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

2494 lines
64 KiB
C
Raw Normal View History

delete extraneous verbiage.
2005-12-16 20:54:36 +03:00
/* $NetBSD: cd.c,v 1.233 2005/12/16 17:54:36 christos Exp $ */
New RCS ID's, take two. they're more aesthecially pleasant, and use 'NetBSD'
1994-06-29 10:39:25 +04:00
Assign my copyrights to TNF.
1998-08-17 04:49:01 +04:00
/*-
Clean up cdgetdisklabel. It was jumping over a return...
2005-12-15 20:56:32 +03:00
* Copyright (c) 1998, 2001, 2003, 2004, 2005 The NetBSD Foundation, Inc.
Assign my copyrights to TNF.
1998-08-17 04:49:01 +04:00
* All rights reserved.
*
* This code is derived from software contributed to The NetBSD Foundation
* by Charles M. Hannum.
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
*
* 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:
Assign my copyrights to TNF.
1998-08-17 04:49:01 +04:00
* This product includes software developed by the NetBSD
* Foundation, Inc. and its contributors.
* 4. Neither the name of The NetBSD Foundation nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
*
Assign my copyrights to TNF.
1998-08-17 04:49:01 +04:00
* 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 SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
*/
/*
* Originally written by Julian Elischer (julian@tfs.com)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
* for TRW Financial Systems for use under the MACH(2.5) operating system.
*
* TRW Financial Systems, in accordance with their agreement with Carnegie
* Mellon University, makes this software available to CMU to distribute
* or use in any manner that they see fit as long as this message is kept with
* the software. For this reason TFS also grants any other persons or
* organisations permission to use or modify this software.
*
* TFS supplies this software to be publicly redistributed
* on the understanding that TFS is not responsible for the correct
* functioning of this software in any circumstances.
*
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
* Ported to run under 386BSD by Julian Elischer (julian@tfs.com) Sept 1992
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
*/
add RCSIDs
2001-11-13 09:54:32 +03:00
#include <sys/cdefs.h>
delete extraneous verbiage.
2005-12-16 20:54:36 +03:00
__KERNEL_RCSID(0, "$NetBSD: cd.c,v 1.233 2005/12/16 17:54:36 christos Exp $");
add RCSIDs
2001-11-13 09:54:32 +03:00
o Make usage of /dev/random dependant on pseudo-device rnd # /dev/random and in-kernel generator in config files. o Add declaration to all architectures. o Clean up copyright message in rnd.c, rnd.h, and rndpool.c to include that this code is derived in part from Ted Tyso's linux code.
1997-10-13 04:46:08 +04:00
#include "rnd.h"
Canonicalize all #includes.
1993-12-17 10:56:32 +03:00
#include <sys/param.h>
#include <sys/systm.h>
On probe or open, retry the READ CAPACITY once if the drive reports NOT READY.
1994-08-12 03:51:26 +04:00
#include <sys/kernel.h>
Canonicalize all #includes.
1993-12-17 10:56:32 +03:00
#include <sys/file.h>
#include <sys/stat.h>
#include <sys/ioctl.h>
#include <sys/buf.h>
move buffer queue related stuffs from buf.h to their own header, bufq.h.
2004-10-28 11:07:35 +04:00
#include <sys/bufq.h>
Canonicalize all #includes.
1993-12-17 10:56:32 +03:00
#include <sys/uio.h>
#include <sys/malloc.h>
#include <sys/errno.h>
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
#include <sys/device.h>
Canonicalize all #includes.
1993-12-17 10:56:32 +03:00
#include <sys/disklabel.h>
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
#include <sys/disk.h>
#include <sys/cdio.h>
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
#include <sys/dvdio.h>
Allows the raw partition to be open()'ed, even when scsipi_start() fails (no media or other ...) so that we can always send ioctl's to the device.
1999-01-26 16:59:44 +03:00
#include <sys/scsiio.h>
scsi prototypes
1996-02-15 00:46:52 +03:00
#include <sys/proc.h>
Eliminate scsi_conf.h.
1996-03-31 00:44:50 +03:00
#include <sys/conf.h>
Allow to detach wdc, atapibus, wd and cd.
1999-09-23 15:04:29 +04:00
#include <sys/vnode.h>
o Make usage of /dev/random dependant on pseudo-device rnd # /dev/random and in-kernel generator in config files. o Add declaration to all architectures. o Clean up copyright message in rnd.c, rnd.h, and rndpool.c to include that this code is derived in part from Ted Tyso's linux code.
1997-10-13 04:46:08 +04:00
#if NRND > 0
Add hooks to insert timing info into the random system
1997-10-10 05:09:03 +04:00
#include <sys/rnd.h>
o Make usage of /dev/random dependant on pseudo-device rnd # /dev/random and in-kernel generator in config files. o Add declaration to all architectures. o Clean up copyright message in rnd.c, rnd.h, and rndpool.c to include that this code is derived in part from Ted Tyso's linux code.
1997-10-13 04:46:08 +04:00
#endif
Canonicalize all #includes.
1993-12-17 10:56:32 +03:00
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
#include <dev/scsipi/scsi_spc.h>
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
#include <dev/scsipi/scsipi_all.h>
#include <dev/scsipi/scsipi_cd.h>
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
#include <dev/scsipi/scsipi_disk.h> /* rw_big and start_stop come */
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
#include <dev/scsipi/scsi_all.h>
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
/* from there */
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
#include <dev/scsipi/scsi_disk.h> /* rw comes from there */
#include <dev/scsipi/scsipiconf.h>
Make the xxstart() functions reentrant again, as some drivers HBA can call scsipi_done() from their scsipi_request(). For this, add a struct scsipi_xfer * argument to scsipi_command(). If not NULL scsipi_command() will use this to enqueue this xfer, otherwise it'll try to allocate a new one. This scsipi_xfer has to be allocated and initialised by scsipi_make_xs() or equivalent. In xxstart(), allocate a scsipi_xfer using scsipi_make_xs(), and if not NULL, dequeue the buffer before calling scsipi_command(). This makes sure that scsipi_command() will not fail, and also makes sure that xxstart() won't be called again between the BUFQ_PEEK() and BUFQ_GET(). Fix "dequeued wrong buf" panics reported by Juergen Hannken-Illjes in private mail and Andreas Wrede on current-users@. Thanks to Jason Thorpe and Chuck Silver for review, and Andreas Wrede for testing the patch.
2004-09-09 23:35:30 +04:00
#include <dev/scsipi/scsipi_base.h>
add support for an ATAPI attachment for 'sd'. fix 'cd' driver's NCD_SCSI bogosity (was using testing wrong macro!) clean up in various ways: * make common atapi_mode_{sense,select}() functions. * put ATAPI data structures in more sensible headers, split up by device type. * include headers a bit more carefully. * pass flags to attachment-specific cd functions, and use them. * get rid of SCSI bits in scsipi_base.h's scsipi_make_xs(), move them into the correct place in scsi_base.c. * fix minor typo in struct name in scsipiconf.h (which was apparently never used except in a #define later in the same file). * use __attribute__ to force 4-byte alignment for xs command store, so that architectures trying to bus_space_write_multi_N() (where N > 1) that data to a controller won't lose. * clean up a few comments in typos, and make a few #defines easier to understand/maintain. * rename cd_link.h to cdvar.h (via repository copy). This is exactly what a 'var' file is supposed to be.
1998-01-15 05:21:27 +03:00
#include <dev/scsipi/cdvar.h>
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
move arch dependence to archs. clean up a bit. deal with weird MAXPARTITONS.
1994-10-14 21:26:22 +03:00
#define CDUNIT(z) DISKUNIT(z)
#define CDPART(z) DISKPART(z)
Allow to detach wdc, atapibus, wd and cd.
1999-09-23 15:04:29 +04:00
#define CDMINOR(unit, part) DISKMINOR(unit, part)
move arch dependence to archs. clean up a bit. deal with weird MAXPARTITONS.
1994-10-14 21:26:22 +03:00
#define MAKECDDEV(maj, unit, part) MAKEDISKDEV(maj, unit, part)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
#define MAXTRACK 99
#define CD_BLOCK_OFFSET 150
#define CD_FRAMES 75
#define CD_SECS 60
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
Second stage of read TOC modifications. Introducing the various forms of TOC reading. Note that the external interfaces haven't changed. Only the formatted toc is requestable. The read msinfo command hasn't been changed in this patch though it could become more clever using the CD_TOC_MSINFO format.
2005-12-11 00:17:21 +03:00
#define CD_TOC_FORM 0 /* formatted TOC, exposed to userland */
#define CD_TOC_MSINFO 1 /* multi-session info */
#define CD_TOC_RAW 2 /* raw TOC as on disc, unprocessed */
#define CD_TOC_PMA 3 /* PMA, used as intermediate (rare use) */
#define CD_TOC_ATIP 4 /* pressed space of recordable */
#define CD_TOC_CDTEXT 5 /* special CD-TEXT, rarely used */
struct cd_formatted_toc {
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
struct ioc_toc_header header;
struct cd_toc_entry entries[MAXTRACK+1]; /* One extra for the */
/* leadout */
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
};
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static void cdstart(struct scsipi_periph *);
Improve handling of memory shortage, to fix problems like: sd3(mpt0:0:1:0): unable to allocate scsipi_xfer sd3: not queued, error 12 Havard Eidnes's analysis of this problem is that the scsipi_xfer pool is competing for resources with other pools, including the the inode and vnode pools which can grow quite large. *_scsipi_cmd(): don't biodone the buffer if scsipi_make_xs() fails, let the caller deal with the problem start function of block devices drivers: dequeue the buffer after the scsipi_command() call. If scsipi_command() fails with ENOMEM don't dequeue the buffer, and schedule a callout to call the start function after some delay. scsipi_init(): prime the scsipi_xfer_pool with one page. This ensure that there is always some scsipi_xfer to play with. If scsipi_command() fails because of pool_get(), we're sure there will be resources available later, when the pending commands have completed. Reviewed by Jason Thorpe and Havard Eidnes. Todo: remove the "unable to allocate scsipi_xfer" and "not queued, error %d" printfs, but I choose to keep them for now, to help make sure the code does what it should.
2004-08-28 00:37:28 +04:00
static void cdrestart(void *);
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static void cdminphys(struct buf *);
static void cdgetdefaultlabel(struct cd_softc *, struct disklabel *);
static void cdgetdisklabel(struct cd_softc *);
Do not manipulate xs->bp in "generic" code -- do it only in the psw_done routine. As part of this, pass down our pre-parsed error code -- though this interface will probably change later to accomodate better error handling.
2004-09-18 03:10:50 +04:00
static void cddone(struct scsipi_xfer *, int);
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static void cdbounce(struct buf *);
static int cd_interpret_sense(struct scsipi_xfer *);
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
static u_long cd_size(struct cd_softc *, int);
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int cd_play(struct cd_softc *, int, int);
static int cd_play_tracks(struct cd_softc *, int, int, int, int);
static int cd_play_msf(struct cd_softc *, int, int, int, int, int, int);
static int cd_pause(struct cd_softc *, int);
static int cd_reset(struct cd_softc *);
static int cd_read_subchannel(struct cd_softc *, int, int, int,
struct cd_sub_channel_info *, int, int);
Second stage of read TOC modifications. Introducing the various forms of TOC reading. Note that the external interfaces haven't changed. Only the formatted toc is requestable. The read msinfo command hasn't been changed in this patch though it could become more clever using the CD_TOC_MSINFO format.
