:00
if ((sc->sc_mode & AUMODE_PLAY) && !sc->sc_pbus && pbus)
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
error = audiostartp(sc);
if (!error &&
Fix three problems spotted by mycroft: - Closing the audio device should stop recording. - Manipulating the pause function could accidentally start playing or recording. - AUDIO_FLUSH could accidentally start playing or recording.
1997-12-03 04:01:19 +03:00
(sc->sc_mode & AUMODE_RECORD) && !sc->sc_rbus && rbus)
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
error = audiostartr(sc);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
New audio subsystem
1995-02-21 04:35:58 +03:00
break;
/*
Fix PR#2060 (add AUDIO_PERROR ioctl to query # of playback drops), also fix semantics of AUDIO_WERROR (now returns a count). Also repair audio driver back to "real-time mode" where user must provide data quickly enough for real time audio or silence is played until user catches up. Add optional "play all" mode (additional bit in mode field) to play all samples provided, with silence when needed to avoid buffer underrun.
1996-02-17 05:28:56 +03:00
* Number of read (write) samples dropped. We don't know where or
New audio subsystem
1995-02-21 04:35:58 +03:00
* when they were dropped.
*/
case AUDIO_RERROR:
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
*(int *)addr = sc->sc_rr.drops;
Fix PR#2060 (add AUDIO_PERROR ioctl to query # of playback drops), also fix semantics of AUDIO_WERROR (now returns a count). Also repair audio driver back to "real-time mode" where user must provide data quickly enough for real time audio or silence is played until user catches up. Add optional "play all" mode (additional bit in mode field) to play all samples provided, with silence when needed to avoid buffer underrun.
1996-02-17 05:28:56 +03:00
break;
case AUDIO_PERROR:
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
*(int *)addr = sc->sc_pr.drops;
New audio subsystem
1995-02-21 04:35:58 +03:00
break;
/*
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
* Offsets into buffer.
New audio subsystem
1995-02-21 04:35:58 +03:00
*/
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
case AUDIO_GETIOFFS:
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
ao = (struct audio_offset *)addr;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
/* figure out where next DMA will start */
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
stamp = sc->sc_rustream == &sc->sc_rr.s
? sc->sc_rr.stamp : sc->sc_rr.fstamp;
offs = sc->sc_rustream->inp - sc->sc_rustream->start;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
ao->samples = stamp;
ao->deltablks =
(stamp / sc->sc_rr.blksize) -
(sc->sc_rr.stamp_last / sc->sc_rr.blksize);
sc->sc_rr.stamp_last = stamp;
ao->offset = offs;
New audio subsystem
1995-02-21 04:35:58 +03:00
break;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
case AUDIO_GETOOFFS:
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
ao = (struct audio_offset *)addr;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
/* figure out where next DMA will start */
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
stamp = sc->sc_pustream == &sc->sc_pr.s
? sc->sc_pr.stamp : sc->sc_pr.fstamp;
offs = sc->sc_pustream->outp - sc->sc_pustream->start
+ sc->sc_pr.blksize;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
ao->samples = stamp;
ao->deltablks =
(stamp / sc->sc_pr.blksize) -
(sc->sc_pr.stamp_last / sc->sc_pr.blksize);
sc->sc_pr.stamp_last = stamp;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (sc->sc_pustream->start + offs >= sc->sc_pustream->end)
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
offs = 0;
ao->offset = offs;
break;
/*
* How many bytes will elapse until mike hears the first
* sample of what we write next?
*/
case AUDIO_WSEEK:
return information about playing buffer, not recording bufer, for AUDIO_WSEEK. PR#35171
2007-03-09 16:20:12 +03:00
*(u_long *)addr = audio_stream_get_used(sc->sc_pustream);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
case AUDIO_SETINFO:
%x -> 0x%x
1997-08-06 11:39:59 +04:00
DPRINTF(("AUDIO_SETINFO mode=0x%x\n", sc->sc_mode));
Fix full-duplex mode (again). Back-out revision 1.170 and and 1.166. Revision 1.170 fixes some problems introduced in revision 1.166. It isn't clear what problem revision 1.166 was attempting to address. It seems bogus anyway, since we later check the modes in audio_check_params().
2003-04-25 07:02:11 +04:00
error = audiosetinfo(sc, (struct audio_info *)addr);
New audio subsystem
1995-02-21 04:35:58 +03:00
break;
case AUDIO_GETINFO:
DPRINTF(("AUDIO_GETINFO\n"));
Add a new ioctl AUDIO_GETBUFINFO. It works like AUDIO_GETINFO, but doesn't obtain the ports, gain and balance related parameters. Those generally require reading from the hardware and therefore are much more expensive to obtain. Modify OSS emulation to use the new ioctl where possible. This reduces CPU usage of mplayer during mp3 playback with my Thinkpad from 20% to < 1% and from 50% to 20% during Xvid playback. Review and comments from jmcneill@
2007-06-11 17:05:46 +04:00
error = audiogetinfo(sc, (struct audio_info *)addr, 0);
break;
case AUDIO_GETBUFINFO:
DPRINTF(("AUDIO_GETBUFINFO\n"));
error = audiogetinfo(sc, (struct audio_info *)addr, 1);
New audio subsystem
1995-02-21 04:35:58 +03:00
break;
case AUDIO_DRAIN:
DPRINTF(("AUDIO_DRAIN\n"));
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
New audio subsystem
1995-02-21 04:35:58 +03:00
error = audio_drain(sc);
if (!error && hw->drain)
error = hw->drain(sc->hw_hdl);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
New audio subsystem
1995-02-21 04:35:58 +03:00
break;
case AUDIO_GETDEV:
DPRINTF(("AUDIO_GETDEV\n"));
error = hw->getdev(sc->hw_hdl, (audio_device_t *)addr);
break;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
New audio subsystem
1995-02-21 04:35:58 +03:00
case AUDIO_GETENC:
DPRINTF(("AUDIO_GETENC\n"));
minor KNF
2005-01-15 19:23:03 +03:00
error = hw->query_encoding(sc->hw_hdl,
(struct audio_encoding *)addr);
New audio subsystem
1995-02-21 04:35:58 +03:00
break;
case AUDIO_GETFD:
DPRINTF(("AUDIO_GETFD\n"));
GETPROPS does not indicate whether the device is currently in full-duplex mode. Reinstate GETFD to return this information.
1997-10-11 15:16:28 +04:00
*(int *)addr = sc->sc_full_duplex;
New audio subsystem
1995-02-21 04:35:58 +03:00
break;
GETPROPS does not indicate whether the device is currently in full-duplex mode. Reinstate GETFD to return this information.
1997-10-11 15:16:28 +04:00
New audio subsystem
1995-02-21 04:35:58 +03:00
case AUDIO_SETFD:
DPRINTF(("AUDIO_SETFD\n"));
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
fd = *(int *)addr;
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
if (audio_get_props(sc) & AUDIO_PROP_FULLDUPLEX) {
audio: Simplify handling of AUDIO_SETFD and committing of encoding mode.
1997-07-28 03:51:48 +04:00
if (hw->setfd)
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
error = hw->setfd(sc->hw_hdl, fd);
audio: Simplify handling of AUDIO_SETFD and committing of encoding mode.
1997-07-28 03:51:48 +04:00
else
error = 0;
if (!error)
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
sc->sc_full_duplex = fd;
audio: Simplify handling of AUDIO_SETFD and committing of encoding mode.
1997-07-28 03:51:48 +04:00
} else {
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
if (fd)
audio: Simplify handling of AUDIO_SETFD and committing of encoding mode.
1997-07-28 03:51:48 +04:00
error = ENOTTY;
else
error = 0;
}
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
break;
case AUDIO_GETPROPS:
DPRINTF(("AUDIO_GETPROPS\n"));
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
*(int *)addr = audio_get_props(sc);
New audio subsystem
1995-02-21 04:35:58 +03:00
break;
default:
Add a new optional method, dev_ioctl, to the audio hardware driver interface. It is called when an unrecognized ioctl() is performed on a device, thus allowing ioctl()s that frob the hardware driver (like loading microcode).
2001-10-03 04:04:47 +04:00
if (hw->dev_ioctl) {
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
error = hw->dev_ioctl(sc->hw_hdl, cmd, addr, flag, l);
Add a new optional method, dev_ioctl, to the audio hardware driver interface. It is called when an unrecognized ioctl() is performed on a device, thus allowing ioctl()s that frob the hardware driver (like loading microcode).
2001-10-03 04:04:47 +04:00
} else {
DPRINTF(("audio_ioctl: unknown ioctl\n"));
error = EINVAL;
}
New audio subsystem
1995-02-21 04:35:58 +03:00
break;
}
Change the cmd argument in {audio,mixer}_ioctl() to u_long.
2000-03-01 03:44:35 +03:00
DPRINTF(("audio_ioctl(%lu,'%c',%lu) result %d\n",
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
IOCPARM_LEN(cmd), (char)IOCGROUP(cmd), cmd&0xff, error));
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
int
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
audio_poll(struct audio_softc *sc, int events, struct lwp *l)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
minor KNF
2005-01-15 19:23:03 +03:00
int revents;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
int used;
New audio subsystem
1995-02-21 04:35:58 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
DPRINTF(("audio_poll: events=0x%x mode=%d\n", events, sc->sc_mode));
New audio subsystem
1995-02-21 04:35:58 +03:00
minor KNF
2005-01-15 19:23:03 +03:00
revents = 0;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
Fix full-duplex mode (again). Back-out revision 1.170 and and 1.166. Revision 1.170 fixes some problems introduced in revision 1.166. It isn't clear what problem revision 1.166 was attempting to address. It seems bogus anyway, since we later check the modes in audio_check_params().
2003-04-25 07:02:11 +04:00
if (events & (POLLIN | POLLRDNORM)) {
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
used = audio_stream_get_used(sc->sc_rustream);
The logic in audio_poll() was incorrect in full duplex mode; among other things, it behaved as if full-duplex audio devices were always ready for writing. Also commented the code in case. This fixes PRs kern/11179 and kern/13829.
2002-06-02 06:00:30 +04:00
/*
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
* If half duplex and playing, audio_read() will generate
The logic in audio_poll() was incorrect in full duplex mode; among other things, it behaved as if full-duplex audio devices were always ready for writing. Also commented the code in case. This fixes PRs kern/11179 and kern/13829.
2002-06-02 06:00:30 +04:00
* silence at the play rate; poll for silence being
* available. Otherwise, poll for recorded sound.
*/
if ((!sc->sc_full_duplex && (sc->sc_mode & AUMODE_PLAY)) ?
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
sc->sc_pr.stamp > sc->sc_wstamp :
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
used > sc->sc_rr.usedlow)
Implement poll(2).
1996-09-07 16:40:22 +04:00
revents |= events & (POLLIN | POLLRDNORM);
Fix full-duplex mode (again). Back-out revision 1.170 and and 1.166. Revision 1.170 fixes some problems introduced in revision 1.166. It isn't clear what problem revision 1.166 was attempting to address. It seems bogus anyway, since we later check the modes in audio_check_params().
2003-04-25 07:02:11 +04:00
}
New audio subsystem
1995-02-21 04:35:58 +03:00
Fix full-duplex mode (again). Back-out revision 1.170 and and 1.166. Revision 1.170 fixes some problems introduced in revision 1.166. It isn't clear what problem revision 1.166 was attempting to address. It seems bogus anyway, since we later check the modes in audio_check_params().
2003-04-25 07:02:11 +04:00
if (events & (POLLOUT | POLLWRNORM)) {
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
used = audio_stream_get_used(sc->sc_pustream);
The logic in audio_poll() was incorrect in full duplex mode; among other things, it behaved as if full-duplex audio devices were always ready for writing. Also commented the code in case. This fixes PRs kern/11179 and kern/13829.
2002-06-02 06:00:30 +04:00
/*
* If half duplex and recording, audio_write() will throw
* away play data, which means we are always ready to write.
* Otherwise, poll for play buffer being below its low water
* mark.
*/
if ((!sc->sc_full_duplex && (sc->sc_mode & AUMODE_RECORD)) ||
- Fix the conditions for POLLOUT to match those elsewhere - Refactor pause attribute handling so that it it's preserved appropriately and reset in audio_clear(). Now audio/jack-devel works with the native driver.
2008-06-03 10:21:17 +04:00
(!(sc->sc_mode & AUMODE_PLAY_ALL) && sc->sc_playdrop > 0) ||
(used <= sc->sc_pr.usedlow))
Implement poll(2).
1996-09-07 16:40:22 +04:00
revents |= events & (POLLOUT | POLLWRNORM);
Fix full-duplex mode (again). Back-out revision 1.170 and and 1.166. Revision 1.170 fixes some problems introduced in revision 1.166. It isn't clear what problem revision 1.166 was attempting to address. It seems bogus anyway, since we later check the modes in audio_check_params().
2003-04-25 07:02:11 +04:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
Implement poll(2).
1996-09-07 16:40:22 +04:00
if (revents == 0) {
if (events & (POLLIN | POLLRDNORM))
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
selrecord(l, &sc->sc_rsel);
Implement poll(2).
1996-09-07 16:40:22 +04:00
if (events & (POLLOUT | POLLWRNORM))
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
selrecord(l, &sc->sc_wsel);
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Implement poll(2).
1996-09-07 16:40:22 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return revents;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
static void
filt_audiordetach(struct knote *kn)
{
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_softc *sc;
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
minor KNF
2005-01-15 19:23:03 +03:00
sc = kn->kn_hook;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
si_ -> sel_
2002-11-26 21:49:40 +03:00
SLIST_REMOVE(&sc->sc_rsel.sel_klist, kn, knote, kn_selnext);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
}
static int
__unused removal on arguments; approved by core.
2006-11-16 04:32:37 +03:00
filt_audioread(struct knote *kn, long hint)
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
{
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_softc *sc;
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
minor KNF
2005-01-15 19:23:03 +03:00
sc = kn->kn_hook;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (!sc->sc_full_duplex && (sc->sc_mode & AUMODE_PLAY))
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
kn->kn_data = sc->sc_pr.stamp - sc->sc_wstamp;
else
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
kn->kn_data = audio_stream_get_used(sc->sc_rustream)
- sc->sc_rr.usedlow;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return kn->kn_data > 0;
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
}
static const struct filterops audioread_filtops =
{ 1, NULL, filt_audiordetach, filt_audioread };
static void
filt_audiowdetach(struct knote *kn)
{
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_softc *sc;
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
minor KNF
2005-01-15 19:23:03 +03:00
sc = kn->kn_hook;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
si_ -> sel_
2002-11-26 21:49:40 +03:00
SLIST_REMOVE(&sc->sc_wsel.sel_klist, kn, knote, kn_selnext);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
}
static int
__unused removal on arguments; approved by core.
2006-11-16 04:32:37 +03:00
filt_audiowrite(struct knote *kn, long hint)
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
{
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_softc *sc;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
audio_stream_t *stream;
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
minor KNF
2005-01-15 19:23:03 +03:00
sc = kn->kn_hook;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
stream = sc->sc_pustream;
kn->kn_data = (stream->end - stream->start)
- audio_stream_get_used(stream);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return kn->kn_data > 0;
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
}
static const struct filterops audiowrite_filtops =
{ 1, NULL, filt_audiowdetach, filt_audiowrite };
int
audio_kqfilter(struct audio_softc *sc, struct knote *kn)
{
struct klist *klist;
switch (kn->kn_filter) {
case EVFILT_READ:
si_ -> sel_
2002-11-26 21:49:40 +03:00
klist = &sc->sc_rsel.sel_klist;
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
kn->kn_fop = &audioread_filtops;
break;
case EVFILT_WRITE:
si_ -> sel_
2002-11-26 21:49:40 +03:00
klist = &sc->sc_wsel.sel_klist;
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
kn->kn_fop = &audiowrite_filtops;
break;
default:
Do not "return 1" from kqfilter for errors. That value is passed directly to the userland caller and results in a mysterious EPERM. Instead, return EINVAL or something else sensible depending on the case.