2005-12-11 00:17:21 +03:00
static int cd_read_toc(struct cd_softc *, int, int, int, void *, int, int, int);
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int cd_get_parms(struct cd_softc *, int);
Second stage of read TOC modifications. Introducing the various forms of TOC reading. Note that the external interfaces haven't changed. Only the formatted toc is requestable. The read msinfo command hasn't been changed in this patch though it could become more clever using the CD_TOC_MSINFO format.
2005-12-11 00:17:21 +03:00
static int cd_load_toc(struct cd_softc *, int, struct cd_formatted_toc *, int);
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int cdreadmsaddr(struct cd_softc *, int *);
static int dvd_auth(struct cd_softc *, dvd_authinfo *);
static int dvd_read_physical(struct cd_softc *, dvd_struct *);
static int dvd_read_copyright(struct cd_softc *, dvd_struct *);
static int dvd_read_disckey(struct cd_softc *, dvd_struct *);
static int dvd_read_bca(struct cd_softc *, dvd_struct *);
static int dvd_read_manufact(struct cd_softc *, dvd_struct *);
static int dvd_read_struct(struct cd_softc *, dvd_struct *);
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
static int cd_mode_sense(struct cd_softc *, u_int8_t, void *, size_t, int,
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
int, int *);
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
static int cd_mode_select(struct cd_softc *, u_int8_t, void *, size_t,
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
int, int);
static int cd_setchan(struct cd_softc *, int, int, int, int, int);
static int cd_getvol(struct cd_softc *, struct ioc_vol *, int);
static int cd_setvol(struct cd_softc *, const struct ioc_vol *, int);
static int cd_set_pa_immed(struct cd_softc *, int);
static int cd_load_unload(struct cd_softc *, struct ioc_load_unload *);
static int cd_setblksize(struct cd_softc *);
static int cdmatch(struct device *, struct cfdata *, void *);
static void cdattach(struct device *, struct device *, void *);
static int cdactivate(struct device *, enum devact);
static int cddetach(struct device *, int);
Eliminate the separate ATAPI and SCSI attachments for "cd".
2003-09-08 05:56:33 +04:00
CFATTACH_DECL(cd, sizeof(struct cd_softc), cdmatch, cdattach, cddetach,
cdactivate);
move the 'cd_cd' declaration back up to where it was before the config changes were made. Though the nature of the declaration had to change, there wasn't a reason to change its location.
1998-01-15 22:56:03 +03:00
extern struct cfdriver cd_cd;
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static const struct scsipi_inquiry_pattern cd_patterns[] = {
Eliminate the separate ATAPI and SCSI attachments for "cd".
2003-09-08 05:56:33 +04:00
{T_CDROM, T_REMOV,
"", "", ""},
{T_WORM, T_REMOV,
"", "", ""},
#if 0
{T_CDROM, T_REMOV, /* more luns */
"PIONEER ", "CD-ROM DRM-600 ", ""},
#endif
{T_DIRECT, T_REMOV,
"NEC CD-ROM DRIVE:260", "", ""},
};
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static dev_type_open(cdopen);
static dev_type_close(cdclose);
static dev_type_read(cdread);
static dev_type_write(cdwrite);
static dev_type_ioctl(cdioctl);
static dev_type_strategy(cdstrategy);
static dev_type_dump(cddump);
static dev_type_size(cdsize);
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 bdevsw cd_bdevsw = {
cdopen, cdclose, cdstrategy, cdioctl, cddump, cdsize, D_DISK
};
const struct cdevsw cd_cdevsw = {
cdopen, cdclose, cdread, cdwrite, cdioctl,
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
nostop, notty, nopoll, nommap, nokqfilter, D_DISK
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
};
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static struct dkdriver cddkdriver = { cdstrategy };
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static const struct scsipi_periphsw cd_switch = {
Retry requests if the CD-ROM drive returns the sense code "Device in Process of Becoming Ready".
2001-01-20 01:47:46 +03:00
cd_interpret_sense, /* use our error handler first */
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
cdstart, /* we have a queue, which is started by this */
NULL, /* we do not have an async handler */
New generic disk framework. Highlights: - New metrics handling. Metrics are now kept in the new `struct disk'. Busy time is now stored as a timeval, and transfer count in bytes. - Storage for disklabels is now dynamically allocated, so that the size of the disk structure is not machine-dependent. - Several new functions for attaching and detaching disks, and handling metrics calculation. Old-style instrumentation is still supported in drivers that did it before. However, old-style instrumentation is being deprecated, and will go away once the userland utilities are updated for the new framework. For usage and architectural details, see the forthcoming disk(9) manual page.
1996-01-08 01:01:38 +03:00
cddone, /* deal with stats at interrupt time */
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
};
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
support for making dev->id_alive be set, this is for iostat to find disk devices. wee bit of a kludge. sub-device attach() routines must now return 1 for successful attach(), 0 otherwise. Other bsd's do this too..
1993-05-04 12:27:29 +04:00
/*
* The routine called by the low level scsi routine when it discovers
* A device suitable for this driver
*/
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cdmatch(struct device *parent, struct cfdata *match, void *aux)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
{
Eliminate the separate ATAPI and SCSI attachments for "cd".
2003-09-08 05:56:33 +04:00
struct scsipibus_attach_args *sa = aux;
int priority;
(void)scsipi_inqmatch(&sa->sa_inqbuf,
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
cd_patterns, sizeof(cd_patterns) / sizeof(cd_patterns[0]),
Eliminate the separate ATAPI and SCSI attachments for "cd".
2003-09-08 05:56:33 +04:00
sizeof(cd_patterns[0]), &priority);
return (priority);
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static void
cdattach(struct device *parent, struct device *self, void *aux)
Eliminate the separate ATAPI and SCSI attachments for "cd".
2003-09-08 05:56:33 +04:00
{
struct cd_softc *cd = (void *)self;
struct scsipibus_attach_args *sa = aux;
struct scsipi_periph *periph = sa->sa_periph;
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
SC_DEBUG(periph, SCSIPI_DB2, ("cdattach: "));
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
Do not manipulate xs->bp in "generic" code -- do it only in the psw_done routine. As part of this, pass down our pre-parsed error code -- though this interface will probably change later to accomodate better error handling.
2004-09-18 03:10:50 +04:00
lockinit(&cd->sc_lock, PRIBIO | PCATCH, "cdlock", 0, 0);
Eliminate the separate ATAPI and SCSI attachments for "cd".
2003-09-08 05:56:33 +04:00
if (scsipi_periph_bustype(sa->sa_periph) == SCSIPI_BUSTYPE_SCSI &&
periph->periph_version == 0)
cd->flags |= CDF_ANCIENT;
- change the way to specify a bufq strategy. (by string rather than by number) - rather than embedding bufq_state in driver softc, have a pointer to the former. - move bufq related functions from kern/subr_disk.c to kern/subr_bufq.c. - rename method to strategy for consistency. - move some definitions which don't need to be exposed to the rest of kernel from sys/bufq.h to sys/bufq_impl.h. (is it better to move it to kern/ or somewhere?) - fix some obvious breakage in dev/qbus/ts.c. (not tested)
2005-10-15 21:29:10 +04:00
bufq_alloc(&cd->buf_queue, "disksort", BUFQ_SORT_RAWBLOCK);
Update for sys/buf.h/disksort_*() changes.
2000-01-22 02:39:55 +03:00
Improve handling of memory shortage, to fix problems like: sd3(mpt0:0:1:0): unable to allocate scsipi_xfer sd3: not queued, error 12 Havard Eidnes's analysis of this problem is that the scsipi_xfer pool is competing for resources with other pools, including the the inode and vnode pools which can grow quite large. *_scsipi_cmd(): don't biodone the buffer if scsipi_make_xs() fails, let the caller deal with the problem start function of block devices drivers: dequeue the buffer after the scsipi_command() call. If scsipi_command() fails with ENOMEM don't dequeue the buffer, and schedule a callout to call the start function after some delay. scsipi_init(): prime the scsipi_xfer_pool with one page. This ensure that there is always some scsipi_xfer to play with. If scsipi_command() fails because of pool_get(), we're sure there will be resources available later, when the pending commands have completed. Reviewed by Jason Thorpe and Havard Eidnes. Todo: remove the "unable to allocate scsipi_xfer" and "not queued, error %d" printfs, but I choose to keep them for now, to help make sure the code does what it should.
2004-08-28 00:37:28 +04:00
callout_init(&cd->sc_callout);
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* Store information needed to contact our base driver
*/
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
cd->sc_periph = periph;
periph->periph_dev = &cd->sc_dev;
periph->periph_switch = &cd_switch;
/*
* Increase our openings to the maximum-per-periph
* supported by the adapter. This will either be
* clamped down or grown by the adapter if necessary.
*/
periph->periph_openings =
SCSIPI_CHAN_MAX_PERIPH(periph->periph_channel);
periph->periph_flags |= PERIPH_GROW_OPENINGS;
new scsi subsystem. changes also in config/mkioconf.c i386/isa/wd.c, fd.c, and all scsi drivers.
1993-04-12 12:19:28 +04:00
New generic disk framework. Highlights: - New metrics handling. Metrics are now kept in the new `struct disk'. Busy time is now stored as a timeval, and transfer count in bytes. - Storage for disklabels is now dynamically allocated, so that the size of the disk structure is not machine-dependent. - Several new functions for attaching and detaching disks, and handling metrics calculation. Old-style instrumentation is still supported in drivers that did it before. However, old-style instrumentation is being deprecated, and will go away once the userland utilities are updated for the new framework. For usage and architectural details, see the forthcoming disk(9) manual page.
1996-01-08 01:01:38 +03:00
/*
* Initialize and attach the disk structure.
*/
scsi prototypes
1996-02-15 00:46:52 +03:00
cd->sc_dk.dk_driver = &cddkdriver;
New generic disk framework. Highlights: - New metrics handling. Metrics are now kept in the new `struct disk'. Busy time is now stored as a timeval, and transfer count in bytes. - Storage for disklabels is now dynamically allocated, so that the size of the disk structure is not machine-dependent. - Several new functions for attaching and detaching disks, and handling metrics calculation. Old-style instrumentation is still supported in drivers that did it before. However, old-style instrumentation is being deprecated, and will go away once the userland utilities are updated for the new framework. For usage and architectural details, see the forthcoming disk(9) manual page.
1996-01-08 01:01:38 +03:00
cd->sc_dk.dk_name = cd->sc_dev.dv_xname;
disk_attach(&cd->sc_dk);
revert previous kprintf change
1996-10-13 03:23:13 +04:00
printf("\n");
Add hooks to insert timing info into the random system
1997-10-10 05:09:03 +04:00
o Make usage of /dev/random dependant on pseudo-device rnd # /dev/random and in-kernel generator in config files. o Add declaration to all architectures. o Clean up copyright message in rnd.c, rnd.h, and rndpool.c to include that this code is derived in part from Ted Tyso's linux code.
1997-10-13 04:46:08 +04:00
#if NRND > 0
Update to slightly altered rnd_attach_source() api
1999-02-28 20:08:05 +03:00
rnd_attach_source(&cd->rnd_source, cd->sc_dev.dv_xname,
RND_TYPE_DISK, 0);
o Make usage of /dev/random dependant on pseudo-device rnd # /dev/random and in-kernel generator in config files. o Add declaration to all architectures. o Clean up copyright message in rnd.c, rnd.h, and rndpool.c to include that this code is derived in part from Ted Tyso's linux code.
1997-10-13 04:46:08 +04:00
#endif
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cdactivate(struct device *self, enum devact act)
Allow to detach wdc, atapibus, wd and cd.
1999-09-23 15:04:29 +04:00
{
int rv = 0;
switch (act) {
case DVACT_ACTIVATE:
rv = EOPNOTSUPP;
break;
case DVACT_DEACTIVATE:
/*
* Nothing to do; we key off the device's DVF_ACTIVE.
*/
break;
}
return (rv);
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cddetach(struct device *self, int flags)
Allow to detach wdc, atapibus, wd and cd.
1999-09-23 15:04:29 +04:00
{
struct cd_softc *cd = (struct cd_softc *) self;
Since __HAVE_OLD_DISKLABEL, DISKMINOR(u, p) isn't continuous over "p" anymore. So we can't pass a range to vdevgone() but have to loop over the partitions explicitely. (otherwise the next device unit will be killed)
2001-05-06 22:30:56 +04:00
int s, bmaj, cmaj, i, mn;
Allow to detach wdc, atapibus, wd and cd.
1999-09-23 15:04:29 +04:00
/* locate the major number */
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
bmaj = bdevsw_lookup_major(&cd_bdevsw);
cmaj = cdevsw_lookup_major(&cd_cdevsw);
Allow to detach wdc, atapibus, wd and cd.
1999-09-23 15:04:29 +04:00
Do not manipulate xs->bp in "generic" code -- do it only in the psw_done routine. As part of this, pass down our pre-parsed error code -- though this interface will probably change later to accomodate better error handling.
2004-09-18 03:10:50 +04:00
/* Nuke the vnodes for any open instances */
for (i = 0; i < MAXPARTITIONS; i++) {
mn = CDMINOR(self->dv_unit, i);
vdevgone(bmaj, mn, mn, VBLK);
vdevgone(cmaj, mn, mn, VCHR);
}
Improve handling of memory shortage, to fix problems like: sd3(mpt0:0:1:0): unable to allocate scsipi_xfer sd3: not queued, error 12 Havard Eidnes's analysis of this problem is that the scsipi_xfer pool is competing for resources with other pools, including the the inode and vnode pools which can grow quite large. *_scsipi_cmd(): don't biodone the buffer if scsipi_make_xs() fails, let the caller deal with the problem start function of block devices drivers: dequeue the buffer after the scsipi_command() call. If scsipi_command() fails with ENOMEM don't dequeue the buffer, and schedule a callout to call the start function after some delay. scsipi_init(): prime the scsipi_xfer_pool with one page. This ensure that there is always some scsipi_xfer to play with. If scsipi_command() fails because of pool_get(), we're sure there will be resources available later, when the pending commands have completed. Reviewed by Jason Thorpe and Havard Eidnes. Todo: remove the "unable to allocate scsipi_xfer" and "not queued, error %d" printfs, but I choose to keep them for now, to help make sure the code does what it should.
2004-08-28 00:37:28 +04:00
/* kill any pending restart */
callout_stop(&cd->sc_callout);
Allow to detach wdc, atapibus, wd and cd.
1999-09-23 15:04:29 +04:00
s = splbio();
/* Kill off any queued buffers. */
- change the way to specify a bufq strategy. (by string rather than by number) - rather than embedding bufq_state in driver softc, have a pointer to the former. - move bufq related functions from kern/subr_disk.c to kern/subr_bufq.c. - rename method to strategy for consistency. - move some definitions which don't need to be exposed to the rest of kernel from sys/bufq.h to sys/bufq_impl.h. (is it better to move it to kern/ or somewhere?) - fix some obvious breakage in dev/qbus/ts.c. (not tested)
2005-10-15 21:29:10 +04:00
bufq_drain(cd->buf_queue);
Allow to detach wdc, atapibus, wd and cd.
1999-09-23 15:04:29 +04:00
- change the way to specify a bufq strategy. (by string rather than by number) - rather than embedding bufq_state in driver softc, have a pointer to the former. - move bufq related functions from kern/subr_disk.c to kern/subr_bufq.c. - rename method to strategy for consistency. - move some definitions which don't need to be exposed to the rest of kernel from sys/bufq.h to sys/bufq_impl.h. (is it better to move it to kern/ or somewhere?) - fix some obvious breakage in dev/qbus/ts.c. (not tested)
2005-10-15 21:29:10 +04:00
bufq_free(cd->buf_queue);
Convert to new device buffer queue interface.
2002-07-22 18:59:43 +04:00
Allow to detach wdc, atapibus, wd and cd.
1999-09-23 15:04:29 +04:00
/* Kill off any pending commands. */
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
scsipi_kill_pending(cd->sc_periph);
Allow to detach wdc, atapibus, wd and cd.
1999-09-23 15:04:29 +04:00
splx(s);
Do not manipulate xs->bp in "generic" code -- do it only in the psw_done routine. As part of this, pass down our pre-parsed error code -- though this interface will probably change later to accomodate better error handling.
2004-09-18 03:10:50 +04:00
lockmgr(&cd->sc_lock, LK_DRAIN, 0);
Allow to detach wdc, atapibus, wd and cd.
1999-09-23 15:04:29 +04:00
/* Detach from the disk list. */
disk_detach(&cd->sc_dk);
#if 0
/* Get rid of the shutdown hook. */
if (cd->sc_sdhook != NULL)
shutdownhook_disestablish(cd->sc_sdhook);
#endif
#if NRND > 0
/* Unhook the entropy source. */
rnd_detach_source(&cd->rnd_source);
#endif
return (0);
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* open the device. Make sure the partition info is a up-to-date as can be.
*/
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
static int
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
cdopen(dev_t dev, int flag, int fmt, struct lwp *l)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
{
Numerous changes. Many bugs fixed, better autoconfig, a few new features.
1994-12-28 22:42:47 +03:00
struct cd_softc *cd;
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
struct scsipi_periph *periph;
struct scsipi_adapter *adapt;
Revamp the locking mechanism slightly.
1995-03-30 03:04:39 +04:00
int unit, part;
int error;
Always allow open() on the raw partition to succeed, even if any of the SCSI command in cdopen() fails. Fix PR kern/30288 by Pavel Cahyna.
2005-07-06 18:28:39 +04:00
int rawpart;
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
unit = CDUNIT(dev);
New device attachment scheme: - split softc size and match/attach out from cfdriver into a new struct cfattach. - new "attach" directive for files.*. May specify the name of the cfattach structure, so that devices may be easily attached to parents with different autoconfiguration semantics.
1996-03-17 03:57:14 +03:00
if (unit >= cd_cd.cd_ndevs)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (ENXIO);
New device attachment scheme: - split softc size and match/attach out from cfdriver into a new struct cfattach. - new "attach" directive for files.*. May specify the name of the cfattach structure, so that devices may be easily attached to parents with different autoconfiguration semantics.
1996-03-17 03:57:14 +03:00
cd = cd_cd.cd_devs[unit];
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
if (cd == NULL)
return (ENXIO);
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
periph = cd->sc_periph;
adapt = periph->periph_channel->chan_adapter;
Allows the raw partition to be open()'ed, even when scsipi_start() fails (no media or other ...) so that we can always send ioctl's to the device.
1999-01-26 16:59:44 +03:00
part = CDPART(dev);
new scsi subsystem. changes also in config/mkioconf.c i386/isa/wd.c, fd.c, and all scsi drivers.
1993-04-12 12:19:28 +04:00
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
SC_DEBUG(periph, SCSIPI_DB1,
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
("cdopen: dev=0x%x (unit %d (of %d), partition %d)\n", dev, unit,
Cleanup the rest of the SCSIDEBUG printfs. From Bernd Ernesti.
1996-05-05 23:52:50 +04:00
cd_cd.cd_ndevs, CDPART(dev)));
new scsi subsystem. changes also in config/mkioconf.c i386/isa/wd.c, fd.c, and all scsi drivers.
1993-04-12 12:19:28 +04:00
Add adapter reference counting for SCSI and ATAPI devices.
1998-11-20 03:35:39 +03:00
/*
* If this is the first open of this device, add a reference
* to the adapter.
*/
if (cd->sc_dk.dk_openmask == 0 &&
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
(error = scsipi_adapter_addref(adapt)) != 0)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (error);
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
Do not manipulate xs->bp in "generic" code -- do it only in the psw_done routine. As part of this, pass down our pre-parsed error code -- though this interface will probably change later to accomodate better error handling.
2004-09-18 03:10:50 +04:00
if ((error = lockmgr(&cd->sc_lock, LK_EXCLUSIVE, NULL)) != 0)
Add adapter reference counting for SCSI and ATAPI devices.
1998-11-20 03:35:39 +03:00
goto bad4;
Always allow open() on the raw partition to succeed, even if any of the SCSI command in cdopen() fails. Fix PR kern/30288 by Pavel Cahyna.
2005-07-06 18:28:39 +04:00
rawpart = (part == RAW_PART && fmt == S_IFCHR);
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
if ((periph->periph_flags & PERIPH_OPEN) != 0) {
Add still more locking.
1994-11-22 06:23:49 +03:00
/*
* If any partition is open, but the disk has been invalidated,
* disallow further opens.
*/
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
if ((periph->periph_flags & PERIPH_MEDIA_LOADED) == 0 &&
Always allow open() on the raw partition to succeed, even if any of the SCSI command in cdopen() fails. Fix PR kern/30288 by Pavel Cahyna.
2005-07-06 18:28:39 +04:00
!rawpart) {
Add a missing unlock.
1995-04-01 14:29:41 +04:00
error = EIO;
goto bad3;
}
Add still more locking.
1994-11-22 06:23:49 +03:00
} else {
Make sure the "raw partition" can always be opened again.
2003-09-13 18:44:50 +04:00
int silent;
Always allow open() on the raw partition to succeed, even if any of the SCSI command in cdopen() fails. Fix PR kern/30288 by Pavel Cahyna.
2005-07-06 18:28:39 +04:00
if (rawpart)
Make sure the "raw partition" can always be opened again.
2003-09-13 18:44:50 +04:00
silent = XS_CTL_SILENT;
else
silent = 0;
Redo the `wait for spin up' code. The Mach 3 method clearly doesn't work. Also, rearrange the first open sequences a bit; SDEV_OPEN is no longer magic.
1995-01-26 14:56:51 +03:00
/* Check that it is still responding and ok. */
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
error = scsipi_test_unit_ready(periph,
Cleanup the scsipi_xfer flags: - `flags' is now gone, replaced with `xs_control' and `xs_status'. - Massive cleanup of the control flags. Now we explicitly say that a job is to complete asynchronously, rather than relying on side-effects, and use a new flag to now that device discovery is being performed. - Do SCSI device discovery interrupt-driven.
1999-10-01 02:57:52 +04:00
XS_CTL_IGNORE_ILLEGAL_REQUEST | XS_CTL_IGNORE_MEDIA_CHANGE |
Don't be silent if there is no media present.
2003-09-13 19:49:04 +04:00
silent);
Add some missing locking, and some general cleanup.
1994-11-21 01:36:43 +03:00
Do a START UNIT only if the TEST UNIT READY reports that the device is not ready. This avoids gratuitously starting the motor on floppy and CD-ROM drives, and eliminates the need for the audio playing test in cdopen(). Therefore, also remove PQUIRK_NOSTARTUNIT.
2003-09-08 22:51:33 +04:00
/*
* Start the pack spinning if necessary. Always allow the
* raw parition to be opened, for raw IOCTLs. Data transfers
* will check for SDEV_MEDIA_LOADED.
*/
In the test for whether to start the unit, I used the wrong error code. ENODEV is only returned when we get "medium not present," which we can fail immediately on. All other "not ready" cases return EIO.
2003-09-09 06:43:34 +04:00
if (error == EIO) {
Make sure the "raw partition" can always be opened again.
2003-09-13 18:44:50 +04:00
int error2;
Do a START UNIT only if the TEST UNIT READY reports that the device is not ready. This avoids gratuitously starting the motor on floppy and CD-ROM drives, and eliminates the need for the audio playing test in cdopen(). Therefore, also remove PQUIRK_NOSTARTUNIT.