2007-12-05 20:19:46 +03:00
return EINVAL;
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
}
kn->kn_hook = sc;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
SLIST_INSERT_HEAD(klist, kn, kn_selnext);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return 0;
merge kqueue branch into -current kqueue provides a stateful and efficient event notification framework currently supported events include socket, file, directory, fifo, pipe, tty and device changes, and monitoring of processes and signals kqueue is supported by all writable filesystems in NetBSD tree (with exception of Coda) and all device drivers supporting poll(2) based on work done by Jonathan Lemon for FreeBSD initial NetBSD port done by Luke Mewburn and Jason Thorpe
2002-10-23 13:10:23 +04:00
}
Change the kernel mmap interface so that the offset to map is an "off_t" and the return value is a "paddr_t" to allow mappings at offsets past 2^31 bytes. Somewhat inspired by FreeBSD, which only changed the offset to a "vm_offset_t". Includes updates for the i386, pc532 and sh3 mmmmap from Jason Thorpe.
2000-06-26 08:55:19 +04:00
paddr_t
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_mmap(struct audio_softc *sc, off_t off, int prot)
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
{
minor KNF
2005-01-15 19:23:03 +03:00
const struct audio_hw_if *hw;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
struct audio_ringbuffer *cb;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
paddr_t rv;
KASSERT(mutex_owned(sc->sc_lock));
kern/45927: fix incorrect assert.
2012-02-22 00:53:34 +04:00
KASSERT(sc->sc_dvlock > 0);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
More printf format munging for off_t: when printing as %lld, cast to long long.
2000-07-19 17:44:24 +04:00
DPRINTF(("audio_mmap: off=%lld, prot=%d\n", (long long)off, prot));
minor KNF
2005-01-15 19:23:03 +03:00
hw = sc->hw_if;
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
if (!(audio_get_props(sc) & AUDIO_PROP_MMAP) || !hw->mappage)
Audio: Remove the machine dependant code I put in audio.c by mistake. This adds another method to audio_hw_if. Also remove a field from audio_hw_if that was not read-only.
1997-07-29 00:56:05 +04:00
return -1;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
#if 0
/* XXX
* The idea here was to use the protection to determine if
* we are mapping the read or write buffer, but it fails.
* The VM system is broken in (at least) two ways.
* 1) If you map memory VM_PROT_WRITE you SIGSEGV
* when writing to it, so VM_PROT_READ|VM_PROT_WRITE
* has to be used for mmapping the play buffer.
* 2) Even if calling mmap() with VM_PROT_READ|VM_PROT_WRITE
* audio_mmap will get called at some point with VM_PROT_READ
* only.
* So, alas, we always map the play buffer for now.
*/
if (prot == (VM_PROT_READ|VM_PROT_WRITE) ||
prot == VM_PROT_WRITE)
cb = &sc->sc_pr;
else if (prot == VM_PROT_READ)
cb = &sc->sc_rr;
else
Audio: Remove the machine dependant code I put in audio.c by mistake. This adds another method to audio_hw_if. Also remove a field from audio_hw_if that was not read-only.
1997-07-29 00:56:05 +04:00
return -1;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
#else
cb = &sc->sc_pr;
#endif
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if ((u_int)off >= cb->s.bufsize)
Audio: Remove the machine dependant code I put in audio.c by mistake. This adds another method to audio_hw_if. Also remove a field from audio_hw_if that was not read-only.
1997-07-29 00:56:05 +04:00
return -1;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
if (!cb->mmapped) {
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
cb->mmapped = true;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
if (cb == &sc->sc_pr) {
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
audio_fill_silence(&cb->s.param, cb->s.start,
cb->s.bufsize);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
sc->sc_pustream = &cb->s;
- Fix the conditions for POLLOUT to match those elsewhere - Refactor pause attribute handling so that it it's preserved appropriately and reset in audio_clear(). Now audio/jack-devel works with the native driver.
2008-06-03 10:21:17 +04:00
if (!sc->sc_pbus && !sc->sc_pr.pause)
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
(void)audiostartp(sc);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
} else {
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
sc->sc_rustream = &cb->s;
- Fix the conditions for POLLOUT to match those elsewhere - Refactor pause attribute handling so that it it's preserved appropriately and reset in audio_clear(). Now audio/jack-devel works with the native driver.
2008-06-03 10:21:17 +04:00
if (!sc->sc_rbus && !sc->sc_rr.pause)
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
(void)audiostartr(sc);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
}
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
rv = hw->mappage(sc->hw_hdl, cb->s.start, off, prot);
mutex_enter(sc->sc_lock);
return rv;
Add mmap method for audio devices. XXX This is just a placeholder to avoid updating conf.h again. Also, clean up some error values.
1996-09-02 01:33:43 +04:00
}
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
int
ANSIfy.
2001-10-03 03:31:54 +04:00
audiostartr(struct audio_softc *sc)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
prototypes for ccd audio and ncr5380
1996-03-07 18:00:07 +03:00
int error;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
KASSERT(mutex_owned(sc->sc_intr_lock));
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
DPRINTF(("audiostartr: start=%p used=%d(hi=%d) mmapped=%d\n",
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
sc->sc_rr.s.start, audio_stream_get_used(&sc->sc_rr.s),
sc->sc_rr.usedhigh, sc->sc_rr.mmapped));
New audio subsystem
1995-02-21 04:35:58 +03:00
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
if (!audio_can_capture(sc))
return EINVAL;
Add trigger_output and trigger_input entry points, which replace init_{in,out}put and start_{in,out}put, and are only called once.
1998-08-10 00:28:07 +04:00
if (sc->hw_if->trigger_input)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
error = sc->hw_if->trigger_input(sc->hw_hdl, sc->sc_rr.s.start,
sc->sc_rr.s.end, sc->sc_rr.blksize,
audio_rint, (void *)sc, &sc->sc_rr.s.param);
Add trigger_output and trigger_input entry points, which replace init_{in,out}put and start_{in,out}put, and are only called once.
1998-08-10 00:28:07 +04:00
else
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
error = sc->hw_if->start_input(sc->hw_hdl, sc->sc_rr.s.start,
Add trigger_output and trigger_input entry points, which replace init_{in,out}put and start_{in,out}put, and are only called once.
1998-08-10 00:28:07 +04:00
sc->sc_rr.blksize, audio_rint, (void *)sc);
prototypes for ccd audio and ncr5380
1996-03-07 18:00:07 +03:00
if (error) {
DPRINTF(("audiostartr failed: %d\n", error));
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
return error;
}
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
sc->sc_rbus = true;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
return 0;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
int
ANSIfy.
2001-10-03 03:31:54 +04:00
audiostartp(struct audio_softc *sc)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
prototypes for ccd audio and ncr5380
1996-03-07 18:00:07 +03:00
int error;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
int used;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
KASSERT(mutex_owned(sc->sc_intr_lock));
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
used = audio_stream_get_used(&sc->sc_pr.s);
* audiosetinfo() call audio_clear() only when blocksizes are actually changed. This fixes a problem on some audio devices and applications which use ossaudio and call SNDCTL_DSP_GETOSPACE repeatedly while playing. * audiostartp() add some debug prints.
2005-01-29 21:16:49 +03:00
DPRINTF(("audiostartp: start=%p used=%d(hi=%d blk=%d) mmapped=%d\n",
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
sc->sc_pr.s.start, used, sc->sc_pr.usedhigh,
* audiosetinfo() call audio_clear() only when blocksizes are actually changed. This fixes a problem on some audio devices and applications which use ossaudio and call SNDCTL_DSP_GETOSPACE repeatedly while playing. * audiostartp() add some debug prints.
2005-01-29 21:16:49 +03:00
sc->sc_pr.blksize, sc->sc_pr.mmapped));
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
if (!audio_can_playback(sc))
return EINVAL;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (!sc->sc_pr.mmapped && used < sc->sc_pr.blksize) {
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
cv_broadcast(&sc->sc_wchan);
* audiosetinfo() call audio_clear() only when blocksizes are actually changed. This fixes a problem on some audio devices and applications which use ossaudio and call SNDCTL_DSP_GETOSPACE repeatedly while playing. * audiostartp() add some debug prints.
2005-01-29 21:16:49 +03:00
DPRINTF(("%s: wakeup and return\n", __func__));
Add trigger_output and trigger_input entry points, which replace init_{in,out}put and start_{in,out}put, and are only called once.
1998-08-10 00:28:07 +04:00
return 0;
If we're short-circuiting audiostartp() because there isn't enough data present, do an extra wakeup() on the write channel. This happens when we change parameters during playback, at which point we flush the buffer, and would stall.
2004-07-08 02:30:22 +04:00
}
Add trigger_output and trigger_input entry points, which replace init_{in,out}put and start_{in,out}put, and are only called once.
1998-08-10 00:28:07 +04:00
* audiosetinfo() call audio_clear() only when blocksizes are actually changed. This fixes a problem on some audio devices and applications which use ossaudio and call SNDCTL_DSP_GETOSPACE repeatedly while playing. * audiostartp() add some debug prints.
2005-01-29 21:16:49 +03:00
if (sc->hw_if->trigger_output) {
DPRINTF(("%s: call trigger_output\n", __func__));
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
error = sc->hw_if->trigger_output(sc->hw_hdl, sc->sc_pr.s.start,
sc->sc_pr.s.end, sc->sc_pr.blksize,
fix a typo in audiostartp(). This change fixes a problem introduced by kent-audio1 merger with some audio hardware such as emuxki.
2005-01-15 07:57:18 +03:00
audio_pint, (void *)sc, &sc->sc_pr.s.param);
* audiosetinfo() call audio_clear() only when blocksizes are actually changed. This fixes a problem on some audio devices and applications which use ossaudio and call SNDCTL_DSP_GETOSPACE repeatedly while playing. * audiostartp() add some debug prints.
2005-01-29 21:16:49 +03:00
} else {
DPRINTF(("%s: call start_output\n", __func__));
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
error = sc->hw_if->start_output(sc->hw_hdl,
__UNCONST(sc->sc_pr.s.outp), sc->sc_pr.blksize,
audio_pint, (void *)sc);
* audiosetinfo() call audio_clear() only when blocksizes are actually changed. This fixes a problem on some audio devices and applications which use ossaudio and call SNDCTL_DSP_GETOSPACE repeatedly while playing. * audiostartp() add some debug prints.
2005-01-29 21:16:49 +03:00
}
Add trigger_output and trigger_input entry points, which replace init_{in,out}put and start_{in,out}put, and are only called once.
1998-08-10 00:28:07 +04:00
if (error) {
DPRINTF(("audiostartp failed: %d\n", error));
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
return error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
sc->sc_pbus = true;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
return 0;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Eliminate sc_silence, and handle output interrupts somewhat differently. From John Kohl, slightly edited.
1996-02-16 05:25:43 +03:00
/*
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
* When the play interrupt routine finds that the write isn't keeping
* the buffer filled it will insert silence in the buffer to make up
* for this. The part of the buffer that is filled with silence
Clarify a comment.
1998-08-05 20:38:09 +04:00
* is kept track of in a very approximate way: it starts at sc_sil_start
* and extends sc_sil_count bytes. If there is already silence in
* the requested area nothing is done; so when the whole buffer is
* silent nothing happens. When the writer starts again sc_sil_count
* is set to 0.
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
*
* XXX
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
* Putting silence into the output buffer should not really be done
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
* from the device interrupt handler. Consider deferring to the soft
* interrupt.
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
*/
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.
2005-12-24 23:27:29 +03:00
static inline void
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_pint_silence(struct audio_softc *sc, struct audio_ringbuffer *cb,
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
uint8_t *inp, int cc)
Eliminate sc_silence, and handle output interrupts somewhat differently. From John Kohl, slightly edited.
1996-02-16 05:25:43 +03:00
{
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
uint8_t *s, *e, *p, *q;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_intr_lock));
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
if (sc->sc_sil_count > 0) {
s = sc->sc_sil_start; /* start of silence */
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
e = s + sc->sc_sil_count; /* end of sil., may be beyond end */
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
p = inp; /* adjusted pointer to area to fill */
if (p < s)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
p += cb->s.end - cb->s.start;
q = p + cc;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
/* Check if there is already silence. */
if (!(s <= p && p < e &&
s <= q && q <= e)) {
if (s <= p)
Add a few more debug messages and clean up debug printing a little.
1998-04-28 13:07:12 +04:00
sc->sc_sil_count = max(sc->sc_sil_count, q-s);
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
DPRINTFN(5,("audio_pint_silence: fill cc=%d inp=%p, "
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
"count=%d size=%d\n",
cc, inp, sc->sc_sil_count,
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
(int)(cb->s.end - cb->s.start)));
audio_fill_silence(&cb->s.param, inp, cc);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
} else {
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
DPRINTFN(5,("audio_pint_silence: already silent "
"cc=%d inp=%p\n", cc, inp));
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
}
} else {
sc->sc_sil_start = inp;
sc->sc_sil_count = cc;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
DPRINTFN(5, ("audio_pint_silence: start fill %p %d\n",
inp, cc));
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
audio_fill_silence(&cb->s.param, inp, cc);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
}
Eliminate sc_silence, and handle output interrupts somewhat differently. From John Kohl, slightly edited.
1996-02-16 05:25:43 +03:00
}
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
static void
audio_softintr_rd(void *cookie)
{
struct audio_softc *sc = cookie;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
proc_t *p;
pid_t pid;
mutex_enter(sc->sc_lock);
cv_broadcast(&sc->sc_rchan);
selnotify(&sc->sc_rsel, 0, NOTE_SUBMIT);
if ((pid = sc->sc_async_audio) != 0) {
DPRINTFN(3, ("audio_softintr_rd: sending SIGIO %d\n", pid));
Network protocol interrupts can now block on locks, so merge the globals proclist_mutex and proclist_lock into a single adaptive mutex (proc_lock). Implications: - Inspecting process state requires thread context, so signals can no longer be sent from a hardware interrupt handler. Signal activity must be deferred to a soft interrupt or kthread. - As the proc state locking is simplified, it's now safe to take exit() and wait() out from under kernel_lock. - The system spends less time at IPL_SCHED, and there is less lock activity.
2008-04-24 19:35:27 +04:00
mutex_enter(proc_lock);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if ((p = proc_find(pid)) != NULL)
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
psignal(p, SIGIO);
Network protocol interrupts can now block on locks, so merge the globals proclist_mutex and proclist_lock into a single adaptive mutex (proc_lock). Implications: - Inspecting process state requires thread context, so signals can no longer be sent from a hardware interrupt handler. Signal activity must be deferred to a soft interrupt or kthread. - As the proc state locking is simplified, it's now safe to take exit() and wait() out from under kernel_lock. - The system spends less time at IPL_SCHED, and there is less lock activity.
2008-04-24 19:35:27 +04:00
mutex_exit(proc_lock);
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
}
static void
audio_softintr_wr(void *cookie)
{
struct audio_softc *sc = cookie;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
proc_t *p;
pid_t pid;
mutex_enter(sc->sc_lock);
cv_broadcast(&sc->sc_wchan);
selnotify(&sc->sc_wsel, 0, NOTE_SUBMIT);
if ((pid = sc->sc_async_audio) != 0) {
DPRINTFN(3, ("audio_softintr_wr: sending SIGIO %d\n", pid));
Network protocol interrupts can now block on locks, so merge the globals proclist_mutex and proclist_lock into a single adaptive mutex (proc_lock). Implications: - Inspecting process state requires thread context, so signals can no longer be sent from a hardware interrupt handler. Signal activity must be deferred to a soft interrupt or kthread. - As the proc state locking is simplified, it's now safe to take exit() and wait() out from under kernel_lock. - The system spends less time at IPL_SCHED, and there is less lock activity.
2008-04-24 19:35:27 +04:00
mutex_enter(proc_lock);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if ((p = proc_find(pid)) != NULL)
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
psignal(p, SIGIO);
Network protocol interrupts can now block on locks, so merge the globals proclist_mutex and proclist_lock into a single adaptive mutex (proc_lock). Implications: - Inspecting process state requires thread context, so signals can no longer be sent from a hardware interrupt handler. Signal activity must be deferred to a soft interrupt or kthread. - As the proc state locking is simplified, it's now safe to take exit() and wait() out from under kernel_lock. - The system spends less time at IPL_SCHED, and there is less lock activity.
2008-04-24 19:35:27 +04:00
mutex_exit(proc_lock);
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
}
New audio subsystem
1995-02-21 04:35:58 +03:00
/*
DMA, not dma nor Dma.
2003-05-03 22:10:37 +04:00
* Called from HW driver module on completion of DMA output.