2003-09-08 22:51:33 +04:00
error2 = scsipi_start(periph, SSS_START, silent);
switch (error2) {
case 0:
error = 0;
break;
In the test for whether to start the unit, I used the wrong error code. ENODEV is only returned when we get "medium not present," which we can fail immediately on. All other "not ready" cases return EIO.
2003-09-09 06:43:34 +04:00
case EIO:
Do a START UNIT only if the TEST UNIT READY reports that the device is not ready. This avoids gratuitously starting the motor on floppy and CD-ROM drives, and eliminates the need for the audio playing test in cdopen(). Therefore, also remove PQUIRK_NOSTARTUNIT.
2003-09-08 22:51:33 +04:00
case EINVAL:
break;
default:
error = error2;
break;
Two changes submitted by Sergey Svishchev <svs@ropnet.ru>: - Don't send START UNIT upon open if the drive is playing audio (PR 11768). - If we receive sense indicating that the media changed, retry (PR 8326).
2001-08-20 15:31:10 +04:00
}
Allows the raw partition to be open()'ed, even when scsipi_start() fails (no media or other ...) so that we can always send ioctl's to the device.
1999-01-26 16:59:44 +03:00
}
Make sure the "raw partition" can always be opened again.
2003-09-13 18:44:50 +04:00
if (error) {
Always allow open() on the raw partition to succeed, even if any of the SCSI command in cdopen() fails. Fix PR kern/30288 by Pavel Cahyna.
2005-07-06 18:28:39 +04:00
if (rawpart)
Make sure the "raw partition" can always be opened again.
2003-09-13 18:44:50 +04:00
goto out;
Do a START UNIT only if the TEST UNIT READY reports that the device is not ready. This avoids gratuitously starting the motor on floppy and CD-ROM drives, and eliminates the need for the audio playing test in cdopen(). Therefore, also remove PQUIRK_NOSTARTUNIT.
2003-09-08 22:51:33 +04:00
goto bad3;
Make sure the "raw partition" can always be opened again.
2003-09-13 18:44:50 +04:00
}
Some devices really do require the START before the PREVENT. This makes no sense.
1995-01-30 14:34:25 +03:00
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
periph->periph_flags |= PERIPH_OPEN;
Some devices really do require the START before the PREVENT. This makes no sense.
1995-01-30 14:34:25 +03:00
/* Lock the pack in. */
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
error = scsipi_prevent(periph, SPAMR_PREVENT_DT,
Cleanup the scsipi_xfer flags: - `flags' is now gone, replaced with `xs_control' and `xs_status'. - Massive cleanup of the control flags. Now we explicitly say that a job is to complete asynchronously, rather than relying on side-effects, and use a new flag to now that device discovery is being performed. - Do SCSI device discovery interrupt-driven.
1999-10-01 02:57:52 +04:00
XS_CTL_IGNORE_ILLEGAL_REQUEST | XS_CTL_IGNORE_MEDIA_CHANGE);
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
SC_DEBUG(periph, SCSIPI_DB1,
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
("cdopen: scsipi_prevent, error=%d\n", error));
Always allow open() on the raw partition to succeed, even if any of the SCSI command in cdopen() fails. Fix PR kern/30288 by Pavel Cahyna.
2005-07-06 18:28:39 +04:00
if (error) {
if (rawpart)
goto out;
Do the PREVENT before the START.
1995-01-23 21:17:22 +03:00
goto bad;
Always allow open() on the raw partition to succeed, even if any of the SCSI command in cdopen() fails. Fix PR kern/30288 by Pavel Cahyna.
2005-07-06 18:28:39 +04:00
}
Do the PREVENT before the START.
1995-01-23 21:17:22 +03:00
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
if ((periph->periph_flags & PERIPH_MEDIA_LOADED) == 0) {
Add some missing locking, and some general cleanup.
1994-11-21 01:36:43 +03:00
/* Load the physical device parameters. */
if (cd_get_parms(cd, 0) != 0) {
Always allow open() on the raw partition to succeed, even if any of the SCSI command in cdopen() fails. Fix PR kern/30288 by Pavel Cahyna.
2005-07-06 18:28:39 +04:00
if (rawpart)
goto out;
Add some missing locking, and some general cleanup.
1994-11-21 01:36:43 +03:00
error = ENXIO;
Always allow open() on the raw partition to succeed, even if any of the SCSI command in cdopen() fails. Fix PR kern/30288 by Pavel Cahyna.
2005-07-06 18:28:39 +04:00
goto bad;
Add some missing locking, and some general cleanup.
1994-11-21 01:36:43 +03:00
}
Always allow open() on the raw partition to succeed, even if any of the SCSI command in cdopen() fails. Fix PR kern/30288 by Pavel Cahyna.
2005-07-06 18:28:39 +04:00
periph->periph_flags |= PERIPH_MEDIA_LOADED;
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
SC_DEBUG(periph, SCSIPI_DB3, ("Params loaded "));
Add some missing locking, and some general cleanup.
1994-11-21 01:36:43 +03:00
/* Fabricate a disk label. */
cdgetdisklabel(cd);
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
SC_DEBUG(periph, SCSIPI_DB3, ("Disklabel fabricated "));
Add some missing locking, and some general cleanup.
1994-11-21 01:36:43 +03:00
}
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
Add still more locking.
1994-11-22 06:23:49 +03:00
/* Check that the partition exists. */
Similar changes to sd.c.
1994-10-20 17:03:38 +03:00
if (part != RAW_PART &&
New generic disk framework. Highlights: - New metrics handling. Metrics are now kept in the new `struct disk'. Busy time is now stored as a timeval, and transfer count in bytes. - Storage for disklabels is now dynamically allocated, so that the size of the disk structure is not machine-dependent. - Several new functions for attaching and detaching disks, and handling metrics calculation. Old-style instrumentation is still supported in drivers that did it before. However, old-style instrumentation is being deprecated, and will go away once the userland utilities are updated for the new framework. For usage and architectural details, see the forthcoming disk(9) manual page.
1996-01-08 01:01:38 +03:00
(part >= cd->sc_dk.dk_label->d_npartitions ||
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
cd->sc_dk.dk_label->d_partitions[part].p_fstype == FS_UNUSED)) {
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
error = ENXIO;
goto bad;
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
Allows the raw partition to be open()'ed, even when scsipi_start() fails (no media or other ...) so that we can always send ioctl's to the device.
1999-01-26 16:59:44 +03:00
out: /* Insure only one open at a time. */
Similar changes to sd.c.
1994-10-20 17:03:38 +03:00
switch (fmt) {
case S_IFCHR:
openpart -> openmask
1994-10-20 17:08:07 +03:00
cd->sc_dk.dk_copenmask |= (1 << part);
Similar changes to sd.c.
1994-10-20 17:03:38 +03:00
break;
case S_IFBLK:
openpart -> openmask
1994-10-20 17:08:07 +03:00
cd->sc_dk.dk_bopenmask |= (1 << part);
Similar changes to sd.c.
1994-10-20 17:03:38 +03:00
break;
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
cd->sc_dk.dk_openmask =
cd->sc_dk.dk_copenmask | cd->sc_dk.dk_bopenmask;
Similar changes to sd.c.
1994-10-20 17:03:38 +03:00
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
SC_DEBUG(periph, SCSIPI_DB3, ("open complete\n"));
Do not manipulate xs->bp in "generic" code -- do it only in the psw_done routine. As part of this, pass down our pre-parsed error code -- though this interface will probably change later to accomodate better error handling.
2004-09-18 03:10:50 +04:00
lockmgr(&cd->sc_lock, LK_RELEASE, NULL);
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (0);
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
periph->periph_flags &= ~PERIPH_MEDIA_LOADED;
Add some missing locking, and some general cleanup.
1994-11-21 01:36:43 +03:00
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
bad:
openpart -> openmask
1994-10-20 17:08:07 +03:00
if (cd->sc_dk.dk_openmask == 0) {
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
scsipi_prevent(periph, SPAMR_ALLOW,
Cleanup the scsipi_xfer flags: - `flags' is now gone, replaced with `xs_control' and `xs_status'. - Massive cleanup of the control flags. Now we explicitly say that a job is to complete asynchronously, rather than relying on side-effects, and use a new flag to now that device discovery is being performed. - Do SCSI device discovery interrupt-driven.
1999-10-01 02:57:52 +04:00
XS_CTL_IGNORE_ILLEGAL_REQUEST | XS_CTL_IGNORE_MEDIA_CHANGE);
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
periph->periph_flags &= ~PERIPH_OPEN;
Add some missing locking, and some general cleanup.
1994-11-21 01:36:43 +03:00
}
Revamp the locking mechanism slightly.
1995-03-30 03:04:39 +04:00
bad3:
Do not manipulate xs->bp in "generic" code -- do it only in the psw_done routine. As part of this, pass down our pre-parsed error code -- though this interface will probably change later to accomodate better error handling.
2004-09-18 03:10:50 +04:00
lockmgr(&cd->sc_lock, LK_RELEASE, NULL);
Add adapter reference counting for SCSI and ATAPI devices.
1998-11-20 03:35:39 +03:00
bad4:
if (cd->sc_dk.dk_openmask == 0)
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
scsipi_adapter_delref(adapt);
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (error);
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* close the device.. only called if we are the LAST
* occurence of an open device
*/
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
static int
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
cdclose(dev_t dev, int flag, int fmt, struct lwp *l)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
{
New device attachment scheme: - split softc size and match/attach out from cfdriver into a new struct cfattach. - new "attach" directive for files.*. May specify the name of the cfattach structure, so that devices may be easily attached to parents with different autoconfiguration semantics.
1996-03-17 03:57:14 +03:00
struct cd_softc *cd = cd_cd.cd_devs[CDUNIT(dev)];
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
struct scsipi_periph *periph = cd->sc_periph;
struct scsipi_adapter *adapt = periph->periph_channel->chan_adapter;
Similar changes to sd.c.
1994-10-20 17:03:38 +03:00
int part = CDPART(dev);
Revamp the locking mechanism slightly.
1995-03-30 03:04:39 +04:00
int error;
Do not manipulate xs->bp in "generic" code -- do it only in the psw_done routine. As part of this, pass down our pre-parsed error code -- though this interface will probably change later to accomodate better error handling.
2004-09-18 03:10:50 +04:00
if ((error = lockmgr(&cd->sc_lock, LK_EXCLUSIVE, NULL)) != 0)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (error);
Similar changes to sd.c.
1994-10-20 17:03:38 +03:00
switch (fmt) {
case S_IFCHR:
openpart -> openmask
1994-10-20 17:08:07 +03:00
cd->sc_dk.dk_copenmask &= ~(1 << part);
Similar changes to sd.c.
1994-10-20 17:03:38 +03:00
break;
case S_IFBLK:
openpart -> openmask
1994-10-20 17:08:07 +03:00
cd->sc_dk.dk_bopenmask &= ~(1 << part);
Similar changes to sd.c.
1994-10-20 17:03:38 +03:00
break;
}
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
cd->sc_dk.dk_openmask =
cd->sc_dk.dk_copenmask | cd->sc_dk.dk_bopenmask;
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
openpart -> openmask
1994-10-20 17:08:07 +03:00
if (cd->sc_dk.dk_openmask == 0) {
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
scsipi_wait_drain(periph);
Add still more locking.