New audio subsystem
1995-02-21 04:35:58 +03:00
* Start output of new block, wrap in ring buffer if needed.
* If no more buffers to play, output zero instead.
* Do a wakeup if necessary.
*/
void
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_pint(void *v)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
stream_fetcher_t null_fetcher;
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_softc *sc;
const struct audio_hw_if *hw;
struct audio_ringbuffer *cb;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
stream_fetcher_t *fetcher;
uint8_t *inp;
int cc, used;
Slight optimization in audio_[pr]int().
1998-08-28 16:07:41 +04:00
int blksize;
prototypes for ccd audio and ncr5380
1996-03-07 18:00:07 +03:00
int error;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
minor KNF
2005-01-15 19:23:03 +03:00
sc = v;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_intr_lock));
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
if (!sc->sc_open)
return; /* ignore interrupt if not open */
Aignore interrupts if the device isn't open.
1997-10-16 20:41:18 +04:00
minor KNF
2005-01-15 19:23:03 +03:00
hw = sc->hw_if;
cb = &sc->sc_pr;
Slight optimization in audio_[pr]int().
1998-08-28 16:07:41 +04:00
blksize = cb->blksize;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
cb->s.outp = audio_stream_add_outp(&cb->s, cb->s.outp, blksize);
cb->stamp += blksize;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
if (cb->mmapped) {
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
DPRINTFN(5, ("audio_pint: mmapped outp=%p cc=%d inp=%p\n",
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
cb->s.outp, blksize, cb->s.inp));
minor KNF
2005-01-15 19:23:03 +03:00
if (hw->trigger_output == NULL)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
(void)hw->start_output(sc->hw_hdl, __UNCONST(cb->s.outp),
Slight optimization in audio_[pr]int().
1998-08-28 16:07:41 +04:00
blksize, audio_pint, (void *)sc);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
return;
}
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
#ifdef AUDIO_INTR_TIME
{
struct timeval tv;
u_long t;
microtime(&tv);
t = tv.tv_usec + 1000000 * tv.tv_sec;
if (sc->sc_pnintr) {
long lastdelta, totdelta;
lastdelta = t - sc->sc_plastintr - sc->sc_pblktime;
Allow option AUDIO_INTR_TIME even if the platform does not support FP in the kernel.
1998-08-04 15:26:14 +04:00
if (lastdelta > sc->sc_pblktime / 3) {
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
printf("audio: play interrupt(%d) off "
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
"relative by %ld us (%lu)\n",
sc->sc_pnintr, lastdelta,
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
sc->sc_pblktime);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
}
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
totdelta = t - sc->sc_pfirstintr -
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
sc->sc_pblktime * sc->sc_pnintr;
Allow option AUDIO_INTR_TIME even if the platform does not support FP in the kernel.
1998-08-04 15:26:14 +04:00
if (totdelta > sc->sc_pblktime) {
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
printf("audio: play interrupt(%d) off "
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
"absolute by %ld us (%lu) (LOST)\n",
sc->sc_pnintr, totdelta,
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
sc->sc_pblktime);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
sc->sc_pnintr++; /* avoid repeated messages */
}
} else
sc->sc_pfirstintr = t;
sc->sc_plastintr = t;
sc->sc_pnintr++;
}
#endif
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
used = audio_stream_get_used(&cb->s);
/*
* "used <= cb->usedlow" should be "used < blksize" ideally.
* Some HW drivers such as uaudio(4) does not call audio_pint()
* at accurate timing. If used < blksize, uaudio(4) already
* request transfer of garbage data.
*/
if (used <= cb->usedlow && !cb->copying && sc->sc_npfilters > 0) {
/* we might have data in filter pipeline */
null_fetcher.fetch_to = null_fetcher_fetch_to;
fetcher = &sc->sc_pfilters[sc->sc_npfilters - 1]->base;
sc->sc_pfilters[0]->set_fetcher(sc->sc_pfilters[0],
&null_fetcher);
used = audio_stream_get_used(sc->sc_pustream);
cc = cb->s.end - cb->s.start;
If one or more playback filters are applied, minimize loading size of the hardware buffer to improve accuracy of AUDIO_GETIOFFS.
2006-05-16 17:46:19 +04:00
if (blksize * 2 < cc)
cc = blksize * 2;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
fetcher->fetch_to(sc, fetcher, &cb->s, cc);
Fixes a bug that audio_offset_t::samples for AUDIO_GETxOFFS has incorrect value. It also affects SNDCTL_DSP_GETOPTR of ossaudio. PR: kern/32405 by Stephane Witzman
2006-05-10 05:56:21 +04:00
cb->fstamp += used - audio_stream_get_used(sc->sc_pustream);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
used = audio_stream_get_used(&cb->s);
}
if (used < blksize) {
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
/* we don't have a full block to use */
if (cb->copying) {
/* writer is in progress, don't disturb */
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
cb->needfill = true;
Add a few more debug messages and clean up debug printing a little.
1998-04-28 13:07:12 +04:00
DPRINTFN(1, ("audio_pint: copying in progress\n"));
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
} else {
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
inp = cb->s.inp;
cc = blksize - (inp - cb->s.start) % blksize;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
if (cb->pause)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
cb->pdrops += cc;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
else {
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
cb->drops += cc;
sc->sc_playdrop += cc;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
}
audio_pint_silence(sc, cb, inp, cc);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
cb->s.inp = audio_stream_add_inp(&cb->s, inp, cc);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
/* Clear next block so we keep ahead of the DMA. */
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
used = audio_stream_get_used(&cb->s);
Fix botches in rev 1.185 (kent-audio1 merge) which cause buffer overrun and "tlb_exception: no copyin/out fault hander" panic in audio_fill_silence() after audio playback on dreamcast. The problem is tracked and fix is provieded by Ryo SHIMIZU, who is the auther of aica(4) audio driver for dreamcast.
2010-11-20 21:03:49 +03:00
if (used + blksize < cb->s.end - cb->s.start)
audio_pint_silence(sc, cb, cb->s.inp, blksize);
New audio subsystem
1995-02-21 04:35:58 +03:00
}
}
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
DPRINTFN(5, ("audio_pint: outp=%p cc=%d\n", cb->s.outp, blksize));
minor KNF
2005-01-15 19:23:03 +03:00
if (hw->trigger_output == NULL) {
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
error = hw->start_output(sc->hw_hdl, __UNCONST(cb->s.outp),
blksize, audio_pint, (void *)sc);
Add trigger_output and trigger_input entry points, which replace init_{in,out}put and start_{in,out}put, and are only called once.
1998-08-10 00:28:07 +04:00
if (error) {
/* XXX does this really help? */
DPRINTF(("audio_pint restart failed: %d\n", error));
audio_clear(sc);
}
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Move some debugging output to more useful levels.
1998-09-01 11:27:06 +04:00
DPRINTFN(2, ("audio_pint: mode=%d pause=%d used=%d lowat=%d\n",
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
sc->sc_mode, cb->pause,
audio_stream_get_used(sc->sc_pustream), cb->usedlow));
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
if ((sc->sc_mode & AUMODE_PLAY) && !cb->pause) {
Defer wakeup() to the soft interrupt also.
2007-02-15 21:12:05 +03:00
if (audio_stream_get_used(sc->sc_pustream) <= cb->usedlow)
Use the softint API.
2007-10-08 20:18:02 +04:00
softint_schedule(sc->sc_sih_wr);
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
/* Possible to return one or more "phantom blocks" now. */
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if (!sc->sc_full_duplex)
Use the softint API.
2007-10-08 20:18:02 +04:00
softint_schedule(sc->sc_sih_rd);
New audio subsystem
1995-02-21 04:35:58 +03:00
}
/*
DMA, not dma nor Dma.
2003-05-03 22:10:37 +04:00
* Called from HW driver module on completion of DMA input.
New audio subsystem
1995-02-21 04:35:58 +03:00
* Mark it as input in the ring buffer (fiddle pointers).
* Do a wakeup if necessary.
*/
void
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_rint(void *v)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
minor KNF
2005-01-15 19:23:03 +03:00
stream_fetcher_t null_fetcher;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
struct audio_softc *sc;
const struct audio_hw_if *hw;
struct audio_ringbuffer *cb;
stream_fetcher_t *last_fetcher;
int cc;
int used;
Slight optimization in audio_[pr]int().
1998-08-28 16:07:41 +04:00
int blksize;
prototypes for ccd audio and ncr5380
1996-03-07 18:00:07 +03:00
int error;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
sc = v;
cb = &sc->sc_rr;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_intr_lock));
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
if (!sc->sc_open)
return; /* ignore interrupt if not open */
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
hw = sc->hw_if;
Slight optimization in audio_[pr]int().
1998-08-28 16:07:41 +04:00
blksize = cb->blksize;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
cb->s.inp = audio_stream_add_inp(&cb->s, cb->s.inp, blksize);
Slight optimization in audio_[pr]int().
1998-08-28 16:07:41 +04:00
cb->stamp += blksize;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
if (cb->mmapped) {
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
DPRINTFN(2, ("audio_rint: mmapped inp=%p cc=%d\n",
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
cb->s.inp, blksize));
minor KNF
2005-01-15 19:23:03 +03:00
if (hw->trigger_input == NULL)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
(void)hw->start_input(sc->hw_hdl, cb->s.inp, blksize,
Add trigger_output and trigger_input entry points, which replace init_{in,out}put and start_{in,out}put, and are only called once.
1998-08-10 00:28:07 +04:00
audio_rint, (void *)sc);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
return;
}
#ifdef AUDIO_INTR_TIME
{
struct timeval tv;
u_long t;
microtime(&tv);
t = tv.tv_usec + 1000000 * tv.tv_sec;
if (sc->sc_rnintr) {
long lastdelta, totdelta;
lastdelta = t - sc->sc_rlastintr - sc->sc_rblktime;
if (lastdelta > sc->sc_rblktime / 5) {
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
printf("audio: record interrupt(%d) off "
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
"relative by %ld us (%lu)\n",
sc->sc_rnintr, lastdelta,
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
sc->sc_rblktime);
New audio subsystem
1995-02-21 04:35:58 +03:00
}
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
totdelta = t - sc->sc_rfirstintr -
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
sc->sc_rblktime * sc->sc_rnintr;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
if (totdelta > sc->sc_rblktime / 2) {
sc->sc_rnintr++;
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
printf("audio: record interrupt(%d) off "
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
"absolute by %ld us (%lu)\n",
sc->sc_rnintr, totdelta,
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
sc->sc_rblktime);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
sc->sc_rnintr++; /* avoid repeated messages */
}
} else
sc->sc_rfirstintr = t;
sc->sc_rlastintr = t;
sc->sc_rnintr++;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
#endif
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (!cb->pause && sc->sc_nrfilters > 0) {
null_fetcher.fetch_to = null_fetcher_fetch_to;
last_fetcher = &sc->sc_rfilters[sc->sc_nrfilters - 1]->base;
sc->sc_rfilters[0]->set_fetcher(sc->sc_rfilters[0],
&null_fetcher);
used = audio_stream_get_used(sc->sc_rustream);
cc = sc->sc_rustream->end - sc->sc_rustream->start;
error = last_fetcher->fetch_to
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
(sc, last_fetcher, sc->sc_rustream, cc);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
cb->fstamp += audio_stream_get_used(sc->sc_rustream) - used;
/* XXX what should do for error? */
}
used = audio_stream_get_used(&sc->sc_rr.s);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
if (cb->pause) {
Add a few more debug messages and clean up debug printing a little.
1998-04-28 13:07:12 +04:00
DPRINTFN(1, ("audio_rint: pdrops %lu\n", cb->pdrops));
Slight optimization in audio_[pr]int().
1998-08-28 16:07:41 +04:00
cb->pdrops += blksize;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
cb->s.outp = audio_stream_add_outp(&cb->s, cb->s.outp, blksize);
} else if (used + blksize > cb->s.end - cb->s.start && !cb->copying) {
Add a few more debug messages and clean up debug printing a little.
1998-04-28 13:07:12 +04:00
DPRINTFN(1, ("audio_rint: drops %lu\n", cb->drops));
Slight optimization in audio_[pr]int().
1998-08-28 16:07:41 +04:00
cb->drops += blksize;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
cb->s.outp = audio_stream_add_outp(&cb->s, cb->s.outp, blksize);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
DPRINTFN(2, ("audio_rint: inp=%p cc=%d\n", cb->s.inp, blksize));
minor KNF
2005-01-15 19:23:03 +03:00
if (hw->trigger_input == NULL) {
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
error = hw->start_input(sc->hw_hdl, cb->s.inp, blksize,
Add trigger_output and trigger_input entry points, which replace init_{in,out}put and start_{in,out}put, and are only called once.
1998-08-10 00:28:07 +04:00
audio_rint, (void *)sc);
if (error) {
/* XXX does this really help? */
DPRINTF(("audio_rint: restart failed: %d\n", error));
audio_clear(sc);
}
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
}
Use the softint API.
2007-10-08 20:18:02 +04:00
softint_schedule(sc->sc_sih_rd);
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Combine set_encoding and set_precision into a single set_format interface.
1997-03-20 09:48:48 +03:00
int
ANSIfy.
2001-10-03 03:31:54 +04:00
audio_check_params(struct audio_params *p)
Combine set_encoding and set_precision into a single set_format interface.
1997-03-20 09:48:48 +03:00
{
minor KNF
2005-01-15 19:23:03 +03:00
In the name of backwards compatibility AUDIO_ENCODING_LINEAR has been renamed AUDIO_ENCODING_SLINEAR and AUDIO_ENCODING_LINEAR reverts to the NetBSD 1.2 sematics. A kernel with COMPAT_12 defined will accept AUDIO_ENCODING_LINEAR and treat it as before, without COMPAT_12 it will be rejected.
1997-07-15 11:46:04 +04:00
if (p->encoding == AUDIO_ENCODING_PCM16) {
if (p->precision == 8)
p->encoding = AUDIO_ENCODING_ULINEAR;
else
p->encoding = AUDIO_ENCODING_SLINEAR;
} else if (p->encoding == AUDIO_ENCODING_PCM8) {
if (p->precision == 8)
p->encoding = AUDIO_ENCODING_ULINEAR;
else
return EINVAL;
}
if (p->encoding == AUDIO_ENCODING_SLINEAR)
Convert to new orthogonal audio encoding scheme and implement some of the new encodings. The change to ioctl AUDIO_GETENC is NOT backwards compatible.
1997-05-07 22:51:31 +04:00
#if BYTE_ORDER == LITTLE_ENDIAN
In the name of backwards compatibility AUDIO_ENCODING_LINEAR has been renamed AUDIO_ENCODING_SLINEAR and AUDIO_ENCODING_LINEAR reverts to the NetBSD 1.2 sematics. A kernel with COMPAT_12 defined will accept AUDIO_ENCODING_LINEAR and treat it as before, without COMPAT_12 it will be rejected.
1997-07-15 11:46:04 +04:00
p->encoding = AUDIO_ENCODING_SLINEAR_LE;
Convert to new orthogonal audio encoding scheme and implement some of the new encodings. The change to ioctl AUDIO_GETENC is NOT backwards compatible.
1997-05-07 22:51:31 +04:00
#else
In the name of backwards compatibility AUDIO_ENCODING_LINEAR has been renamed AUDIO_ENCODING_SLINEAR and AUDIO_ENCODING_LINEAR reverts to the NetBSD 1.2 sematics. A kernel with COMPAT_12 defined will accept AUDIO_ENCODING_LINEAR and treat it as before, without COMPAT_12 it will be rejected.
1997-07-15 11:46:04 +04:00
p->encoding = AUDIO_ENCODING_SLINEAR_BE;
Convert to new orthogonal audio encoding scheme and implement some of the new encodings. The change to ioctl AUDIO_GETENC is NOT backwards compatible.
1997-05-07 22:51:31 +04:00
#endif
if (p->encoding == AUDIO_ENCODING_ULINEAR)
#if BYTE_ORDER == LITTLE_ENDIAN
p->encoding = AUDIO_ENCODING_ULINEAR_LE;
#else
p->encoding = AUDIO_ENCODING_ULINEAR_BE;
#endif
Combine set_encoding and set_precision into a single set_format interface.
1997-03-20 09:48:48 +03:00
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
switch (p->encoding) {
Combine set_encoding and set_precision into a single set_format interface.