1994-11-22 06:23:49 +03:00
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
scsipi_prevent(periph, SPAMR_ALLOW,
Cleanup the scsipi_xfer flags: - `flags' is now gone, replaced with `xs_control' and `xs_status'. - Massive cleanup of the control flags. Now we explicitly say that a job is to complete asynchronously, rather than relying on side-effects, and use a new flag to now that device discovery is being performed. - Do SCSI device discovery interrupt-driven.
1999-10-01 02:57:52 +04:00
XS_CTL_IGNORE_ILLEGAL_REQUEST | XS_CTL_IGNORE_MEDIA_CHANGE |
XS_CTL_IGNORE_NOT_READY);
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
periph->periph_flags &= ~PERIPH_OPEN;
Add adapter reference counting for SCSI and ATAPI devices.
1998-11-20 03:35:39 +03:00
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
scsipi_wait_drain(periph);
When closing, wait for pending xfers to drain before unlocking the door, and wait again before deleting the reference to the adapter.
1998-12-08 03:18:46 +03:00
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
scsipi_adapter_delref(adapt);
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
Add still more locking.
1994-11-22 06:23:49 +03:00
Do not manipulate xs->bp in "generic" code -- do it only in the psw_done routine. As part of this, pass down our pre-parsed error code -- though this interface will probably change later to accomodate better error handling.
2004-09-18 03:10:50 +04:00
lockmgr(&cd->sc_lock, LK_RELEASE, NULL);
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (0);
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* Actually translate the requested transfer into one the physical driver can
* understand. The transfer is described by a buf and will include only one
* physical transfer.
*/
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static void
cdstrategy(struct buf *bp)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
{
New device attachment scheme: - split softc size and match/attach out from cfdriver into a new struct cfattach. - new "attach" directive for files.*. May specify the name of the cfattach structure, so that devices may be easily attached to parents with different autoconfiguration semantics.
1996-03-17 03:57:14 +03:00
struct cd_softc *cd = cd_cd.cd_devs[CDUNIT(bp->b_dev)];
Fix a bug in disksort_*() which caused non-optimal ordering when multiple active partitions were on a single spindle. Add a b_rawblkno member to struct buf which contains the non-partition-relative block number to sort by.
2000-02-07 23:16:47 +03:00
struct disklabel *lp;
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
struct scsipi_periph *periph = cd->sc_periph;
Fix a bug in disksort_*() which caused non-optimal ordering when multiple active partitions were on a single spindle. Add a b_rawblkno member to struct buf which contains the non-partition-relative block number to sort by.
2000-02-07 23:16:47 +03:00
daddr_t blkno;
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
int s;
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
SC_DEBUG(cd->sc_periph, SCSIPI_DB2, ("cdstrategy "));
SC_DEBUG(cd->sc_periph, SCSIPI_DB1,
Fix debug message output args to match yamt's src/sys/sys/buf.h changes.
2004-09-26 13:00:37 +04:00
("%d bytes @ blk %" PRId64 "\n", bp->b_bcount, bp->b_blkno));
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* If the device has been made invalid, error out
* maybe the media changed
*/
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
if ((periph->periph_flags & PERIPH_MEDIA_LOADED) == 0) {
if (periph->periph_flags & PERIPH_OPEN)
Return ENODEV instead of EIO when we are trying to open a device without media in the drive. restrict "opening of empty drive" to character devices only (reading a block device returns a short read instead of ENODEV, which can lead to confusion).
1999-01-29 14:17:58 +03:00
bp->b_error = EIO;
else
bp->b_error = ENODEV;
fixed various bugs like cdattach() returning garbage.
1993-04-12 16:10:28 +04:00
goto bad;
}
Fix a bug in disksort_*() which caused non-optimal ordering when multiple active partitions were on a single spindle. Add a b_rawblkno member to struct buf which contains the non-partition-relative block number to sort by.
2000-02-07 23:16:47 +03:00
lp = cd->sc_dk.dk_label;
Allows the raw partition to be open()'ed, even when scsipi_start() fails (no media or other ...) so that we can always send ioctl's to the device.
1999-01-26 16:59:44 +03:00
/*
Abort transfer if b_blkno is negative. Closes PR kern/5553 by Johan Danielsson.
1999-02-10 15:29:50 +03:00
* The transfer must be a whole number of blocks, offset must not
* be negative.
Allows the raw partition to be open()'ed, even when scsipi_start() fails (no media or other ...) so that we can always send ioctl's to the device.
1999-01-26 16:59:44 +03:00
*/
Fix a bug in disksort_*() which caused non-optimal ordering when multiple active partitions were on a single spindle. Add a b_rawblkno member to struct buf which contains the non-partition-relative block number to sort by.
2000-02-07 23:16:47 +03:00
if ((bp->b_bcount % lp->d_secsize) != 0 ||
Abort transfer if b_blkno is negative. Closes PR kern/5553 by Johan Danielsson.
1999-02-10 15:29:50 +03:00
bp->b_blkno < 0 ) {
Allows the raw partition to be open()'ed, even when scsipi_start() fails (no media or other ...) so that we can always send ioctl's to the device.
1999-01-26 16:59:44 +03:00
bp->b_error = EINVAL;
goto bad;
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* If it's a null transfer, return immediately
*/
if (bp->b_bcount == 0)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
goto done;
Always boundary check I/O.
1995-01-13 13:46:32 +03:00
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
Always boundary check I/O.
1995-01-13 13:46:32 +03:00
* Do bounds checking, adjust transfer. if error, process.
* If end of partition, just return.
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
*/
Check RAW_PART against the media size instead of the disklabel. Add the media size in 512-byte sectors to the softc, to avoid some 64 bit computations. Bump the capacity stored in softcs for disks to 64 bits.
2003-04-04 02:18:23 +04:00
if (CDPART(bp->b_dev) == RAW_PART) {
if (bounds_check_with_mediasize(bp, DEV_BSIZE,
cd->params.disksize512) <= 0)
goto done;
} else {
Change bounds_check_with_label() to take a pointer to the disk structure, rather than the label itself. This paves the way for some future changes.
2003-05-11 03:12:28 +04:00
if (bounds_check_with_label(&cd->sc_dk, bp,
Check RAW_PART against the media size instead of the disklabel. Add the media size in 512-byte sectors to the softc, to avoid some 64 bit computations. Bump the capacity stored in softcs for disks to 64 bits.
2003-04-04 02:18:23 +04:00
(cd->flags & (CDF_WLABEL|CDF_LABELLING)) != 0) <= 0)
goto done;
}
Similar changes to sd.c.
1994-10-20 17:03:38 +03:00
Fix a bug in disksort_*() which caused non-optimal ordering when multiple active partitions were on a single spindle. Add a b_rawblkno member to struct buf which contains the non-partition-relative block number to sort by.
2000-02-07 23:16:47 +03:00
/*
* Now convert the block number to absolute and put it in
* terms of the device's logical block size.
*/
blkno = bp->b_blkno / (lp->d_secsize / DEV_BSIZE);
if (CDPART(bp->b_dev) != RAW_PART)
blkno += lp->d_partitions[CDPART(bp->b_dev)].p_offset;
bp->b_rawblkno = blkno;
Add support for smaller sector sizes so we can mount ffs filesystems.
2001-08-16 02:21:01 +04:00
/*
* If the disklabel sector size does not match the device
* sector size we may need to do some extra work.
*/
if (lp->d_secsize != cd->params.blksize) {
/*
* If the xfer is not a multiple of the device block size
* or it is not block aligned, we need to bounce it.
*/
if ((bp->b_bcount % cd->params.blksize) != 0 ||
((blkno * lp->d_secsize) % cd->params.blksize) != 0) {
struct buf *nbp;
void *bounce = NULL;
long count;
if ((bp->b_flags & B_READ) == 0) {
/* XXXX We don't support bouncing writes. */
bp->b_error = EACCES;
goto bad;
}
count = ((blkno * lp->d_secsize) % cd->params.blksize);
/* XXX Store starting offset in bp->b_rawblkno */
bp->b_rawblkno = count;
count += bp->b_bcount;
count = roundup(count, cd->params.blksize);
blkno = ((blkno * lp->d_secsize) / cd->params.blksize);
s = splbio();
nbp = pool_get(&bufpool, PR_NOWAIT);
splx(s);
if (!nbp) {
/* No memory -- fail the iop. */
bp->b_error = ENOMEM;
goto bad;
}
bounce = malloc(count, M_DEVBUF, M_NOWAIT);
if (!bounce) {
/* No memory -- fail the iop. */
s = splbio();
pool_put(&bufpool, nbp);
splx(s);
bp->b_error = ENOMEM;
goto bad;
}
/* Set up the IOP to the bounce buffer. */
Add a new BUF_INIT() macro which initializes b_dep and b_interlock, and use it. This fixes a few places where either b_dep or b_interlock were not properly initialized.
2003-02-25 23:35:31 +03:00
BUF_INIT(nbp);
Add support for smaller sector sizes so we can mount ffs filesystems.
2001-08-16 02:21:01 +04:00
nbp->b_error = 0;
nbp->b_proc = bp->b_proc;
nbp->b_vp = NULLVP;
nbp->b_bcount = count;
nbp->b_bufsize = count;
nbp->b_data = bounce;
nbp->b_rawblkno = blkno;
/* We need to do a read-modify-write operation */
nbp->b_flags = bp->b_flags | B_READ | B_CALL;
nbp->b_iodone = cdbounce;
/* Put ptr to orig buf in b_private and use new buf */
nbp->b_private = bp;
store a i/o priority hint in struct buf for buffer queue discipline.
2004-01-10 17:39:50 +03:00
BIO_COPYPRIO(nbp, bp);
Add support for smaller sector sizes so we can mount ffs filesystems.
2001-08-16 02:21:01 +04:00
bp = nbp;
} else {
/* Xfer is aligned -- just adjust the start block */
bp->b_rawblkno = (blkno * lp->d_secsize) /
cd->params.blksize;
}
}
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
s = splbio();
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
* Place it in the queue of disk activities for this disk.
*
* XXX Only do disksort() if the current operating mode does not
* XXX include tagged queueing.
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
*/
- change the way to specify a bufq strategy. (by string rather than by number) - rather than embedding bufq_state in driver softc, have a pointer to the former. - move bufq related functions from kern/subr_disk.c to kern/subr_bufq.c. - rename method to strategy for consistency. - move some definitions which don't need to be exposed to the rest of kernel from sys/bufq.h to sys/bufq_impl.h. (is it better to move it to kern/ or somewhere?) - fix some obvious breakage in dev/qbus/ts.c. (not tested)
2005-10-15 21:29:10 +04:00
BUFQ_PUT(cd->buf_queue, bp);
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* Tell the device to get going on the transfer if it's
* not doing anything, otherwise just wait for completion
*/
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
cdstart(cd->sc_periph);
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
splx(s);
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
return;
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
bad:
bp->b_flags |= B_ERROR;
done:
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* Correctly set the buf to indicate a completed xfer
*/
bp->b_resid = bp->b_bcount;
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
biodone(bp);
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* cdstart looks to see if there is a buf waiting for the device
* and that the device is not already busy. If both are true,
* It deques the buf and creates a scsi command to perform the
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
* transfer in the buf. The transfer request will call scsipi_done
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
* on completion, which will in turn call this routine again
* so that the next queued transfer is performed.
* The bufs are queued by the strategy routine (cdstrategy)
*
* This routine is also called after other non-queued requests
* have been made of the scsi driver, to ensure that the queue
* continues to be drained.
*
* must be called at the correct (highish) spl level
Fix comment: xxstart() can also be called from xxrestart()
2004-09-07 00:38:14 +04:00
* cdstart() is called at splbio from cdstrategy, cdrestart and scsipi_done
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
*/
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
static void
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
cdstart(struct scsipi_periph *periph)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
{
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
struct cd_softc *cd = (void *)periph->periph_dev;
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
struct buf *bp = 0;
USe more appropriate macro/struct names for READ/WRITE (6) and READ/WRITE (10).