1997-03-20 09:48:48 +03:00
case AUDIO_ENCODING_ULAW:
case AUDIO_ENCODING_ALAW:
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
if (p->precision != 8)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return EINVAL;
Combine set_encoding and set_precision into a single set_format interface.
1997-03-20 09:48:48 +03:00
break;
X68k built-in voice synthesizer supports 4bit ADPCM.
2001-05-02 16:49:41 +04:00
case AUDIO_ENCODING_ADPCM:
if (p->precision != 4 && p->precision != 8)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return EINVAL;
X68k built-in voice synthesizer supports 4bit ADPCM.
2001-05-02 16:49:41 +04:00
break;
In the name of backwards compatibility AUDIO_ENCODING_LINEAR has been renamed AUDIO_ENCODING_SLINEAR and AUDIO_ENCODING_LINEAR reverts to the NetBSD 1.2 sematics. A kernel with COMPAT_12 defined will accept AUDIO_ENCODING_LINEAR and treat it as before, without COMPAT_12 it will be rejected.
1997-07-15 11:46:04 +04:00
case AUDIO_ENCODING_SLINEAR_LE:
case AUDIO_ENCODING_SLINEAR_BE:
Convert to new orthogonal audio encoding scheme and implement some of the new encodings. The change to ioctl AUDIO_GETENC is NOT backwards compatible.
1997-05-07 22:51:31 +04:00
case AUDIO_ENCODING_ULINEAR_LE:
case AUDIO_ENCODING_ULINEAR_BE:
Allow 24 and 32bit linear audio. Actually, the silence filler can do any multiple of 8 bits now, but I didn't allow the parameter check to accept more than 32 bit to avoid confusion of drivers that fail to check the parameters themselves thoroughly. This should be changed later.
2001-09-03 22:51:43 +04:00
/* XXX is: our zero-fill can handle any multiple of 8 */
if (p->precision != 8 && p->precision != 16 &&
p->precision != 24 && p->precision != 32)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return EINVAL;
if (p->precision == 8 && p->encoding == AUDIO_ENCODING_SLINEAR_BE)
p->encoding = AUDIO_ENCODING_SLINEAR_LE;
if (p->precision == 8 && p->encoding == AUDIO_ENCODING_ULINEAR_BE)
p->encoding = AUDIO_ENCODING_ULINEAR_LE;
if (p->validbits > p->precision)
return EINVAL;
Combine set_encoding and set_precision into a single set_format interface.
1997-03-20 09:48:48 +03:00
break;
TMS320AV110 (mpeg layer 2 decoder chip) and ZBUS Melody board, which uses it. No optionboards yet for Melody, only audio driver. MPEG data types for audio.c
1997-10-17 03:57:56 +04:00
case AUDIO_ENCODING_MPEG_L1_STREAM:
case AUDIO_ENCODING_MPEG_L1_PACKETS:
case AUDIO_ENCODING_MPEG_L1_SYSTEM:
case AUDIO_ENCODING_MPEG_L2_STREAM:
case AUDIO_ENCODING_MPEG_L2_PACKETS:
case AUDIO_ENCODING_MPEG_L2_SYSTEM:
Add AUDIO_ENCODING_AC3 encoding type.
2011-09-06 05:16:43 +04:00
case AUDIO_ENCODING_AC3:
TMS320AV110 (mpeg layer 2 decoder chip) and ZBUS Melody board, which uses it. No optionboards yet for Melody, only audio driver. MPEG data types for audio.c
1997-10-17 03:57:56 +04:00
break;
Combine set_encoding and set_precision into a single set_format interface.
1997-03-20 09:48:48 +03:00
default:
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return EINVAL;
Combine set_encoding and set_precision into a single set_format interface.
1997-03-20 09:48:48 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
/* sanity check # of channels*/
if (p->channels < 1 || p->channels > AUDIO_MAX_CHANNELS)
return EINVAL;
Fix divide by 0 bug in audio driver.
1997-07-28 03:06:04 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return 0;
Combine set_encoding and set_precision into a single set_format interface.
1997-03-20 09:48:48 +03:00
}
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
int
audio_set_defaults(struct audio_softc *sc, u_int mode)
{
struct audio_info ai;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
/* default parameters */
sc->sc_rparams = audio_default;
sc->sc_pparams = audio_default;
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
sc->sc_blkset = false;
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
AUDIO_INITINFO(&ai);
ai.record.sample_rate = sc->sc_rparams.sample_rate;
ai.record.encoding = sc->sc_rparams.encoding;
ai.record.channels = sc->sc_rparams.channels;
ai.record.precision = sc->sc_rparams.precision;
- Fix the conditions for POLLOUT to match those elsewhere - Refactor pause attribute handling so that it it's preserved appropriately and reset in audio_clear(). Now audio/jack-devel works with the native driver.
2008-06-03 10:21:17 +04:00
ai.record.pause = false;
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
ai.play.sample_rate = sc->sc_pparams.sample_rate;
ai.play.encoding = sc->sc_pparams.encoding;
ai.play.channels = sc->sc_pparams.channels;
ai.play.precision = sc->sc_pparams.precision;
- Fix the conditions for POLLOUT to match those elsewhere - Refactor pause attribute handling so that it it's preserved appropriately and reset in audio_clear(). Now audio/jack-devel works with the native driver.
2008-06-03 10:21:17 +04:00
ai.play.pause = false;
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
ai.mode = mode;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return audiosetinfo(sc, &ai);
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
}
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
int
ANSIfy.
2001-10-03 03:31:54 +04:00
au_set_lr_value(struct audio_softc *sc, mixer_ctrl_t *ct, int l, int r)
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
{
minor KNF
2005-01-15 19:23:03 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
ct->type = AUDIO_MIXER_VALUE;
ct->un.value.num_channels = 2;
ct->un.value.level[AUDIO_MIXER_LEVEL_LEFT] = l;
ct->un.value.level[AUDIO_MIXER_LEVEL_RIGHT] = r;
if (sc->hw_if->set_port(sc->hw_hdl, ct) == 0)
return 0;
ct->un.value.num_channels = 1;
ct->un.value.level[AUDIO_MIXER_LEVEL_MONO] = (l+r)/2;
return sc->hw_if->set_port(sc->hw_hdl, ct);
}
int
ANSIfy.
2001-10-03 03:31:54 +04:00
au_set_gain(struct audio_softc *sc, struct au_mixer_ports *ports,
int gain, int balance)
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
{
mixer_ctrl_t ct;
int i, error;
int l, r;
u_int mask;
int nset;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
if (balance == AUDIO_MID_BALANCE) {
l = r = gain;
} else if (balance < AUDIO_MID_BALANCE) {
l = gain;
Fix LR contrary problem of play.balance and record.balance. The change is based on PR kern/18901 by SUNAKAWA Hirokatsu.
2002-11-08 08:02:07 +03:00
r = (balance * gain) / AUDIO_MID_BALANCE;
} else {
r = gain;
l = ((AUDIO_RIGHT_BALANCE - balance) * gain)
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
/ AUDIO_MID_BALANCE;
}
DPRINTF(("au_set_gain: gain=%d balance=%d, l=%d r=%d\n",
gain, balance, l, r));
if (ports->index == -1) {
usemaster:
if (ports->master == -1)
return 0; /* just ignore it silently */
ct.dev = ports->master;
error = au_set_lr_value(sc, &ct, l, r);
} else {
ct.dev = ports->index;
if (ports->isenum) {
ct.type = AUDIO_MIXER_ENUM;
error = sc->hw_if->get_port(sc->hw_hdl, &ct);
if (error)
return error;
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
if (ports->isdual) {
if (ports->cur_port == -1)
ct.dev = ports->master;
else
ct.dev = ports->miport[ports->cur_port];
error = au_set_lr_value(sc, &ct, l, r);
} else {
for(i = 0; i < ports->nports; i++)
if (ports->misel[i] == ct.un.ord) {
ct.dev = ports->miport[i];
if (ct.dev == -1 ||
au_set_lr_value(sc, &ct, l, r))
goto usemaster;
else
break;
}
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
}
} else {
ct.type = AUDIO_MIXER_SET;
error = sc->hw_if->get_port(sc->hw_hdl, &ct);
if (error)
return error;
mask = ct.un.mask;
nset = 0;
for(i = 0; i < ports->nports; i++) {
if (ports->misel[i] & mask) {
ct.dev = ports->miport[i];
if (ct.dev != -1 &&
au_set_lr_value(sc, &ct, l, r) == 0)
nset++;
}
}
if (nset == 0)
goto usemaster;
}
}
if (!error)
mixer_signal(sc);
return error;
}
int
ANSIfy.
2001-10-03 03:31:54 +04:00
au_get_lr_value(struct audio_softc *sc, mixer_ctrl_t *ct, int *l, int *r)
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
{
int error;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
ct->un.value.num_channels = 2;
if (sc->hw_if->get_port(sc->hw_hdl, ct) == 0) {
*l = ct->un.value.level[AUDIO_MIXER_LEVEL_LEFT];
*r = ct->un.value.level[AUDIO_MIXER_LEVEL_RIGHT];
} else {
ct->un.value.num_channels = 1;
error = sc->hw_if->get_port(sc->hw_hdl, ct);
if (error)
return error;
*r = *l = ct->un.value.level[AUDIO_MIXER_LEVEL_MONO];
}
return 0;
}
void
ANSIfy.
2001-10-03 03:31:54 +04:00
au_get_gain(struct audio_softc *sc, struct au_mixer_ports *ports,
u_int *pgain, u_char *pbalance)
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
{
mixer_ctrl_t ct;
int i, l, r, n;
minor KNF
2005-01-15 19:23:03 +03:00
int lgain, rgain;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
minor KNF
2005-01-15 19:23:03 +03:00
lgain = AUDIO_MAX_GAIN / 2;
rgain = AUDIO_MAX_GAIN / 2;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
if (ports->index == -1) {
usemaster:
if (ports->master == -1)
goto bad;
ct.dev = ports->master;
ct.type = AUDIO_MIXER_VALUE;
if (au_get_lr_value(sc, &ct, &lgain, &rgain))
goto bad;
} else {
ct.dev = ports->index;
if (ports->isenum) {
ct.type = AUDIO_MIXER_ENUM;
if (sc->hw_if->get_port(sc->hw_hdl, &ct))
goto bad;
ct.type = AUDIO_MIXER_VALUE;
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
if (ports->isdual) {
if (ports->cur_port == -1)
ct.dev = ports->master;
else
ct.dev = ports->miport[ports->cur_port];
au_get_lr_value(sc, &ct, &lgain, &rgain);
} else {
for(i = 0; i < ports->nports; i++)
if (ports->misel[i] == ct.un.ord) {
ct.dev = ports->miport[i];
if (ct.dev == -1 ||
au_get_lr_value(sc, &ct,
&lgain, &rgain))
goto usemaster;
else
break;
}
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
}
} else {
ct.type = AUDIO_MIXER_SET;
if (sc->hw_if->get_port(sc->hw_hdl, &ct))
goto bad;
ct.type = AUDIO_MIXER_VALUE;
lgain = rgain = n = 0;
for(i = 0; i < ports->nports; i++) {
if (ports->misel[i] & ct.un.mask) {
ct.dev = ports->miport[i];
if (ct.dev == -1 ||
au_get_lr_value(sc, &ct, &l, &r))
goto usemaster;
else {
lgain += l;
rgain += r;
n++;
}
}
}
if (n != 0) {
lgain /= n;
rgain /= n;
}
}
}
bad:
if (lgain == rgain) { /* handles lgain==rgain==0 */
*pgain = lgain;
*pbalance = AUDIO_MID_BALANCE;
} else if (lgain < rgain) {
*pgain = rgain;
Fix LR contrary problem of play.balance and record.balance. The change is based on PR kern/18901 by SUNAKAWA Hirokatsu.
2002-11-08 08:02:07 +03:00
/* balance should be > AUDIO_MID_BALANCE */
*pbalance = AUDIO_RIGHT_BALANCE -
(AUDIO_MID_BALANCE * lgain) / rgain;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
} else /* lgain > rgain */ {
*pgain = lgain;
Fix LR contrary problem of play.balance and record.balance. The change is based on PR kern/18901 by SUNAKAWA Hirokatsu.
2002-11-08 08:02:07 +03:00
/* balance should be < AUDIO_MID_BALANCE */
*pbalance = (AUDIO_MID_BALANCE * rgain) / lgain;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
}
}
int
ANSIfy.
2001-10-03 03:31:54 +04:00
au_set_port(struct audio_softc *sc, struct au_mixer_ports *ports, u_int port)
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
{
mixer_ctrl_t ct;
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
int i, error, use_mixerout;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
use_mixerout = 1;
if (port == 0) {
if (ports->allports == 0)
return 0; /* Allow this special case. */
else if (ports->isdual) {
if (ports->cur_port == -1) {
return 0;
} else {
port = ports->aumask[ports->cur_port];
ports->cur_port = -1;
use_mixerout = 0;
}
}
}
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
if (ports->index == -1)
return EINVAL;
ct.dev = ports->index;
if (ports->isenum) {
if (port & (port-1))
return EINVAL; /* Only one port allowed */
ct.type = AUDIO_MIXER_ENUM;
error = EINVAL;
for(i = 0; i < ports->nports; i++)
if (ports->aumask[i] == port) {
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
if (ports->isdual && use_mixerout) {
ct.un.ord = ports->mixerout;
ports->cur_port = i;
} else {
ct.un.ord = ports->misel[i];
}
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
error = sc->hw_if->set_port(sc->hw_hdl, &ct);
break;
}
} else {
ct.type = AUDIO_MIXER_SET;
ct.un.mask = 0;
for(i = 0; i < ports->nports; i++)
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
if (ports->aumask[i] & port)
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
ct.un.mask |= ports->misel[i];
if (port != 0 && ct.un.mask == 0)
error = EINVAL;
else
error = sc->hw_if->set_port(sc->hw_hdl, &ct);
}
if (!error)
mixer_signal(sc);
return error;
}
int
ANSIfy.
2001-10-03 03:31:54 +04:00
au_get_port(struct audio_softc *sc, struct au_mixer_ports *ports)
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
{
mixer_ctrl_t ct;
int i, aumask;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
if (ports->index == -1)
return 0;
ct.dev = ports->index;
ct.type = ports->isenum ? AUDIO_MIXER_ENUM : AUDIO_MIXER_SET;
if (sc->hw_if->get_port(sc->hw_hdl, &ct))
return 0;
aumask = 0;
if (ports->isenum) {
Fix a couple of long-standing bugs in the volume control(s) part of the audio device interface: 1) When attempting to match the appropriate mixer control, we weren't checking the class label, but only the second level label, so for devices that had both an "inputs.cd" and a "record.cd", for example, we could never do the right thing except by chance. For this reason, evidently, all the record masters were labeled (by the underlying drivers) either "record.record" or "record.volume", to distinguish from "outputs.master". We'll now accept "record.master", and document that as the the new preferred way. 2) More importantly, the model was deficient. Selecting a port on many chips completely disables most of the level controls, which doesn't play nice with other applications which are trying to use the interface. Now, selecting a port simply sets which mixer input control shall be changed, setting state in the audio layer. In other words, the "mixerout" port is really selected all the time, enabling the final stage mixer, and setting "gain" sets the level of the appropriate input. It should be possible for separate applications to each control the mic, dac, and cd inputs at the same time using this interface, simply by reiterating their port selections with each change, but applications that don't bother to do that aren't any worse off than they were before. The user is expected to set the master output with another application, dedicated to that task. Though it is now meaningful to select "no port" with the audio device interface, to manipulate the master output, there's no particular reason for an audio device consumer to do that. (I added the capability in order to restore the initial state of the audio device, for testing purposes. It might also be useful to users of broken binary- only applications.) Observe that the mixer device interface (and so, "mixerctl") still retains all capabilities, including the ability to set the actual input port on the chip, overriding the level controls. No change is being made to the mixer device interface. The mixer device simply presents all the controls on the chip, with no attempt at abstraction, so there are no bugs there. The upshot is, that applications that have been trying to use the audio device interface to change the volume, such as mplayer, now "just work". I've tested these changes extensively with "eso" and "eap" since first proposing them on tech-kern last January, and somewhat with "esm" and a few others. This closes both PR kern/10221, and PR kern/17159.