2004-12-08 02:07:31 +03:00
struct scsipi_rw_10 cmd_big;
struct scsi_rw_6 cmd_small;
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
struct scsipi_generic *cmdp;
Make the xxstart() functions reentrant again, as some drivers HBA can call scsipi_done() from their scsipi_request(). For this, add a struct scsipi_xfer * argument to scsipi_command(). If not NULL scsipi_command() will use this to enqueue this xfer, otherwise it'll try to allocate a new one. This scsipi_xfer has to be allocated and initialised by scsipi_make_xs() or equivalent. In xxstart(), allocate a scsipi_xfer using scsipi_make_xs(), and if not NULL, dequeue the buffer before calling scsipi_command(). This makes sure that scsipi_command() will not fail, and also makes sure that xxstart() won't be called again between the BUFQ_PEEK() and BUFQ_GET(). Fix "dequeued wrong buf" panics reported by Juergen Hannken-Illjes in private mail and Andreas Wrede on current-users@. Thanks to Jason Thorpe and Chuck Silver for review, and Andreas Wrede for testing the patch.
2004-09-09 23:35:30 +04:00
struct scsipi_xfer *xs;
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
int flags, nblks, cmdlen, error;
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
SC_DEBUG(periph, SCSIPI_DB2, ("cdstart "));
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
Allow DIOCWDINFO even if label is not `writable'.
1995-03-23 14:33:18 +03:00
* Check if the device has room for another command
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
*/
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
while (periph->periph_active < periph->periph_openings) {
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* there is excess capacity, but a special waits
* It'll need the adapter as soon as we clear out of the
* way and let it run (user level wait).
*/
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
if (periph->periph_flags & PERIPH_WAITING) {
periph->periph_flags &= ~PERIPH_WAITING;
wakeup((caddr_t)periph);
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
return;
}
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
Return ENODEV instead of EIO when we are trying to open a device without media in the drive. restrict "opening of empty drive" to character devices only (reading a block device returns a short read instead of ENODEV, which can lead to confusion).
1999-01-29 14:17:58 +03:00
* If the device has become invalid, abort all the
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
* reads and writes until all files have been closed and
Numerous changes. Many bugs fixed, better autoconfig, a few new features.
1994-12-28 22:42:47 +03:00
* re-opened
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
*/
Improve handling of memory shortage, to fix problems like: sd3(mpt0:0:1:0): unable to allocate scsipi_xfer sd3: not queued, error 12 Havard Eidnes's analysis of this problem is that the scsipi_xfer pool is competing for resources with other pools, including the the inode and vnode pools which can grow quite large. *_scsipi_cmd(): don't biodone the buffer if scsipi_make_xs() fails, let the caller deal with the problem start function of block devices drivers: dequeue the buffer after the scsipi_command() call. If scsipi_command() fails with ENOMEM don't dequeue the buffer, and schedule a callout to call the start function after some delay. scsipi_init(): prime the scsipi_xfer_pool with one page. This ensure that there is always some scsipi_xfer to play with. If scsipi_command() fails because of pool_get(), we're sure there will be resources available later, when the pending commands have completed. Reviewed by Jason Thorpe and Havard Eidnes. Todo: remove the "unable to allocate scsipi_xfer" and "not queued, error %d" printfs, but I choose to keep them for now, to help make sure the code does what it should.
2004-08-28 00:37:28 +04:00
if (__predict_false(
(periph->periph_flags & PERIPH_MEDIA_LOADED) == 0)) {
- change the way to specify a bufq strategy. (by string rather than by number) - rather than embedding bufq_state in driver softc, have a pointer to the former. - move bufq related functions from kern/subr_disk.c to kern/subr_bufq.c. - rename method to strategy for consistency. - move some definitions which don't need to be exposed to the rest of kernel from sys/bufq.h to sys/bufq_impl.h. (is it better to move it to kern/ or somewhere?) - fix some obvious breakage in dev/qbus/ts.c. (not tested)
2005-10-15 21:29:10 +04:00
if ((bp = BUFQ_GET(cd->buf_queue)) != NULL) {
Improve handling of memory shortage, to fix problems like: sd3(mpt0:0:1:0): unable to allocate scsipi_xfer sd3: not queued, error 12 Havard Eidnes's analysis of this problem is that the scsipi_xfer pool is competing for resources with other pools, including the the inode and vnode pools which can grow quite large. *_scsipi_cmd(): don't biodone the buffer if scsipi_make_xs() fails, let the caller deal with the problem start function of block devices drivers: dequeue the buffer after the scsipi_command() call. If scsipi_command() fails with ENOMEM don't dequeue the buffer, and schedule a callout to call the start function after some delay. scsipi_init(): prime the scsipi_xfer_pool with one page. This ensure that there is always some scsipi_xfer to play with. If scsipi_command() fails because of pool_get(), we're sure there will be resources available later, when the pending commands have completed. Reviewed by Jason Thorpe and Havard Eidnes. Todo: remove the "unable to allocate scsipi_xfer" and "not queued, error %d" printfs, but I choose to keep them for now, to help make sure the code does what it should.
2004-08-28 00:37:28 +04:00
bp->b_error = EIO;
bp->b_flags |= B_ERROR;
bp->b_resid = bp->b_bcount;
biodone(bp);
continue;
} else {
return;
}
resolve confusion over who owns the buf after calling scsi_scsi_cmd()
1994-06-16 19:57:39 +04:00
}
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
Improve handling of memory shortage, to fix problems like: sd3(mpt0:0:1:0): unable to allocate scsipi_xfer sd3: not queued, error 12 Havard Eidnes's analysis of this problem is that the scsipi_xfer pool is competing for resources with other pools, including the the inode and vnode pools which can grow quite large. *_scsipi_cmd(): don't biodone the buffer if scsipi_make_xs() fails, let the caller deal with the problem start function of block devices drivers: dequeue the buffer after the scsipi_command() call. If scsipi_command() fails with ENOMEM don't dequeue the buffer, and schedule a callout to call the start function after some delay. scsipi_init(): prime the scsipi_xfer_pool with one page. This ensure that there is always some scsipi_xfer to play with. If scsipi_command() fails because of pool_get(), we're sure there will be resources available later, when the pending commands have completed. Reviewed by Jason Thorpe and Havard Eidnes. Todo: remove the "unable to allocate scsipi_xfer" and "not queued, error %d" printfs, but I choose to keep them for now, to help make sure the code does what it should.
2004-08-28 00:37:28 +04:00
/*
* See if there is a buf with work for us to do..
*/
- change the way to specify a bufq strategy. (by string rather than by number) - rather than embedding bufq_state in driver softc, have a pointer to the former. - move bufq related functions from kern/subr_disk.c to kern/subr_bufq.c. - rename method to strategy for consistency. - move some definitions which don't need to be exposed to the rest of kernel from sys/bufq.h to sys/bufq_impl.h. (is it better to move it to kern/ or somewhere?) - fix some obvious breakage in dev/qbus/ts.c. (not tested)
2005-10-15 21:29:10 +04:00
if ((bp = BUFQ_PEEK(cd->buf_queue)) == NULL)
Improve handling of memory shortage, to fix problems like: sd3(mpt0:0:1:0): unable to allocate scsipi_xfer sd3: not queued, error 12 Havard Eidnes's analysis of this problem is that the scsipi_xfer pool is competing for resources with other pools, including the the inode and vnode pools which can grow quite large. *_scsipi_cmd(): don't biodone the buffer if scsipi_make_xs() fails, let the caller deal with the problem start function of block devices drivers: dequeue the buffer after the scsipi_command() call. If scsipi_command() fails with ENOMEM don't dequeue the buffer, and schedule a callout to call the start function after some delay. scsipi_init(): prime the scsipi_xfer_pool with one page. This ensure that there is always some scsipi_xfer to play with. If scsipi_command() fails because of pool_get(), we're sure there will be resources available later, when the pending commands have completed. Reviewed by Jason Thorpe and Havard Eidnes. Todo: remove the "unable to allocate scsipi_xfer" and "not queued, error %d" printfs, but I choose to keep them for now, to help make sure the code does what it should.
2004-08-28 00:37:28 +04:00
return;
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
Fix a bug in disksort_*() which caused non-optimal ordering when multiple active partitions were on a single spindle. Add a b_rawblkno member to struct buf which contains the non-partition-relative block number to sort by.
2000-02-07 23:16:47 +03:00
* We have a buf, now we should make a command.
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
*/
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
Add support for smaller sector sizes so we can mount ffs filesystems.
2001-08-16 02:21:01 +04:00
nblks = howmany(bp->b_bcount, cd->params.blksize);
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
In both cd.c and sd.c: If the read or write request can fit into a 6-byte cdb, then use a 6-byte cdb, otherwise use the 10-byte as before. In sd.c: In sdattach(), make a note if the device is "ancient" (i.e. inqbuf.version & SID_ANSII == 0). Implement sdminphys(): if the device is "ancient", shorten the transfer so it will fit into a 6-byte cdb. These changes have eliminated the rejected read/write requests on my Sun 3/60 with 2 ESDI disks behind an Emulex MD21.
1995-12-07 22:11:32 +03:00
* Fill out the scsi command. If the transfer will
* fit in a "small" cdb, use it.
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
*/
Fix a bug in disksort_*() which caused non-optimal ordering when multiple active partitions were on a single spindle. Add a b_rawblkno member to struct buf which contains the non-partition-relative block number to sort by.
2000-02-07 23:16:47 +03:00
if (((bp->b_rawblkno & 0x1fffff) == bp->b_rawblkno) &&
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
((nblks & 0xff) == nblks) &&
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
!(periph->periph_quirks & PQUIRK_ONLYBIG)) {
In both cd.c and sd.c: If the read or write request can fit into a 6-byte cdb, then use a 6-byte cdb, otherwise use the 10-byte as before. In sd.c: In sdattach(), make a note if the device is "ancient" (i.e. inqbuf.version & SID_ANSII == 0). Implement sdminphys(): if the device is "ancient", shorten the transfer so it will fit into a 6-byte cdb. These changes have eliminated the rejected read/write requests on my Sun 3/60 with 2 ESDI disks behind an Emulex MD21.
1995-12-07 22:11:32 +03:00
/*
* We can fit in a small cdb.
*/
bzero -> memset
2001-07-18 22:21:04 +04:00
memset(&cmd_small, 0, sizeof(cmd_small));
In both cd.c and sd.c: If the read or write request can fit into a 6-byte cdb, then use a 6-byte cdb, otherwise use the 10-byte as before. In sd.c: In sdattach(), make a note if the device is "ancient" (i.e. inqbuf.version & SID_ANSII == 0). Implement sdminphys(): if the device is "ancient", shorten the transfer so it will fit into a 6-byte cdb. These changes have eliminated the rejected read/write requests on my Sun 3/60 with 2 ESDI disks behind an Emulex MD21.
1995-12-07 22:11:32 +03:00
cmd_small.opcode = (bp->b_flags & B_READ) ?
USe more appropriate macro/struct names for READ/WRITE (6) and READ/WRITE (10).
2004-12-08 02:07:31 +03:00
SCSI_READ_6_COMMAND : SCSI_WRITE_6_COMMAND;
Fix a bug in disksort_*() which caused non-optimal ordering when multiple active partitions were on a single spindle. Add a b_rawblkno member to struct buf which contains the non-partition-relative block number to sort by.