2004-01-31 03:07:56 +03:00
if (ports->isdual && ports->cur_port != -1) {
if (ports->mixerout == ct.un.ord)
aumask = ports->aumask[ports->cur_port];
else
ports->cur_port = -1;
}
if (aumask == 0)
for(i = 0; i < ports->nports; i++)
if (ports->misel[i] == ct.un.ord)
aumask = ports->aumask[i];
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
} else {
for(i = 0; i < ports->nports; i++)
if (ct.un.mask & ports->misel[i])
aumask |= ports->aumask[i];
}
return aumask;
}
Make some functions non-static.
1995-04-18 03:04:31 +04:00
int
ANSIfy.
2001-10-03 03:31:54 +04:00
audiosetinfo(struct audio_softc *sc, struct audio_info *ai)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
stream_filter_list_t pfilters, rfilters;
audio_params_t pp, rp;
struct audio_prinfo *r, *p;
const struct audio_hw_if *hw;
audio_stream_t *oldpus, *oldrus;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
int setmode;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
int error;
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
int np, nr;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
unsigned int blks;
Make a slight change to when blocksize, hi- and lo-water marks are recalculated. The new semantics is more intuitive(?) and makes Doom work :-)
1997-08-19 03:20:09 +04:00
int oldpblksize, oldrblksize;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
u_int gain;
Replace the Mach-derived boolean_t type with the C99 bool type. A future commit will replace use of TRUE and FALSE with true and false.
2007-02-22 01:59:35 +03:00
bool rbus, pbus;
bool cleared, modechange, pausechange;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
u_char balance;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
hw = sc->hw_if;
minor KNF
2005-01-15 19:23:03 +03:00
if (hw == NULL) /* HW has not attached */
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return ENXIO;
New audio subsystem
1995-02-21 04:35:58 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
DPRINTF(("%s sc=%p ai=%p\n", __func__, sc, ai));
r = &ai->record;
p = &ai->play;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
rbus = sc->sc_rbus;
pbus = sc->sc_pbus;
error = 0;
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
cleared = false;
modechange = false;
pausechange = false;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
pp = sc->sc_pparams; /* Temporary encoding storage in */
rp = sc->sc_rparams; /* case setting the modes fails. */
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
nr = np = 0;
Make a slight change to when blocksize, hi- and lo-water marks are recalculated. The new semantics is more intuitive(?) and makes Doom work :-)
1997-08-19 03:20:09 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(p->sample_rate)) {
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
pp.sample_rate = p->sample_rate;
np++;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(r->sample_rate)) {
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
rp.sample_rate = r->sample_rate;
nr++;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(p->encoding)) {
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
pp.encoding = p->encoding;
np++;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(r->encoding)) {
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
rp.encoding = r->encoding;
nr++;
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(p->precision)) {
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
pp.precision = p->precision;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
/* we don't have API to specify validbits */
pp.validbits = p->precision;
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
np++;
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(r->precision)) {
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
rp.precision = r->precision;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
/* we don't have API to specify validbits */
rp.validbits = r->precision;
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
nr++;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(p->channels)) {
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
pp.channels = p->channels;
np++;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(r->channels)) {
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
rp.channels = r->channels;
nr++;
}
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
if (!audio_can_capture(sc))
nr = 0;
if (!audio_can_playback(sc))
np = 0;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
#ifdef AUDIO_DEBUG
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (audiodebug && nr > 0)
minor KNF
2005-01-15 19:23:03 +03:00
audio_print_params("audiosetinfo() Setting record params:", &rp);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (audiodebug && np > 0)
minor KNF
2005-01-15 19:23:03 +03:00
audio_print_params("audiosetinfo() Setting play params:", &pp);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
#endif
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (nr > 0 && (error = audio_check_params(&rp)))
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
return error;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (np > 0 && (error = audio_check_params(&pp)))
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
return error;
Set oldpblksize and oldrblksize early enough so that zero-initialisation isn't necessary. The GCC warning was not a false-positive, AFAICS.
2006-11-01 13:13:37 +03:00
oldpblksize = sc->sc_pr.blksize;
oldrblksize = sc->sc_rr.blksize;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
setmode = 0;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (nr > 0) {
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
if (!cleared) {
deal with PR#46232: ensure locking around audio_clear() is handled consistently and avoid locking against self.
2012-03-26 22:26:10 +04:00
audio_clear_intr_unlocked(sc);
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
cleared = true;
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
}
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
modechange = true;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
setmode |= AUMODE_RECORD;
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (np > 0) {
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
if (!cleared) {
deal with PR#46232: ensure locking around audio_clear() is handled consistently and avoid locking against self.
2012-03-26 22:26:10 +04:00
audio_clear_intr_unlocked(sc);
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
cleared = true;
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
}
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
modechange = true;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
setmode |= AUMODE_PLAY;
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(ai->mode)) {
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
if (!cleared) {
deal with PR#46232: ensure locking around audio_clear() is handled consistently and avoid locking against self.
2012-03-26 22:26:10 +04:00
audio_clear_intr_unlocked(sc);
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
cleared = true;
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
}
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
modechange = true;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
sc->sc_mode = ai->mode;
if (sc->sc_mode & AUMODE_PLAY_ALL)
sc->sc_mode |= AUMODE_PLAY;
if ((sc->sc_mode & AUMODE_PLAY) && !sc->sc_full_duplex)
/* Play takes precedence */
sc->sc_mode &= ~AUMODE_RECORD;
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
oldpus = sc->sc_pustream;
oldrus = sc->sc_rustream;
Call set_params if usemode changes.
1998-08-10 05:13:20 +04:00
if (modechange) {
Sampling rate and mono-stereo conversion described in http://mail-index.netbsd.org/tech-kern/2002/03/04/0005.html auconv.c: Add conversion functions audio.c: Sample alignment, calling conversion functions, etc. audio_if.h: Add four hw_* members to "struct audio_params" audiovar.h: Add conversion buffers, etc. auich and uaudio: Add conversion request code to *_set_params().
2002-03-07 17:37:02 +03:00
int indep;
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
indep = audio_get_props(sc) & AUDIO_PROP_INDEPENDENT;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
if (!indep) {
if (setmode == AUMODE_RECORD)
pp = rp;
else if (setmode == AUMODE_PLAY)
rp = pp;
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
memset(&pfilters, 0, sizeof(pfilters));
memset(&rfilters, 0, sizeof(rfilters));
pfilters.append = stream_filter_list_append;
pfilters.prepend = stream_filter_list_prepend;
pfilters.set = stream_filter_list_set;
rfilters.append = stream_filter_list_append;
rfilters.prepend = stream_filter_list_prepend;
rfilters.set = stream_filter_list_set;
Sampling rate and mono-stereo conversion described in http://mail-index.netbsd.org/tech-kern/2002/03/04/0005.html auconv.c: Add conversion functions audio.c: Sample alignment, calling conversion functions, etc. audio_if.h: Add four hw_* members to "struct audio_params" audiovar.h: Add conversion buffers, etc. auich and uaudio: Add conversion request code to *_set_params().
2002-03-07 17:37:02 +03:00
/* Some device drivers change channels/sample_rate and change
* no channels/sample_rate. */
Mask off extra usemode bits.
1998-08-09 09:44:51 +04:00
error = hw->set_params(sc->hw_hdl, setmode,
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
sc->sc_mode & (AUMODE_PLAY | AUMODE_RECORD), &pp, &rp,
&pfilters, &rfilters);
if (error) {
DPRINTF(("%s: hw->set_params() failed with %d\n",
__func__, error));
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
goto cleanup;
Sampling rate and mono-stereo conversion described in http://mail-index.netbsd.org/tech-kern/2002/03/04/0005.html auconv.c: Add conversion functions audio.c: Sample alignment, calling conversion functions, etc. audio_if.h: Add four hw_* members to "struct audio_params" audiovar.h: Add conversion buffers, etc. auich and uaudio: Add conversion request code to *_set_params().
2002-03-07 17:37:02 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
audio_check_params(&pp);
audio_check_params(&rp);
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
if (!indep) {
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
/* XXX for !indep device, we have to use the same
* parameters for the hardware, not userland */
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
if (setmode == AUMODE_RECORD) {
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
pp = rp;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
} else if (setmode == AUMODE_PLAY) {
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
rp = pp;
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
}
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (sc->sc_pr.mmapped && pfilters.req_size > 0) {
DPRINTF(("%s: mmapped, and filters are requested.\n",
__func__));
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
error = EINVAL;
goto cleanup;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
}
/* construct new filter chain */
if (setmode & AUMODE_PLAY) {
error = audio_setup_pfilters(sc, &pp, &pfilters);
if (error)
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
goto cleanup;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
}
if (setmode & AUMODE_RECORD) {
error = audio_setup_rfilters(sc, &rp, &rfilters);
if (error)
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
goto cleanup;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
}
DPRINTF(("%s: filter setup is completed.\n", __func__));
/* userland formats */
Always copy back both the record and play parameters, in case the driver had to change one behind our back.
1998-08-09 08:54:44 +04:00
sc->sc_pparams = pp;
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
sc->sc_rparams = rp;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Make a slight change to when blocksize, hi- and lo-water marks are recalculated. The new semantics is more intuitive(?) and makes Doom work :-)
1997-08-19 03:20:09 +04:00
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
/* Play params can affect the record params, so recalculate blksize. */
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (nr > 0 || np > 0) {
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
audio_calc_blksize(sc, AUMODE_RECORD);
audio_calc_blksize(sc, AUMODE_PLAY);
}
#ifdef AUDIO_DEBUG
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (audiodebug > 1 && nr > 0)
minor KNF
2005-01-15 19:23:03 +03:00
audio_print_params("audiosetinfo() After setting record params:", &sc->sc_rparams);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (audiodebug > 1 && np > 0)
minor KNF
2005-01-15 19:23:03 +03:00
audio_print_params("audiosetinfo() After setting play params:", &sc->sc_pparams);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
#endif
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(p->port)) {
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
if (!cleared) {
deal with PR#46232: ensure locking around audio_clear() is handled consistently and avoid locking against self.
2012-03-26 22:26:10 +04:00
audio_clear_intr_unlocked(sc);
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
cleared = true;
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
}
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
error = au_set_port(sc, &sc->sc_outports, p->port);
Check for errors while changing gain.
1996-02-20 14:47:22 +03:00
if (error)
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
goto cleanup;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(r->port)) {
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
if (!cleared) {
deal with PR#46232: ensure locking around audio_clear() is handled consistently and avoid locking against self.
2012-03-26 22:26:10 +04:00
audio_clear_intr_unlocked(sc);
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
cleared = true;
A bit of cleanup: * Separate the code to set the default parameters into a new function, audio_set_defaults(). Make it use audiosetinfo(), which properly initializes the block size and whatnot. Use this in both audioattach() and the /dev/audio case of audio_open(). * Do not force a reinitialization when /dev/sound is opened. * Do all of the block size sanity checks in auto_init_ringbuffer(), not in both audio_calc_blksize() and audiosetinfo(). * Fix a bug in audiosetinfo() that caused the block size to not be recalculated immediately if we set it to 0. * For AUDIO_GET[IO]OFFS, modify the deltablks calculation so that it gives us the number of block boundaries crossed.
2003-10-02 11:15:20 +04:00
}
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
error = au_set_port(sc, &sc->sc_inports, r->port);
Check for errors while changing gain.
1996-02-20 14:47:22 +03:00
if (error)
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
goto cleanup;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(p->gain)) {
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
au_get_gain(sc, &sc->sc_outports, &gain, &balance);
error = au_set_gain(sc, &sc->sc_outports, p->gain, balance);
Check for errors while changing gain.
1996-02-20 14:47:22 +03:00
if (error)
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
goto cleanup;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(r->gain)) {
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
au_get_gain(sc, &sc->sc_inports, &gain, &balance);
error = au_set_gain(sc, &sc->sc_inports, r->gain, balance);
if (error)
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
goto cleanup;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
}
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED_CH(p->balance)) {
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
au_get_gain(sc, &sc->sc_outports, &gain, &balance);
error = au_set_gain(sc, &sc->sc_outports, gain, p->balance);
if (error)
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
goto cleanup;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED_CH(r->balance)) {
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
au_get_gain(sc, &sc->sc_inports, &gain, &balance);
error = au_set_gain(sc, &sc->sc_inports, gain, r->balance);
if (error)
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
goto cleanup;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(ai->monitor_gain) && sc->sc_monitor_port != -1) {
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
mixer_ctrl_t ct;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
ct.dev = sc->sc_monitor_port;
Set mixer type for rec and play level set. PR#1027
1995-05-08 20:06:38 +04:00
ct.type = AUDIO_MIXER_VALUE;
New audio subsystem
1995-02-21 04:35:58 +03:00
ct.un.value.num_channels = 1;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
ct.un.value.level[AUDIO_MIXER_LEVEL_MONO] = ai->monitor_gain;
Fix two small bugs: (1) Fix sense of a strcmp() test so that the monitor subdevice of the mixer is properly identified. (2) Use hw_if->set_port() instead of hw_if->get_port() when changinh the monitor level in audiosetinfo(). Fixes PR kern/7661.
1999-05-30 04:21:08 +04:00
error = sc->hw_if->set_port(sc->hw_hdl, &ct);
Check for errors while changing gain.
1996-02-20 14:47:22 +03:00
if (error)
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
goto cleanup;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED_CH(p->pause)) {
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
sc->sc_pr.pause = p->pause;
Make sure buffer sizes are initialized correctly even when the pause state is explicitly set. Without this, the machine would crash in the audio interupt when the driver needs to divide by the block size (e.g., cs4281.c/cs4280.c). Idea for the fix by yamt.
2005-06-11 12:14:19 +04:00
pbus = !p->pause;
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
pausechange = true;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED_CH(r->pause)) {
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
sc->sc_rr.pause = r->pause;
Make sure buffer sizes are initialized correctly even when the pause state is explicitly set. Without this, the machine would crash in the audio interupt when the driver needs to divide by the block size (e.g., cs4281.c/cs4280.c). Idea for the fix by yamt.
2005-06-11 12:14:19 +04:00
rbus = !r->pause;
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
pausechange = true;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(ai->blocksize)) {
* audiosetinfo() call audio_clear() only when blocksizes are actually changed. This fixes a problem on some audio devices and applications which use ossaudio and call SNDCTL_DSP_GETOSPACE repeatedly while playing. * audiostartp() add some debug prints.
2005-01-29 21:16:49 +03:00
int pblksize, rblksize;
Make sure to recalculate the block size after changing the sampling rate. Also, nuke the global variable audio_blocksize.
1997-03-20 19:13:55 +03:00
/* Block size specified explicitly. */
Make a slight change to when blocksize, hi- and lo-water marks are recalculated. The new semantics is more intuitive(?) and makes Doom work :-)
1997-08-19 03:20:09 +04:00
if (ai->blocksize == 0) {
* audiosetinfo() call audio_clear() only when blocksizes are actually changed. This fixes a problem on some audio devices and applications which use ossaudio and call SNDCTL_DSP_GETOSPACE repeatedly while playing. * audiostartp() add some debug prints.
2005-01-29 21:16:49 +03:00
if (!cleared) {
deal with PR#46232: ensure locking around audio_clear() is handled consistently and avoid locking against self.
2012-03-26 22:26:10 +04:00
audio_clear_intr_unlocked(sc);
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
cleared = true;
* audiosetinfo() call audio_clear() only when blocksizes are actually changed. This fixes a problem on some audio devices and applications which use ossaudio and call SNDCTL_DSP_GETOSPACE repeatedly while playing. * audiostartp() add some debug prints.
2005-01-29 21:16:49 +03:00
}
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
sc->sc_blkset = false;
Make a slight change to when blocksize, hi- and lo-water marks are recalculated. The new semantics is more intuitive(?) and makes Doom work :-)
1997-08-19 03:20:09 +04:00
audio_calc_blksize(sc, AUMODE_RECORD);
audio_calc_blksize(sc, AUMODE_PLAY);
} else {
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
sc->sc_blkset = true;
* audiosetinfo() call audio_clear() only when blocksizes are actually changed. This fixes a problem on some audio devices and applications which use ossaudio and call SNDCTL_DSP_GETOSPACE repeatedly while playing. * audiostartp() add some debug prints.
2005-01-29 21:16:49 +03:00
/* check whether new blocksize changes actually */
if (hw->round_blocksize == NULL) {
if (!cleared) {
deal with PR#46232: ensure locking around audio_clear() is handled consistently and avoid locking against self.