2000-02-07 23:16:47 +03:00
_lto3b(bp->b_rawblkno, cmd_small.addr);
In both cd.c and sd.c: If the read or write request can fit into a 6-byte cdb, then use a 6-byte cdb, otherwise use the 10-byte as before. In sd.c: In sdattach(), make a note if the device is "ancient" (i.e. inqbuf.version & SID_ANSII == 0). Implement sdminphys(): if the device is "ancient", shorten the transfer so it will fit into a 6-byte cdb. These changes have eliminated the rejected read/write requests on my Sun 3/60 with 2 ESDI disks behind an Emulex MD21.
1995-12-07 22:11:32 +03:00
cmd_small.length = nblks & 0xff;
cmdlen = sizeof(cmd_small);
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
cmdp = (struct scsipi_generic *)&cmd_small;
Eliminate the separate ATAPI and SCSI attachments for "cd".
2003-09-08 05:56:33 +04:00
} else {
In both cd.c and sd.c: If the read or write request can fit into a 6-byte cdb, then use a 6-byte cdb, otherwise use the 10-byte as before. In sd.c: In sdattach(), make a note if the device is "ancient" (i.e. inqbuf.version & SID_ANSII == 0). Implement sdminphys(): if the device is "ancient", shorten the transfer so it will fit into a 6-byte cdb. These changes have eliminated the rejected read/write requests on my Sun 3/60 with 2 ESDI disks behind an Emulex MD21.
1995-12-07 22:11:32 +03:00
/*
* Need a large cdb.
*/
bzero -> memset
2001-07-18 22:21:04 +04:00
memset(&cmd_big, 0, sizeof(cmd_big));
In both cd.c and sd.c: If the read or write request can fit into a 6-byte cdb, then use a 6-byte cdb, otherwise use the 10-byte as before. In sd.c: In sdattach(), make a note if the device is "ancient" (i.e. inqbuf.version & SID_ANSII == 0). Implement sdminphys(): if the device is "ancient", shorten the transfer so it will fit into a 6-byte cdb. These changes have eliminated the rejected read/write requests on my Sun 3/60 with 2 ESDI disks behind an Emulex MD21.
1995-12-07 22:11:32 +03:00
cmd_big.opcode = (bp->b_flags & B_READ) ?
USe more appropriate macro/struct names for READ/WRITE (6) and READ/WRITE (10).
2004-12-08 02:07:31 +03:00
READ_10 : WRITE_10;
Fix a bug in disksort_*() which caused non-optimal ordering when multiple active partitions were on a single spindle. Add a b_rawblkno member to struct buf which contains the non-partition-relative block number to sort by.
2000-02-07 23:16:47 +03:00
_lto4b(bp->b_rawblkno, cmd_big.addr);
Define a full set of [234][bl]tol() and lto[234][bl]() conversion functions, inlined. Use sized types in protocol structures. Make the definition of scsi_sense_data less ugly.
1996-03-19 06:05:15 +03:00
_lto2b(nblks, cmd_big.length);
In both cd.c and sd.c: If the read or write request can fit into a 6-byte cdb, then use a 6-byte cdb, otherwise use the 10-byte as before. In sd.c: In sdattach(), make a note if the device is "ancient" (i.e. inqbuf.version & SID_ANSII == 0). Implement sdminphys(): if the device is "ancient", shorten the transfer so it will fit into a 6-byte cdb. These changes have eliminated the rejected read/write requests on my Sun 3/60 with 2 ESDI disks behind an Emulex MD21.
1995-12-07 22:11:32 +03:00
cmdlen = sizeof(cmd_big);
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
cmdp = (struct scsipi_generic *)&cmd_big;
In both cd.c and sd.c: If the read or write request can fit into a 6-byte cdb, then use a 6-byte cdb, otherwise use the 10-byte as before. In sd.c: In sdattach(), make a note if the device is "ancient" (i.e. inqbuf.version & SID_ANSII == 0). Implement sdminphys(): if the device is "ancient", shorten the transfer so it will fit into a 6-byte cdb. These changes have eliminated the rejected read/write requests on my Sun 3/60 with 2 ESDI disks behind an Emulex MD21.
1995-12-07 22:11:32 +03:00
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
New generic disk framework. Highlights: - New metrics handling. Metrics are now kept in the new `struct disk'. Busy time is now stored as a timeval, and transfer count in bytes. - Storage for disklabels is now dynamically allocated, so that the size of the disk structure is not machine-dependent. - Several new functions for attaching and detaching disks, and handling metrics calculation. Old-style instrumentation is still supported in drivers that did it before. However, old-style instrumentation is being deprecated, and will go away once the userland utilities are updated for the new framework. For usage and architectural details, see the forthcoming disk(9) manual page.
1996-01-08 01:01:38 +03:00
/* Instrumentation. */
disk_busy(&cd->sc_dk);
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
/*
* Figure out what flags to use.
*/
Remove the old device buffer queue interface. Approved by: Jason R. Thorpe <thorpej@wasabisystems.com>
2002-08-30 19:43:36 +04:00
flags = XS_CTL_NOSLEEP|XS_CTL_ASYNC|XS_CTL_SIMPLE_TAG;
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
if (bp->b_flags & B_READ)
Use simple tags for read and write, unless B_ORDERED is set, in which case we use an ordered tag.
2001-04-28 08:11:31 +04:00
flags |= XS_CTL_DATA_IN;
Don't forget to set XS_CTL_DATA_OUT (that was removed in previous).
2001-04-28 13:33:25 +04:00
else
flags |= XS_CTL_DATA_OUT;
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* Call the routine that chats with the adapter.
* Note: we cannot sleep as we may be an interrupt
*/
Make the xxstart() functions reentrant again, as some drivers HBA can call scsipi_done() from their scsipi_request(). For this, add a struct scsipi_xfer * argument to scsipi_command(). If not NULL scsipi_command() will use this to enqueue this xfer, otherwise it'll try to allocate a new one. This scsipi_xfer has to be allocated and initialised by scsipi_make_xs() or equivalent. In xxstart(), allocate a scsipi_xfer using scsipi_make_xs(), and if not NULL, dequeue the buffer before calling scsipi_command(). This makes sure that scsipi_command() will not fail, and also makes sure that xxstart() won't be called again between the BUFQ_PEEK() and BUFQ_GET(). Fix "dequeued wrong buf" panics reported by Juergen Hannken-Illjes in private mail and Andreas Wrede on current-users@. Thanks to Jason Thorpe and Chuck Silver for review, and Andreas Wrede for testing the patch.
2004-09-09 23:35:30 +04:00
xs = scsipi_make_xs(periph, cmdp, cmdlen,
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
(u_char *)bp->b_data, bp->b_bcount,
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
CDRETRIES, 30000, bp, flags);
Make the xxstart() functions reentrant again, as some drivers HBA can call scsipi_done() from their scsipi_request(). For this, add a struct scsipi_xfer * argument to scsipi_command(). If not NULL scsipi_command() will use this to enqueue this xfer, otherwise it'll try to allocate a new one. This scsipi_xfer has to be allocated and initialised by scsipi_make_xs() or equivalent. In xxstart(), allocate a scsipi_xfer using scsipi_make_xs(), and if not NULL, dequeue the buffer before calling scsipi_command(). This makes sure that scsipi_command() will not fail, and also makes sure that xxstart() won't be called again between the BUFQ_PEEK() and BUFQ_GET(). Fix "dequeued wrong buf" panics reported by Juergen Hannken-Illjes in private mail and Andreas Wrede on current-users@. Thanks to Jason Thorpe and Chuck Silver for review, and Andreas Wrede for testing the patch.
2004-09-09 23:35:30 +04:00
if (__predict_false(xs == NULL)) {
Improve handling of memory shortage, to fix problems like: sd3(mpt0:0:1:0): unable to allocate scsipi_xfer sd3: not queued, error 12 Havard Eidnes's analysis of this problem is that the scsipi_xfer pool is competing for resources with other pools, including the the inode and vnode pools which can grow quite large. *_scsipi_cmd(): don't biodone the buffer if scsipi_make_xs() fails, let the caller deal with the problem start function of block devices drivers: dequeue the buffer after the scsipi_command() call. If scsipi_command() fails with ENOMEM don't dequeue the buffer, and schedule a callout to call the start function after some delay. scsipi_init(): prime the scsipi_xfer_pool with one page. This ensure that there is always some scsipi_xfer to play with. If scsipi_command() fails because of pool_get(), we're sure there will be resources available later, when the pending commands have completed. Reviewed by Jason Thorpe and Havard Eidnes. Todo: remove the "unable to allocate scsipi_xfer" and "not queued, error %d" printfs, but I choose to keep them for now, to help make sure the code does what it should.
2004-08-28 00:37:28 +04:00
/*
* out of memory. Keep this buffer in the queue, and
* retry later.
*/
callout_reset(&cd->sc_callout, hz / 2, cdrestart,
periph);
return;
}
Make the xxstart() functions reentrant again, as some drivers HBA can call scsipi_done() from their scsipi_request(). For this, add a struct scsipi_xfer * argument to scsipi_command(). If not NULL scsipi_command() will use this to enqueue this xfer, otherwise it'll try to allocate a new one. This scsipi_xfer has to be allocated and initialised by scsipi_make_xs() or equivalent. In xxstart(), allocate a scsipi_xfer using scsipi_make_xs(), and if not NULL, dequeue the buffer before calling scsipi_command(). This makes sure that scsipi_command() will not fail, and also makes sure that xxstart() won't be called again between the BUFQ_PEEK() and BUFQ_GET(). Fix "dequeued wrong buf" panics reported by Juergen Hannken-Illjes in private mail and Andreas Wrede on current-users@. Thanks to Jason Thorpe and Chuck Silver for review, and Andreas Wrede for testing the patch.
2004-09-09 23:35:30 +04:00
/*
* need to dequeue the buffer before queuing the command,
* because cdstart may be called recursively from the
* HBA driver
*/
Improve handling of memory shortage, to fix problems like: sd3(mpt0:0:1:0): unable to allocate scsipi_xfer sd3: not queued, error 12 Havard Eidnes's analysis of this problem is that the scsipi_xfer pool is competing for resources with other pools, including the the inode and vnode pools which can grow quite large. *_scsipi_cmd(): don't biodone the buffer if scsipi_make_xs() fails, let the caller deal with the problem start function of block devices drivers: dequeue the buffer after the scsipi_command() call. If scsipi_command() fails with ENOMEM don't dequeue the buffer, and schedule a callout to call the start function after some delay. scsipi_init(): prime the scsipi_xfer_pool with one page. This ensure that there is always some scsipi_xfer to play with. If scsipi_command() fails because of pool_get(), we're sure there will be resources available later, when the pending commands have completed. Reviewed by Jason Thorpe and Havard Eidnes. Todo: remove the "unable to allocate scsipi_xfer" and "not queued, error %d" printfs, but I choose to keep them for now, to help make sure the code does what it should.
2004-08-28 00:37:28 +04:00
#ifdef DIAGNOSTIC
- change the way to specify a bufq strategy. (by string rather than by number) - rather than embedding bufq_state in driver softc, have a pointer to the former. - move bufq related functions from kern/subr_disk.c to kern/subr_bufq.c. - rename method to strategy for consistency. - move some definitions which don't need to be exposed to the rest of kernel from sys/bufq.h to sys/bufq_impl.h. (is it better to move it to kern/ or somewhere?) - fix some obvious breakage in dev/qbus/ts.c. (not tested)
2005-10-15 21:29:10 +04:00
if (BUFQ_GET(cd->buf_queue) != bp)
Improve handling of memory shortage, to fix problems like: sd3(mpt0:0:1:0): unable to allocate scsipi_xfer sd3: not queued, error 12 Havard Eidnes's analysis of this problem is that the scsipi_xfer pool is competing for resources with other pools, including the the inode and vnode pools which can grow quite large. *_scsipi_cmd(): don't biodone the buffer if scsipi_make_xs() fails, let the caller deal with the problem start function of block devices drivers: dequeue the buffer after the scsipi_command() call. If scsipi_command() fails with ENOMEM don't dequeue the buffer, and schedule a callout to call the start function after some delay. scsipi_init(): prime the scsipi_xfer_pool with one page. This ensure that there is always some scsipi_xfer to play with. If scsipi_command() fails because of pool_get(), we're sure there will be resources available later, when the pending commands have completed. Reviewed by Jason Thorpe and Havard Eidnes. Todo: remove the "unable to allocate scsipi_xfer" and "not queued, error %d" printfs, but I choose to keep them for now, to help make sure the code does what it should.