2012-03-26 22:26:10 +04:00
audio_clear_intr_unlocked(sc);
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
cleared = true;
* audiosetinfo() call audio_clear() only when blocksizes are actually changed. This fixes a problem on some audio devices and applications which use ossaudio and call SNDCTL_DSP_GETOSPACE repeatedly while playing. * audiostartp() add some debug prints.
2005-01-29 21:16:49 +03:00
}
sc->sc_pr.blksize = ai->blocksize;
sc->sc_rr.blksize = ai->blocksize;
} else {
pblksize = hw->round_blocksize(sc->hw_hdl,
ai->blocksize, AUMODE_PLAY, &sc->sc_pr.s.param);
rblksize = hw->round_blocksize(sc->hw_hdl,
ai->blocksize, AUMODE_RECORD, &sc->sc_rr.s.param);
if (pblksize != sc->sc_pr.blksize ||
rblksize != sc->sc_rr.blksize) {
if (!cleared) {
deal with PR#46232: ensure locking around audio_clear() is handled consistently and avoid locking against self.
2012-03-26 22:26:10 +04:00
audio_clear_intr_unlocked(sc);
TRUE -> true, FALSE -> false
2007-03-01 20:31:35 +03:00
cleared = true;
* audiosetinfo() call audio_clear() only when blocksizes are actually changed. This fixes a problem on some audio devices and applications which use ossaudio and call SNDCTL_DSP_GETOSPACE repeatedly while playing. * audiostartp() add some debug prints.
2005-01-29 21:16:49 +03:00
}
sc->sc_pr.blksize = ai->blocksize;
sc->sc_rr.blksize = ai->blocksize;
}
}
Make a slight change to when blocksize, hi- and lo-water marks are recalculated. The new semantics is more intuitive(?) and makes Doom work :-)
1997-08-19 03:20:09 +04:00
}
Make sure to recalculate the block size after changing the sampling rate. Also, nuke the global variable audio_blocksize.
1997-03-20 19:13:55 +03:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(ai->mode)) {
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
if (sc->sc_mode & AUMODE_PLAY)
New audio subsystem
1995-02-21 04:35:58 +03:00
audio_init_play(sc);
Make mode a flag set. Full-dpulex fixes from John Kohl.
1995-07-07 05:52:30 +04:00
if (sc->sc_mode & AUMODE_RECORD)
New audio subsystem
1995-02-21 04:35:58 +03:00
audio_init_record(sc);
}
Make sure to recalculate the block size after changing the sampling rate. Also, nuke the global variable audio_blocksize.
1997-03-20 19:13:55 +03:00
audio: Simplify handling of AUDIO_SETFD and committing of encoding mode.
1997-07-28 03:51:48 +04:00
if (hw->commit_settings) {
error = hw->commit_settings(sc->hw_hdl);
if (error)
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
goto cleanup;
audio: Simplify handling of AUDIO_SETFD and committing of encoding mode.
1997-07-28 03:51:48 +04:00
}
New audio subsystem
1995-02-21 04:35:58 +03:00
When resuming the audio device, call audiosetinfo with the last known good parameters. Fixes an issue with suspending the audio device while playback is paused.
2007-12-13 17:02:53 +03:00
sc->sc_lastinfo = *ai;
sc->sc_lastinfovalid = true;
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
cleanup:
Make sure buffer sizes are initialized correctly even when the pause state is explicitly set. Without this, the machine would crash in the audio interupt when the driver needs to divide by the block size (e.g., cs4281.c/cs4280.c). Idea for the fix by yamt.
2005-06-11 12:14:19 +04:00
if (cleared || pausechange) {
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
int init_error;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
init_error = audio_initbufs(sc);
if (init_error) goto err;
Make a slight change to when blocksize, hi- and lo-water marks are recalculated. The new semantics is more intuitive(?) and makes Doom work :-)
1997-08-19 03:20:09 +04:00
if (sc->sc_pr.blksize != oldpblksize ||
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
sc->sc_rr.blksize != oldrblksize ||
sc->sc_pustream != oldpus ||
sc->sc_rustream != oldrus)
Make a slight change to when blocksize, hi- and lo-water marks are recalculated. The new semantics is more intuitive(?) and makes Doom work :-)
1997-08-19 03:20:09 +04:00
audio_calcwater(sc);
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
if ((sc->sc_mode & AUMODE_PLAY) &&
pbus && !sc->sc_pbus)
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
init_error = audiostartp(sc);
if (!init_error &&
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
(sc->sc_mode & AUMODE_RECORD) &&
rbus && !sc->sc_rbus)
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
init_error = audiostartr(sc);
- Change audio_hw_if a little: set_param now sets the play and record modes at the same time instead by using two different calls. This enables it to check more easily if the combined mode is all right. - Improve the error checking in audio.c. - Add a new audio property, AUDIO_PROP_INDEPENDENT, show if the play and record settings are independent. - Fix some buglets in audio.c.
1997-08-25 02:31:23 +04:00
err:
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
if (init_error)
return init_error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
Restructure a little to make sure AUDIO_FLUSH restarts the looping DMA.
1997-08-07 03:08:26 +04:00
/* Change water marks after initializing the buffers. */
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(ai->hiwat)) {
Restructure a little to make sure AUDIO_FLUSH restarts the looping DMA.
1997-08-07 03:08:26 +04:00
blks = ai->hiwat;
if (blks > sc->sc_pr.maxblks)
blks = sc->sc_pr.maxblks;
Sanity check hi/lo water mark settings better.
1997-10-08 02:40:43 +04:00
if (blks < 2)
blks = 2;
Restructure a little to make sure AUDIO_FLUSH restarts the looping DMA.
1997-08-07 03:08:26 +04:00
sc->sc_pr.usedhigh = blks * sc->sc_pr.blksize;
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(ai->lowat)) {
Restructure a little to make sure AUDIO_FLUSH restarts the looping DMA.
1997-08-07 03:08:26 +04:00
blks = ai->lowat;
if (blks > sc->sc_pr.maxblks - 1)
blks = sc->sc_pr.maxblks - 1;
sc->sc_pr.usedlow = blks * sc->sc_pr.blksize;
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
if (SPECIFIED(ai->hiwat) || SPECIFIED(ai->lowat)) {
Sanity check hi/lo water mark settings better.
1997-10-08 02:40:43 +04:00
if (sc->sc_pr.usedlow > sc->sc_pr.usedhigh - sc->sc_pr.blksize)
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
sc->sc_pr.usedlow =
Fold lines to 80 chars.
1998-12-28 01:52:23 +03:00
sc->sc_pr.usedhigh - sc->sc_pr.blksize;
Sanity check hi/lo water mark settings better.
1997-10-08 02:40:43 +04:00
}
Restructure a little to make sure AUDIO_FLUSH restarts the looping DMA.
1997-08-07 03:08:26 +04:00
Fix for kern/17446. Restore the audio settings if an error occured.
2006-11-01 09:39:19 +03:00
return error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Make some functions non-static.
1995-04-18 03:04:31 +04:00
int
Add a new ioctl AUDIO_GETBUFINFO. It works like AUDIO_GETINFO, but doesn't obtain the ports, gain and balance related parameters. Those generally require reading from the hardware and therefore are much more expensive to obtain. Modify OSS emulation to use the new ioctl where possible. This reduces CPU usage of mplayer during mp3 playback with my Thinkpad from 20% to < 1% and from 50% to 20% during Xvid playback. Review and comments from jmcneill@
2007-06-11 17:05:46 +04:00
audiogetinfo(struct audio_softc *sc, struct audio_info *ai, int buf_only_mode)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_prinfo *r, *p;
const struct audio_hw_if *hw;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
minor KNF
2005-01-15 19:23:03 +03:00
r = &ai->record;
p = &ai->play;
hw = sc->hw_if;
if (hw == NULL) /* HW has not attached */
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return ENXIO;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
Change the interface between the high and low level audio drivers so that all audio encoding parameters are set at once. This simplifies the interface and make error checking easier.
1997-04-30 01:01:33 +04:00
p->sample_rate = sc->sc_pparams.sample_rate;
r->sample_rate = sc->sc_rparams.sample_rate;
p->channels = sc->sc_pparams.channels;
r->channels = sc->sc_rparams.channels;
p->precision = sc->sc_pparams.precision;
r->precision = sc->sc_rparams.precision;
p->encoding = sc->sc_pparams.encoding;
r->encoding = sc->sc_rparams.encoding;
New audio subsystem
1995-02-21 04:35:58 +03:00
Add a new ioctl AUDIO_GETBUFINFO. It works like AUDIO_GETINFO, but doesn't obtain the ports, gain and balance related parameters. Those generally require reading from the hardware and therefore are much more expensive to obtain. Modify OSS emulation to use the new ioctl where possible. This reduces CPU usage of mplayer during mp3 playback with my Thinkpad from 20% to < 1% and from 50% to 20% during Xvid playback. Review and comments from jmcneill@
2007-06-11 17:05:46 +04:00
if (buf_only_mode) {
r->port = 0;
p->port = 0;
New audio subsystem
1995-02-21 04:35:58 +03:00
Add a new ioctl AUDIO_GETBUFINFO. It works like AUDIO_GETINFO, but doesn't obtain the ports, gain and balance related parameters. Those generally require reading from the hardware and therefore are much more expensive to obtain. Modify OSS emulation to use the new ioctl where possible. This reduces CPU usage of mplayer during mp3 playback with my Thinkpad from 20% to < 1% and from 50% to 20% during Xvid playback. Review and comments from jmcneill@
2007-06-11 17:05:46 +04:00
r->avail_ports = 0;
p->avail_ports = 0;
New audio subsystem
1995-02-21 04:35:58 +03:00
Add a new ioctl AUDIO_GETBUFINFO. It works like AUDIO_GETINFO, but doesn't obtain the ports, gain and balance related parameters. Those generally require reading from the hardware and therefore are much more expensive to obtain. Modify OSS emulation to use the new ioctl where possible. This reduces CPU usage of mplayer during mp3 playback with my Thinkpad from 20% to < 1% and from 50% to 20% during Xvid playback. Review and comments from jmcneill@
2007-06-11 17:05:46 +04:00
r->gain = 0;
r->balance = 0;
p->gain = 0;
p->balance = 0;
} else {
r->port = au_get_port(sc, &sc->sc_inports);
p->port = au_get_port(sc, &sc->sc_outports);
r->avail_ports = sc->sc_inports.allports;
p->avail_ports = sc->sc_outports.allports;
au_get_gain(sc, &sc->sc_inports, &r->gain, &r->balance);
au_get_gain(sc, &sc->sc_outports, &p->gain, &p->balance);
}
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
Add a new ioctl AUDIO_GETBUFINFO. It works like AUDIO_GETINFO, but doesn't obtain the ports, gain and balance related parameters. Those generally require reading from the hardware and therefore are much more expensive to obtain. Modify OSS emulation to use the new ioctl where possible. This reduces CPU usage of mplayer during mp3 playback with my Thinkpad from 20% to < 1% and from 50% to 20% during Xvid playback. Review and comments from jmcneill@
2007-06-11 17:05:46 +04:00
if (sc->sc_monitor_port != -1 && buf_only_mode == 0) {
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
mixer_ctrl_t ct;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
ct.dev = sc->sc_monitor_port;
ct.type = AUDIO_MIXER_VALUE;
ct.un.value.num_channels = 1;
if (sc->hw_if->get_port(sc->hw_hdl, &ct))
ai->monitor_gain = 0;
else
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
ai->monitor_gain =
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
ct.un.value.level[AUDIO_MIXER_LEVEL_MONO];
} else
ai->monitor_gain = 0;
New audio subsystem
1995-02-21 04:35:58 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
p->seek = audio_stream_get_used(sc->sc_pustream);
r->seek = audio_stream_get_used(sc->sc_rustream);
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
/*
* XXX samples should be a value for userland data.
* But drops is a value for HW data.
*/
p->samples = (sc->sc_pustream == &sc->sc_pr.s
? sc->sc_pr.stamp : sc->sc_pr.fstamp) - sc->sc_pr.drops;
r->samples = (sc->sc_rustream == &sc->sc_rr.s
? sc->sc_rr.stamp : sc->sc_rr.fstamp) - sc->sc_rr.drops;
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
p->eof = sc->sc_eof;
r->eof = 0;
New audio subsystem
1995-02-21 04:35:58 +03:00
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
p->pause = sc->sc_pr.pause;
r->pause = sc->sc_rr.pause;
New audio subsystem
1995-02-21 04:35:58 +03:00
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
p->error = sc->sc_pr.drops != 0;
r->error = sc->sc_rr.drops != 0;
New audio subsystem
1995-02-21 04:35:58 +03:00
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
p->waiting = r->waiting = 0; /* open never hangs */
New audio subsystem
1995-02-21 04:35:58 +03:00
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
p->open = (sc->sc_open & AUOPEN_WRITE) != 0;
r->open = (sc->sc_open & AUOPEN_READ) != 0;
p->active = sc->sc_pbus;
r->active = sc->sc_rbus;
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
p->buffer_size = sc->sc_pustream ? sc->sc_pustream->bufsize : 0;
r->buffer_size = sc->sc_rustream ? sc->sc_rustream->bufsize : 0;
Make the audio API (almost) SunOS compatible. The changes is to allow some limited mixer manipulation through the audio device (instead of the mixer device). This rendered 4 methods in audio_hw_if unused so garbage collect these.
1997-10-19 11:41:33 +04:00
Changes to the sudio system: - It is now possible to handle devices that want "looping" DMA, e.g. the SoundBlaster correctly. The WSS and SB drivers use this. To do this several new methods were introduced in audio_hw_if. - Different silence handling (forced by previous change). - The audio driver can now be mmap()-ed, but due to problems in the VM system only for writing for now. - The OSS (Linux) audio emulation takes advantage of some of the new features.
1997-07-27 05:16:32 +04:00
ai->blocksize = sc->sc_pr.blksize;
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
if (sc->sc_pr.blksize > 0) {
ai->hiwat = sc->sc_pr.usedhigh / sc->sc_pr.blksize;
ai->lowat = sc->sc_pr.usedlow / sc->sc_pr.blksize;
} else
ai->hiwat = ai->lowat = 0;
Make mode a flag set. Full-dpulex fixes from John Kohl.
1995-07-07 05:52:30 +04:00
ai->mode = sc->sc_mode;
New audio subsystem
1995-02-21 04:35:58 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return 0;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
/*
* Mixer driver
*/
int
__unused removal on arguments; approved by core.
2006-11-16 04:32:37 +03:00
mixer_open(dev_t dev, struct audio_softc *sc, int flags,
int ifmt, struct lwp *l)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
minor KNF
2005-01-15 19:23:03 +03:00
if (sc->hw_if == NULL)
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return ENXIO;
New audio subsystem
1995-02-21 04:35:58 +03:00
Add support for detaching audio devices.
1999-09-09 14:24:39 +04:00
DPRINTF(("mixer_open: flags=0x%x sc=%p\n", flags, sc));
New audio subsystem
1995-02-21 04:35:58 +03:00
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return 0;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Implement delivery of SIGIO on changes to /dev/mixer.
1997-08-11 05:38:12 +04:00
/*
* Remove a process from those to be signalled on mixer activity.
*/
static void
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mixer_remove(struct audio_softc *sc)
Implement delivery of SIGIO on changes to /dev/mixer.
1997-08-11 05:38:12 +04:00
{
struct mixer_asyncs **pm, *m;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
pid_t pid;
In mixer_remove() it's better to check if lwp is not NULL before trying to access a children node of it. Fixes PR 32823, by Yours Truly. I guess I should read the code before sending a PR.
2006-02-25 14:32:51 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
Implement delivery of SIGIO on changes to /dev/mixer.
1997-08-11 05:38:12 +04:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
pid = curproc->p_pid;
minor KNF
2005-01-15 19:23:03 +03:00
for (pm = &sc->sc_async_mixer; *pm; pm = &(*pm)->next) {
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if ((*pm)->pid == pid) {
Implement delivery of SIGIO on changes to /dev/mixer.
1997-08-11 05:38:12 +04:00
m = *pm;
*pm = m->next;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
kmem_free(m, sizeof(*m));
Implement delivery of SIGIO on changes to /dev/mixer.
1997-08-11 05:38:12 +04:00
return;
}
}
}
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
/*
ANSIfy.
2001-10-03 03:31:54 +04:00
* Signal all processes waiting for the mixer.
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
*/
static void
ANSIfy.
2001-10-03 03:31:54 +04:00
mixer_signal(struct audio_softc *sc)
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
{
struct mixer_asyncs *m;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
proc_t *p;
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
for (m = sc->sc_async_mixer; m; m = m->next) {
Network protocol interrupts can now block on locks, so merge the globals proclist_mutex and proclist_lock into a single adaptive mutex (proc_lock). Implications: - Inspecting process state requires thread context, so signals can no longer be sent from a hardware interrupt handler. Signal activity must be deferred to a soft interrupt or kthread. - As the proc state locking is simplified, it's now safe to take exit() and wait() out from under kernel_lock. - The system spends less time at IPL_SCHED, and there is less lock activity.
2008-04-24 19:35:27 +04:00
mutex_enter(proc_lock);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if ((p = proc_find(m->pid)) != NULL)
psignal(p, SIGIO);
Network protocol interrupts can now block on locks, so merge the globals proclist_mutex and proclist_lock into a single adaptive mutex (proc_lock). Implications: - Inspecting process state requires thread context, so signals can no longer be sent from a hardware interrupt handler. Signal activity must be deferred to a soft interrupt or kthread. - As the proc state locking is simplified, it's now safe to take exit() and wait() out from under kernel_lock. - The system spends less time at IPL_SCHED, and there is less lock activity.
2008-04-24 19:35:27 +04:00
mutex_exit(proc_lock);
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
}
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
}
New audio subsystem
1995-02-21 04:35:58 +03:00
/*
* Close a mixer device
*/
/* ARGSUSED */
int
__unused removal on arguments; approved by core.
2006-11-16 04:32:37 +03:00
mixer_close(struct audio_softc *sc, int flags, int ifmt,
- sprinkle __unused on function decls. - fix a couple of unused bugs - no more -Wno-unused for i386
2006-10-12 05:30:41 +04:00
struct lwp *l)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
minor KNF
2005-01-15 19:23:03 +03:00
DPRINTF(("mixer_close: sc %p\n", sc));
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mixer_remove(sc);
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return 0;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
int
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.
2007-03-04 08:59:00 +03:00
mixer_ioctl(struct audio_softc *sc, u_long cmd, void *addr, int flag,
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
struct lwp *l)
New audio subsystem
1995-02-21 04:35:58 +03:00
{
minor KNF
2005-01-15 19:23:03 +03:00
const struct audio_hw_if *hw;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
struct mixer_asyncs *ma;
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
mixer_ctrl_t *mc;
minor KNF
2005-01-15 19:23:03 +03:00
int error;
New audio subsystem
1995-02-21 04:35:58 +03:00
Change the cmd argument in {audio,mixer}_ioctl() to u_long.
2000-03-01 03:44:35 +03:00
DPRINTF(("mixer_ioctl(%lu,'%c',%lu)\n",
IOCPARM_LEN(cmd), (char)IOCGROUP(cmd), cmd&0xff));
minor KNF
2005-01-15 19:23:03 +03:00
hw = sc->hw_if;
error = EINVAL;
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
/* we can return cached values if we are sleeping */
if (cmd != AUDIO_MIXER_READ)
Split device_t and softc for audio(4), and other related cosmetic changes.
2008-03-04 21:23:44 +03:00
device_active(sc->dev, DVA_SYSTEM);
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
New audio subsystem
1995-02-21 04:35:58 +03:00
switch (cmd) {
Implement delivery of SIGIO on changes to /dev/mixer.
1997-08-11 05:38:12 +04:00
case FIOASYNC:
if (*(int *)addr) {
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
ma = kmem_alloc(sizeof(struct mixer_asyncs), KM_SLEEP);
mutex_enter(sc->sc_lock);
} else {
ma = NULL;
}
mixer_remove(sc); /* remove old entry */
if (ma != NULL) {
Implement delivery of SIGIO on changes to /dev/mixer.
1997-08-11 05:38:12 +04:00
ma->next = sc->sc_async_mixer;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
ma->pid = curproc->p_pid;
Implement delivery of SIGIO on changes to /dev/mixer.
1997-08-11 05:38:12 +04:00
sc->sc_async_mixer = ma;
}
error = 0;
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
case AUDIO_GETDEV:
audio: Simplify handling of AUDIO_SETFD and committing of encoding mode.
1997-07-28 03:51:48 +04:00
DPRINTF(("AUDIO_GETDEV\n"));
error = hw->getdev(sc->hw_hdl, (audio_device_t *)addr);
break;
whitespace and folding fix.
2002-03-23 20:17:10 +03:00
New audio subsystem
1995-02-21 04:35:58 +03:00
case AUDIO_MIXER_DEVINFO:
audio: Simplify handling of AUDIO_SETFD and committing of encoding mode.
1997-07-28 03:51:48 +04:00
DPRINTF(("AUDIO_MIXER_DEVINFO\n"));
Add a field in mixer_devinfo_t that indicates the minimum delta to change mixer value. A value of 0 indicates that the driver doesn't supply a delta. The change is backwards compatible.
2000-12-29 13:00:08 +03:00
((mixer_devinfo_t *)addr)->un.v.delta = 0; /* default */
audio: Simplify handling of AUDIO_SETFD and committing of encoding mode.
1997-07-28 03:51:48 +04:00
error = hw->query_devinfo(sc->hw_hdl, (mixer_devinfo_t *)addr);
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
case AUDIO_MIXER_READ:
audio: Simplify handling of AUDIO_SETFD and committing of encoding mode.
1997-07-28 03:51:48 +04:00
DPRINTF(("AUDIO_MIXER_READ\n"));
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
mc = (mixer_ctrl_t *)addr;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if (device_is_active(sc->sc_dev))
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
error = hw->get_port(sc->hw_hdl, mc);
else if (mc->dev >= sc->sc_nmixer_states)
error = ENXIO;
else {
int dev = mc->dev;
memcpy(mc, &sc->sc_mixer_state[dev],
sizeof(mixer_ctrl_t));
error = 0;
}
audio: Simplify handling of AUDIO_SETFD and committing of encoding mode.
1997-07-28 03:51:48 +04:00
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
case AUDIO_MIXER_WRITE:
audio: Simplify handling of AUDIO_SETFD and committing of encoding mode.
1997-07-28 03:51:48 +04:00
DPRINTF(("AUDIO_MIXER_WRITE\n"));
error = hw->set_port(sc->hw_hdl, (mixer_ctrl_t *)addr);
if (!error && hw->commit_settings)
error = hw->commit_settings(sc->hw_hdl);
Implement /dev/audioctl. It behaves much like /dev/sound, but only allows ioctl() and can be opened at any time.
1997-08-19 01:19:02 +04:00
if (!error)
mixer_signal(sc);
audio: Simplify handling of AUDIO_SETFD and committing of encoding mode.
1997-07-28 03:51:48 +04:00
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
default:
Fix pasto in type of dev_ioctl method.
2001-10-04 00:48:41 +04:00
if (hw->dev_ioctl)
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
error = hw->dev_ioctl(sc->hw_hdl, cmd, addr, flag, l);
Fix pasto in type of dev_ioctl method.
2001-10-04 00:48:41 +04:00
else
Add a new optional method, dev_ioctl, to the audio hardware driver interface. It is called when an unrecognized ioctl() is performed on a device, thus allowing ioctl()s that frob the hardware driver (like loading microcode).
2001-10-03 04:04:47 +04:00
error = EINVAL;
audio: Simplify handling of AUDIO_SETFD and committing of encoding mode.
1997-07-28 03:51:48 +04:00
break;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Change the cmd argument in {audio,mixer}_ioctl() to u_long.
2000-03-01 03:44:35 +03:00
DPRINTF(("mixer_ioctl(%lu,'%c',%lu) result %d\n",
IOCPARM_LEN(cmd), (char)IOCGROUP(cmd), cmd&0xff, error));
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return error;
New audio subsystem
1995-02-21 04:35:58 +03:00
}
Handle the case of MIDI devices with no audio devices.
1998-08-28 11:44:12 +04:00
#endif /* NAUDIO > 0 */
#include "midi.h"
Make midi_attach_mi() and audioprint() available also if MIDIBUS (but not necessarily MIDI) is defined. Fixes PR port-i386/6615 (Krister Walfridsson <cato@ulysses.df.lth.se>). (XXX this should probably go into a separate file to get rid of the #ifdef mess)
1998-12-20 17:26:44 +03:00
#if NAUDIO == 0 && (NMIDI > 0 || NMIDIBUS > 0)
Make this compile again with no audio but with midi devices.
1998-09-27 20:43:56 +04:00
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/device.h>
#include <sys/audioio.h>
#include <dev/audio_if.h>
#endif
Make midi_attach_mi() and audioprint() available also if MIDIBUS (but not necessarily MIDI) is defined. Fixes PR port-i386/6615 (Krister Walfridsson <cato@ulysses.df.lth.se>). (XXX this should probably go into a separate file to get rid of the #ifdef mess)
1998-12-20 17:26:44 +03:00
#if NAUDIO > 0 || (NMIDI > 0 || NMIDIBUS > 0)
Handle the case of MIDI devices with no audio devices.
1998-08-28 11:44:12 +04:00
int
ANSIfy.
2001-10-03 03:31:54 +04:00
audioprint(void *aux, const char *pnp)
Handle the case of MIDI devices with no audio devices.
1998-08-28 11:44:12 +04:00
{
minor KNF
2005-01-15 19:23:03 +03:00
struct audio_attach_args *arg;
Handle the case of MIDI devices with no audio devices.
1998-08-28 11:44:12 +04:00
const char *type;
if (pnp != NULL) {
minor KNF
2005-01-15 19:23:03 +03:00
arg = aux;
Handle the case of MIDI devices with no audio devices.
1998-08-28 11:44:12 +04:00
switch (arg->type) {
case AUDIODEV_TYPE_AUDIO:
type = "audio";
break;
case AUDIODEV_TYPE_MIDI:
type = "midi";
break;
case AUDIODEV_TYPE_OPL:
type = "opl";
break;
case AUDIODEV_TYPE_MPU:
type = "mpu";
break;
default:
panic("audioprint: unknown type %d", arg->type);
}
Use aprint_normal() in cfprint routines.
2003-01-01 03:10:15 +03:00
aprint_normal("%s at %s", type, pnp);
Handle the case of MIDI devices with no audio devices.
1998-08-28 11:44:12 +04:00
}
merge kent-audio1 branch, which introduces audio filter pipeline to the MI audio framework Summary of changes: * struct audio_params - remove sw_code, factor, factor_denom, hw_sample_rate, hw_encoding ,hw_precision, and hw_channels. Conversion information is conveyed by stream_filter_list_t. - change the type of sample_rate: u_long -> u_int - add `validbits,' which represents the valid data size in precision bits. It is required in order to distinguish 24/32bit from 24/24bit or 32/32bit. * audio_hw_if - add two parameters to set_params() stream_filter_list_t *pfil, stream_filter_list *rfil A HW driver should set filter recipes for requested formats - constify audio_params parameters of trigger_output() and trigger_input(). They represent audio formats for the hardware. - make open() and close() optional - add int (AUMODE_PLAY or AUMODE_RECORD) and audio_params_t parameters to round_blocksize() * sw_code is replaced with stream_filter_t. stream_filer_t converts audio data in an input buffer and writes into another output buffer unlike sw_code, which converts data in single buffer. converters in dev/auconv.c, dev/mulaw.c, dev/aurateconv.c, dev/tc/bba.c, dev/ic/msm6258.c, and arch/arm/iomd/vidcaudio.c are reimplemented as stream_filter_t * MI audio - audiosetinfo() builds filter pipelines from stream_filter_list_t filled by audio_hw_if::set_params() - audiosetinfo() returns with EINVAL if mmapped and set_params() requests filters - audio_write(), audio_pint(), and audio_rint() invoke a filter pipeline. - ioctl() for FIONREAD, AUDIO_WSEEK, AUDIO_GETIOFFS, AUDIO_GETOOFFS, and audio_prinfo::{seek,samples} for AUDIO_GETINFO handle values for a buffer nearest to userland. * add `struct device *' parameter to ac97_attach() * all of audio HW drivers follow audio_hw_if and ac97 changes
2005-01-11 01:01:36 +03:00
return UNCONF;
Handle the case of MIDI devices with no audio devices.
1998-08-28 11:44:12 +04:00
}
Make midi_attach_mi() and audioprint() available also if MIDIBUS (but not necessarily MIDI) is defined. Fixes PR port-i386/6615 (Krister Walfridsson <cato@ulysses.df.lth.se>). (XXX this should probably go into a separate file to get rid of the #ifdef mess)
1998-12-20 17:26:44 +03:00
#endif /* NAUDIO > 0 || (NMIDI > 0 || NMIDIBUS > 0) */
Halt input/output on soft suspend if currently running, and restore them on soft resume.
2006-06-19 14:19:08 +04:00
add a helper function to get the hw device_t from an audio_softc
2014-11-18 04:50:12 +03:00
#if NAUDIO > 0
device_t
audio_get_device(struct audio_softc *sc)
{
return sc->sc_dev;
}
#endif
Protect audio_powerhook with '#if NAUDIO > 0'. PR kern/33832.
2006-06-27 06:07:08 +04:00
#if NAUDIO > 0
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
static void
audio_mixer_capture(struct audio_softc *sc)
Halt input/output on soft suspend if currently running, and restore them on soft resume.
2006-06-19 14:19:08 +04:00
{
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
mixer_devinfo_t mi;
mixer_ctrl_t *mc;
Halt input/output on soft suspend if currently running, and restore them on soft resume.
2006-06-19 14:19:08 +04:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
Halt input/output on soft suspend if currently running, and restore them on soft resume.
2006-06-19 14:19:08 +04:00
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
for (mi.index = 0;; mi.index++) {
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
if (sc->hw_if->query_devinfo(sc->hw_hdl, &mi) != 0)
break;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mi.index < sc->sc_nmixer_states);
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
if (mi.type == AUDIO_MIXER_CLASS)
continue;
mc = &sc->sc_mixer_state[mi.index];
mc->dev = mi.index;
mc->type = mi.type;
When collecting mixer values while suspending, initialize the number of channels from the mixer device info. Fixes PR kern/37512. OK: jmcneill
2007-12-11 03:08:14 +03:00
mc->un.value.num_channels = mi.un.v.num_channels;
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
(void)sc->hw_if->get_port(sc->hw_hdl, mc);
Halt input/output on soft suspend if currently running, and restore them on soft resume.
2006-06-19 14:19:08 +04:00
}
return;
}
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
static void
audio_mixer_restore(struct audio_softc *sc)
{
mixer_devinfo_t mi;
mixer_ctrl_t *mc;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
for (mi.index = 0; ; mi.index++) {
if (sc->hw_if->query_devinfo(sc->hw_hdl, &mi) != 0)
break;
if (mi.type == AUDIO_MIXER_CLASS)
continue;
mc = &sc->sc_mixer_state[mi.index];
(void)sc->hw_if->set_port(sc->hw_hdl, mc);
}
if (sc->hw_if->commit_settings)
sc->hw_if->commit_settings(sc->hw_hdl);
return;
}
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#ifdef AUDIO_PM_IDLE
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
static void
audio_idle(void *arg)
{
device_t dv = arg;
struct audio_softc *sc = device_private(dv);
#ifdef PNP_DEBUG
extern int pnp_debug_idle;
if (pnp_debug_idle)
printf("%s: idle handler called\n", device_xname(dv));
#endif
Don't count poll(2) and select(2) as device activity. Don't activate the device on write(2) if it is suspended and was not idle.
2007-12-13 17:40:36 +03:00
sc->sc_idle = true;
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
/* XXX joerg Make pmf_device_suspend handle children? */
In pmf(9), improve the implementation of device self-suspension and make suspension by self, by drvctl(8), and by ACPI system sleep play nice together. Start solidifying some temporary API changes. 1. Extract a new header file, <sys/device_if.h>, from <sys/device.h> and #include it from <sys/pmf.h> instead of <sys/device.h> to break the circular dependency between <sys/device.h> and <sys/pmf.h>. 2. Introduce pmf_qual_t, an aggregate of qualifications on a PMF suspend/resume call. Start to replace instances of PMF_FN_PROTO, PMF_FN_ARGS, et cetera, with a pmf_qual_t. 3. Introduce the notion of a "suspensor," an entity that holds a device in suspension. More than one suspensor may hold a device at once. A device stays suspended as long as at least one suspensor holds it. A device resumes when the last suspensor releases it. Currently, the kernel defines three suspensors, 3a the system-suspensor: for system suspension, initiated by 'sysctl -w machdep.sleep_state=3', by lid closure, by power-button press, et cetera, 3b the drvctl-suspensor: for device suspension by /dev/drvctl ioctl, e.g., drvctl -S sip0. 3c the system self-suspensor: for device drivers that suspend themselves and their children. Several drivers for network interfaces put the network device to sleep while it is not administratively up, that is, after the kernel calls if_stop(, 1). The self-suspensor should not be used directly. See the description of suspensor delegates, below. A suspensor can have one or more "delegates". A suspensor can release devices that its delegates hold suspended. Right now, only the system self-suspensor has delegates. For each device that a self-suspending driver attaches, it creates the device's self-suspensor, a delegate of the system self-suspensor. Suspensors stop a system-wide suspend/resume cycle from waking devices that the operator put to sleep with drvctl before the cycle. They also help self-suspension to work more simply, safely, and in accord with expectations. 4. Add the notion of device activation level, devact_level_t, and a routine for checking the current activation level, device_activation(). Current activation levels are DEVACT_LEVEL_BUS, DEVACT_LEVEL_DRIVER, and DEVACT_LEVEL_CLASS, which respectively indicate that the device's bus is active, that the bus and device are active, and that the bus, device, and the functions of the device's class (network, audio) are active. Suspend/resume calls can be qualified with a devact_level_t. The power-management framework treats a devact_level_t that qualifies a device suspension as the device's current activation level; it only runs hooks to reduce the activation level from the presumed current level to the fully suspended state. The framework treats a devact_level_t qualifying device resumption as the target activation level; it only runs hooks to raise the activation level to the target. 5. Use pmf_qual_t, devact_level_t, and self-suspensors in several drivers. 6. Temporarily add an unused power-management workqueue that I will remove or replace, soon.
2009-09-16 20:34:49 +04:00
if (!pmf_device_suspend(dv, PMF_Q_SELF))
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
return;
In pmf(9), improve the implementation of device self-suspension and make suspension by self, by drvctl(8), and by ACPI system sleep play nice together. Start solidifying some temporary API changes. 1. Extract a new header file, <sys/device_if.h>, from <sys/device.h> and #include it from <sys/pmf.h> instead of <sys/device.h> to break the circular dependency between <sys/device.h> and <sys/pmf.h>. 2. Introduce pmf_qual_t, an aggregate of qualifications on a PMF suspend/resume call. Start to replace instances of PMF_FN_PROTO, PMF_FN_ARGS, et cetera, with a pmf_qual_t. 3. Introduce the notion of a "suspensor," an entity that holds a device in suspension. More than one suspensor may hold a device at once. A device stays suspended as long as at least one suspensor holds it. A device resumes when the last suspensor releases it. Currently, the kernel defines three suspensors, 3a the system-suspensor: for system suspension, initiated by 'sysctl -w machdep.sleep_state=3', by lid closure, by power-button press, et cetera, 3b the drvctl-suspensor: for device suspension by /dev/drvctl ioctl, e.g., drvctl -S sip0. 3c the system self-suspensor: for device drivers that suspend themselves and their children. Several drivers for network interfaces put the network device to sleep while it is not administratively up, that is, after the kernel calls if_stop(, 1). The self-suspensor should not be used directly. See the description of suspensor delegates, below. A suspensor can have one or more "delegates". A suspensor can release devices that its delegates hold suspended. Right now, only the system self-suspensor has delegates. For each device that a self-suspending driver attaches, it creates the device's self-suspensor, a delegate of the system self-suspensor. Suspensors stop a system-wide suspend/resume cycle from waking devices that the operator put to sleep with drvctl before the cycle. They also help self-suspension to work more simply, safely, and in accord with expectations. 4. Add the notion of device activation level, devact_level_t, and a routine for checking the current activation level, device_activation(). Current activation levels are DEVACT_LEVEL_BUS, DEVACT_LEVEL_DRIVER, and DEVACT_LEVEL_CLASS, which respectively indicate that the device's bus is active, that the bus and device are active, and that the bus, device, and the functions of the device's class (network, audio) are active. Suspend/resume calls can be qualified with a devact_level_t. The power-management framework treats a devact_level_t that qualifies a device suspension as the device's current activation level; it only runs hooks to reduce the activation level from the presumed current level to the fully suspended state. The framework treats a devact_level_t qualifying device resumption as the target activation level; it only runs hooks to raise the activation level to the target. 5. Use pmf_qual_t, devact_level_t, and self-suspensors in several drivers. 6. Temporarily add an unused power-management workqueue that I will remove or replace, soon.
2009-09-16 20:34:49 +04:00
if (!pmf_device_suspend(sc->sc_dev, PMF_Q_SELF))
pmf_device_resume(dv, PMF_Q_SELF);
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
}
static void
audio_activity(device_t dv, devactive_t type)
{
struct audio_softc *sc = device_private(dv);
if (type != DVA_SYSTEM)
return;
callout_schedule(&sc->sc_idle_counter, audio_idle_timeout * hz);
Don't count poll(2) and select(2) as device activity. Don't activate the device on write(2) if it is suspended and was not idle.
2007-12-13 17:40:36 +03:00
sc->sc_idle = false;
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
if (!device_is_active(dv)) {
/* XXX joerg How to deal with a failing resume... */
In pmf(9), improve the implementation of device self-suspension and make suspension by self, by drvctl(8), and by ACPI system sleep play nice together. Start solidifying some temporary API changes. 1. Extract a new header file, <sys/device_if.h>, from <sys/device.h> and #include it from <sys/pmf.h> instead of <sys/device.h> to break the circular dependency between <sys/device.h> and <sys/pmf.h>. 2. Introduce pmf_qual_t, an aggregate of qualifications on a PMF suspend/resume call. Start to replace instances of PMF_FN_PROTO, PMF_FN_ARGS, et cetera, with a pmf_qual_t. 3. Introduce the notion of a "suspensor," an entity that holds a device in suspension. More than one suspensor may hold a device at once. A device stays suspended as long as at least one suspensor holds it. A device resumes when the last suspensor releases it. Currently, the kernel defines three suspensors, 3a the system-suspensor: for system suspension, initiated by 'sysctl -w machdep.sleep_state=3', by lid closure, by power-button press, et cetera, 3b the drvctl-suspensor: for device suspension by /dev/drvctl ioctl, e.g., drvctl -S sip0. 3c the system self-suspensor: for device drivers that suspend themselves and their children. Several drivers for network interfaces put the network device to sleep while it is not administratively up, that is, after the kernel calls if_stop(, 1). The self-suspensor should not be used directly. See the description of suspensor delegates, below. A suspensor can have one or more "delegates". A suspensor can release devices that its delegates hold suspended. Right now, only the system self-suspensor has delegates. For each device that a self-suspending driver attaches, it creates the device's self-suspensor, a delegate of the system self-suspensor. Suspensors stop a system-wide suspend/resume cycle from waking devices that the operator put to sleep with drvctl before the cycle. They also help self-suspension to work more simply, safely, and in accord with expectations. 4. Add the notion of device activation level, devact_level_t, and a routine for checking the current activation level, device_activation(). Current activation levels are DEVACT_LEVEL_BUS, DEVACT_LEVEL_DRIVER, and DEVACT_LEVEL_CLASS, which respectively indicate that the device's bus is active, that the bus and device are active, and that the bus, device, and the functions of the device's class (network, audio) are active. Suspend/resume calls can be qualified with a devact_level_t. The power-management framework treats a devact_level_t that qualifies a device suspension as the device's current activation level; it only runs hooks to reduce the activation level from the presumed current level to the fully suspended state. The framework treats a devact_level_t qualifying device resumption as the target activation level; it only runs hooks to raise the activation level to the target. 5. Use pmf_qual_t, devact_level_t, and self-suspensors in several drivers. 6. Temporarily add an unused power-management workqueue that I will remove or replace, soon.
2009-09-16 20:34:49 +04:00
pmf_device_resume(sc->sc_dev, PMF_Q_SELF);
pmf_device_resume(dv, PMF_Q_SELF);
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
}
}
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#endif
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
static bool
A pointer typedef entails trading too much flexibility to declare const and non-const types, and the kernel uses both const and non-const PMF qualifiers and device suspensors, so change the pmf_qual_t and device_suspensor_t typedefs from "pointers to const" to non-pointer, non-const types.
2010-02-25 01:37:54 +03:00
audio_suspend(device_t dv, const pmf_qual_t *qual)
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
{
struct audio_softc *sc = device_private(dv);
const struct audio_hw_if *hwp = sc->hw_if;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
audio_mixer_capture(sc);
Correct copy/paste/modify issue in audio_suspend().
2011-11-25 07:13:06 +04:00
mutex_enter(sc->sc_intr_lock);
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
if (sc->sc_pbus == true)
hwp->halt_output(sc->hw_hdl);
if (sc->sc_rbus == true)
hwp->halt_input(sc->hw_hdl);
Correct copy/paste/modify issue in audio_suspend().
2011-11-25 07:13:06 +04:00
mutex_exit(sc->sc_intr_lock);
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#ifdef AUDIO_PM_IDLE
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
callout_halt(&sc->sc_idle_counter, sc->sc_lock);
Disable audio idle timer. It was a fun experiment, but causes more issues than it was worth.
2008-02-13 21:31:51 +03:00
#endif
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
return true;
}
static bool
A pointer typedef entails trading too much flexibility to declare const and non-const types, and the kernel uses both const and non-const PMF qualifiers and device suspensors, so change the pmf_qual_t and device_suspensor_t typedefs from "pointers to const" to non-pointer, non-const types.
2010-02-25 01:37:54 +03:00
audio_resume(device_t dv, const pmf_qual_t *qual)
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
{
struct audio_softc *sc = device_private(dv);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
When resuming the audio device, call audiosetinfo with the last known good parameters. Fixes an issue with suspending the audio device while playback is paused.
2007-12-13 17:02:53 +03:00
if (sc->sc_lastinfovalid)
audiosetinfo(sc, &sc->sc_lastinfo);
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
audio_mixer_restore(sc);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_intr_lock);
- Fix the conditions for POLLOUT to match those elsewhere - Refactor pause attribute handling so that it it's preserved appropriately and reset in audio_clear(). Now audio/jack-devel works with the native driver.
2008-06-03 10:21:17 +04:00
if ((sc->sc_pbus == true) && !sc->sc_pr.pause)
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
audiostartp(sc);
- Fix the conditions for POLLOUT to match those elsewhere - Refactor pause attribute handling so that it it's preserved appropriately and reset in audio_clear(). Now audio/jack-devel works with the native driver.
2008-06-03 10:21:17 +04:00
if ((sc->sc_rbus == true) && !sc->sc_rr.pause)
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
audiostartr(sc);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_intr_lock);
mutex_exit(sc->sc_lock);
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
return true;
}
static void
audio_volume_down(device_t dv)
{
struct audio_softc *sc = device_private(dv);
pmf: in audio_volume_up and audio_volume_down, honour the mixer control's delta field rather than arbitrary '32' when stepping up or down master volume.
2008-06-03 06:16:18 +04:00
mixer_devinfo_t mi;
int newgain;
u_int gain;
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
u_char balance;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
pmf: in audio_volume_up and audio_volume_down, honour the mixer control's delta field rather than arbitrary '32' when stepping up or down master volume.
2008-06-03 06:16:18 +04:00
if (sc->sc_outports.index == -1 && sc->sc_outports.master != -1) {
mi.index = sc->sc_outports.master;
mi.un.v.delta = 0;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if (sc->hw_if->query_devinfo(sc->hw_hdl, &mi) == 0) {
au_get_gain(sc, &sc->sc_outports, &gain, &balance);
newgain = gain - mi.un.v.delta;
if (newgain < AUDIO_MIN_GAIN)
newgain = AUDIO_MIN_GAIN;
au_set_gain(sc, &sc->sc_outports, newgain, balance);
}
pmf: in audio_volume_up and audio_volume_down, honour the mixer control's delta field rather than arbitrary '32' when stepping up or down master volume.
2008-06-03 06:16:18 +04:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
}
static void
audio_volume_up(device_t dv)
{
struct audio_softc *sc = device_private(dv);
pmf: in audio_volume_up and audio_volume_down, honour the mixer control's delta field rather than arbitrary '32' when stepping up or down master volume.
2008-06-03 06:16:18 +04:00
mixer_devinfo_t mi;
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
u_int gain, newgain;
u_char balance;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
pmf: in audio_volume_up and audio_volume_down, honour the mixer control's delta field rather than arbitrary '32' when stepping up or down master volume.
2008-06-03 06:16:18 +04:00
if (sc->sc_outports.index == -1 && sc->sc_outports.master != -1) {
mi.index = sc->sc_outports.master;
mi.un.v.delta = 0;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
if (sc->hw_if->query_devinfo(sc->hw_hdl, &mi) == 0) {
au_get_gain(sc, &sc->sc_outports, &gain, &balance);
newgain = gain + mi.un.v.delta;
if (newgain > AUDIO_MAX_GAIN)
newgain = AUDIO_MAX_GAIN;
au_set_gain(sc, &sc->sc_outports, newgain, balance);
}
pmf: in audio_volume_up and audio_volume_down, honour the mixer control's delta field rather than arbitrary '32' when stepping up or down master volume.
2008-06-03 06:16:18 +04:00
}
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
}
static void
audio_volume_toggle(device_t dv)
{
struct audio_softc *sc = device_private(dv);
u_int gain, newgain;
u_char balance;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_enter(sc->sc_lock);
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
au_get_gain(sc, &sc->sc_outports, &gain, &balance);
if (gain != 0) {
sc->sc_lastgain = gain;
newgain = 0;
Fix audio_volume_toggle. How did this ever work for me? Thinkpad T61 mutes audio automatically when the mute key is pressed.
2008-04-01 04:52:11 +04:00
} else
newgain = sc->sc_lastgain;
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
au_set_gain(sc, &sc->sc_outports, newgain, balance);
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
mutex_exit(sc->sc_lock);
Merge jmcneill-pm branch.
2007-12-09 23:27:42 +03:00
}
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
static int
audio_get_props(struct audio_softc *sc)
{
const struct audio_hw_if *hw;
int props;
Merge jmcneill-audiomp3 branch, which is derived from ad-audiomp2. From the original ad-audiomp branch notes: Add MP locking to the audio drivers. Making the audio drivers MP safe is necessary before efforts can be made to make the VM system MP safe. The are two locks per device instance, an ISR lock and a character device lock. The ISR lock replaces calls to splaudio()/splx(), and will be held across calls to device methods which were called at splaudio() before (e.g. trigger_output). The character device lock is held across calls to nearly all of the methods, excluding some only used for initialization, e.g. get_locks. Welcome to 5.99.57.
2011-11-24 03:07:28 +04:00
KASSERT(mutex_owned(sc->sc_lock));
Add support for playback- or capture-only devices by adding AUDIO_PROP_PLAYBACK and AUDIO_PROP_CAPTURE properties. From jmcneill@. Fixes PR#42050
2009-09-29 19:58:54 +04:00
hw = sc->hw_if;
props = hw->get_props(sc->hw_hdl);
/*
* if neither playback nor capture properties are reported,
* assume both are supported by the device driver
*/
if ((props & (AUDIO_PROP_PLAYBACK|AUDIO_PROP_CAPTURE)) == 0)
props |= (AUDIO_PROP_PLAYBACK | AUDIO_PROP_CAPTURE);
return props;
}
static bool
audio_can_playback(struct audio_softc *sc)
{
return audio_get_props(sc) & AUDIO_PROP_PLAYBACK ? true : false;
}
static bool
audio_can_capture(struct audio_softc *sc)
{
return audio_get_props(sc) & AUDIO_PROP_CAPTURE ? true : false;
}
Protect audio_powerhook with '#if NAUDIO > 0'. PR kern/33832.
2006-06-27 06:07:08 +04:00
#endif /* NAUDIO > 0 */