2004-08-28 00:37:28 +04:00
panic("cdstart(): dequeued wrong buf");
#else
- change the way to specify a bufq strategy. (by string rather than by number) - rather than embedding bufq_state in driver softc, have a pointer to the former. - move bufq related functions from kern/subr_disk.c to kern/subr_bufq.c. - rename method to strategy for consistency. - move some definitions which don't need to be exposed to the rest of kernel from sys/bufq.h to sys/bufq_impl.h. (is it better to move it to kern/ or somewhere?) - fix some obvious breakage in dev/qbus/ts.c. (not tested)
2005-10-15 21:29:10 +04:00
BUFQ_GET(cd->buf_queue);
Improve handling of memory shortage, to fix problems like: sd3(mpt0:0:1:0): unable to allocate scsipi_xfer sd3: not queued, error 12 Havard Eidnes's analysis of this problem is that the scsipi_xfer pool is competing for resources with other pools, including the the inode and vnode pools which can grow quite large. *_scsipi_cmd(): don't biodone the buffer if scsipi_make_xs() fails, let the caller deal with the problem start function of block devices drivers: dequeue the buffer after the scsipi_command() call. If scsipi_command() fails with ENOMEM don't dequeue the buffer, and schedule a callout to call the start function after some delay. scsipi_init(): prime the scsipi_xfer_pool with one page. This ensure that there is always some scsipi_xfer to play with. If scsipi_command() fails because of pool_get(), we're sure there will be resources available later, when the pending commands have completed. Reviewed by Jason Thorpe and Havard Eidnes. Todo: remove the "unable to allocate scsipi_xfer" and "not queued, error %d" printfs, but I choose to keep them for now, to help make sure the code does what it should.
2004-08-28 00:37:28 +04:00
#endif
In places where we've already called scsipi_make_xs(), call scsipi_execute_xs() directly rather than going through scsipi_command().
2004-09-18 03:35:13 +04:00
error = scsipi_execute_xs(xs);
Make the xxstart() functions reentrant again, as some drivers HBA can call scsipi_done() from their scsipi_request(). For this, add a struct scsipi_xfer * argument to scsipi_command(). If not NULL scsipi_command() will use this to enqueue this xfer, otherwise it'll try to allocate a new one. This scsipi_xfer has to be allocated and initialised by scsipi_make_xs() or equivalent. In xxstart(), allocate a scsipi_xfer using scsipi_make_xs(), and if not NULL, dequeue the buffer before calling scsipi_command(). This makes sure that scsipi_command() will not fail, and also makes sure that xxstart() won't be called again between the BUFQ_PEEK() and BUFQ_GET(). Fix "dequeued wrong buf" panics reported by Juergen Hannken-Illjes in private mail and Andreas Wrede on current-users@. Thanks to Jason Thorpe and Chuck Silver for review, and Andreas Wrede for testing the patch.
2004-09-09 23:35:30 +04:00
/* with a scsipi_xfer preallocated, scsipi_command can't fail */
KASSERT(error == 0);
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
Improve handling of memory shortage, to fix problems like: sd3(mpt0:0:1:0): unable to allocate scsipi_xfer sd3: not queued, error 12 Havard Eidnes's analysis of this problem is that the scsipi_xfer pool is competing for resources with other pools, including the the inode and vnode pools which can grow quite large. *_scsipi_cmd(): don't biodone the buffer if scsipi_make_xs() fails, let the caller deal with the problem start function of block devices drivers: dequeue the buffer after the scsipi_command() call. If scsipi_command() fails with ENOMEM don't dequeue the buffer, and schedule a callout to call the start function after some delay. scsipi_init(): prime the scsipi_xfer_pool with one page. This ensure that there is always some scsipi_xfer to play with. If scsipi_command() fails because of pool_get(), we're sure there will be resources available later, when the pending commands have completed. Reviewed by Jason Thorpe and Havard Eidnes. Todo: remove the "unable to allocate scsipi_xfer" and "not queued, error %d" printfs, but I choose to keep them for now, to help make sure the code does what it should.
2004-08-28 00:37:28 +04:00
static void
cdrestart(void *v)
{
int s = splbio();
cdstart((struct scsipi_periph *)v);
splx(s);
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static void
Do not manipulate xs->bp in "generic" code -- do it only in the psw_done routine. As part of this, pass down our pre-parsed error code -- though this interface will probably change later to accomodate better error handling.
2004-09-18 03:10:50 +04:00
cddone(struct scsipi_xfer *xs, int error)
New generic disk framework. Highlights: - New metrics handling. Metrics are now kept in the new `struct disk'. Busy time is now stored as a timeval, and transfer count in bytes. - Storage for disklabels is now dynamically allocated, so that the size of the disk structure is not machine-dependent. - Several new functions for attaching and detaching disks, and handling metrics calculation. Old-style instrumentation is still supported in drivers that did it before. However, old-style instrumentation is being deprecated, and will go away once the userland utilities are updated for the new framework. For usage and architectural details, see the forthcoming disk(9) manual page.
1996-01-08 01:01:38 +03:00
{
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
struct cd_softc *cd = (void *)xs->xs_periph->periph_dev;
Do not manipulate xs->bp in "generic" code -- do it only in the psw_done routine. As part of this, pass down our pre-parsed error code -- though this interface will probably change later to accomodate better error handling.
2004-09-18 03:10:50 +04:00
struct buf *bp = xs->bp;
New generic disk framework. Highlights: - New metrics handling. Metrics are now kept in the new `struct disk'. Busy time is now stored as a timeval, and transfer count in bytes. - Storage for disklabels is now dynamically allocated, so that the size of the disk structure is not machine-dependent. - Several new functions for attaching and detaching disks, and handling metrics calculation. Old-style instrumentation is still supported in drivers that did it before. However, old-style instrumentation is being deprecated, and will go away once the userland utilities are updated for the new framework. For usage and architectural details, see the forthcoming disk(9) manual page.
1996-01-08 01:01:38 +03:00
Do not manipulate xs->bp in "generic" code -- do it only in the psw_done routine. As part of this, pass down our pre-parsed error code -- though this interface will probably change later to accomodate better error handling.
2004-09-18 03:10:50 +04:00
if (bp) {
bp->b_error = error;
bp->b_resid = xs->resid;
if (error)
bp->b_flags |= B_ERROR;
disk_unbusy(&cd->sc_dk, bp->b_bcount - bp->b_resid,
(bp->b_flags & B_READ));
o Make usage of /dev/random dependant on pseudo-device rnd # /dev/random and in-kernel generator in config files. o Add declaration to all architectures. o Clean up copyright message in rnd.c, rnd.h, and rndpool.c to include that this code is derived in part from Ted Tyso's linux code.
1997-10-13 04:46:08 +04:00
#if NRND > 0
Do not manipulate xs->bp in "generic" code -- do it only in the psw_done routine. As part of this, pass down our pre-parsed error code -- though this interface will probably change later to accomodate better error handling.
2004-09-18 03:10:50 +04:00
rnd_add_uint32(&cd->rnd_source, bp->b_rawblkno);
o Make usage of /dev/random dependant on pseudo-device rnd # /dev/random and in-kernel generator in config files. o Add declaration to all architectures. o Clean up copyright message in rnd.c, rnd.h, and rndpool.c to include that this code is derived in part from Ted Tyso's linux code.
1997-10-13 04:46:08 +04:00
#endif
Do not manipulate xs->bp in "generic" code -- do it only in the psw_done routine. As part of this, pass down our pre-parsed error code -- though this interface will probably change later to accomodate better error handling.
2004-09-18 03:10:50 +04:00
biodone(bp);
Add hooks to insert timing info into the random system
1997-10-10 05:09:03 +04:00
}
New generic disk framework. Highlights: - New metrics handling. Metrics are now kept in the new `struct disk'. Busy time is now stored as a timeval, and transfer count in bytes. - Storage for disklabels is now dynamically allocated, so that the size of the disk structure is not machine-dependent. - Several new functions for attaching and detaching disks, and handling metrics calculation. Old-style instrumentation is still supported in drivers that did it before. However, old-style instrumentation is being deprecated, and will go away once the userland utilities are updated for the new framework. For usage and architectural details, see the forthcoming disk(9) manual page.
1996-01-08 01:01:38 +03:00
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static void
cdbounce(struct buf *bp)
Add support for smaller sector sizes so we can mount ffs filesystems.
2001-08-16 02:21:01 +04:00
{
struct buf *obp = (struct buf *)bp->b_private;
if (bp->b_flags & B_ERROR) {
/* EEK propagate the error and free the memory */
goto done;
}
if (obp->b_flags & B_READ) {
/* Copy data to the final destination and free the buf. */
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
memcpy(obp->b_data, bp->b_data+obp->b_rawblkno,
Add support for smaller sector sizes so we can mount ffs filesystems.
2001-08-16 02:21:01 +04:00
obp->b_bcount);
} else {
/*
* XXXX This is a CD-ROM -- READ ONLY -- why do we bother with
* XXXX any of this write stuff?
*/
if (bp->b_flags & B_READ) {
struct cd_softc *cd = cd_cd.cd_devs[CDUNIT(bp->b_dev)];
struct buf *nbp;
int s;
/* Read part of RMW complete. */
memcpy(bp->b_data+obp->b_rawblkno, obp->b_data,
obp->b_bcount);
s = splbio();
/* We need to alloc a new buf. */
nbp = pool_get(&bufpool, PR_NOWAIT);
if (!nbp) {
splx(s);
/* No buf available. */
bp->b_flags |= B_ERROR;
bp->b_error = ENOMEM;
bp->b_resid = bp->b_bcount;
}
/* Set up the IOP to the bounce buffer. */
Add a new BUF_INIT() macro which initializes b_dep and b_interlock, and use it. This fixes a few places where either b_dep or b_interlock were not properly initialized.
2003-02-25 23:35:31 +03:00
BUF_INIT(nbp);
Add support for smaller sector sizes so we can mount ffs filesystems.
2001-08-16 02:21:01 +04:00
nbp->b_error = 0;
nbp->b_proc = bp->b_proc;
nbp->b_vp = NULLVP;
nbp->b_bcount = bp->b_bcount;
nbp->b_bufsize = bp->b_bufsize;
nbp->b_data = bp->b_data;
nbp->b_rawblkno = bp->b_rawblkno;
/* We need to do a read-modify-write operation */
nbp->b_flags = obp->b_flags | B_CALL;
nbp->b_iodone = cdbounce;
/* Put ptr to orig buf in b_private and use new buf */
nbp->b_private = obp;
/*
* Place it in the queue of disk activities for this
* disk.
*
* XXX Only do disksort() if the current operating mode
* XXX does not include tagged queueing.
*/
- change the way to specify a bufq strategy. (by string rather than by number) - rather than embedding bufq_state in driver softc, have a pointer to the former. - move bufq related functions from kern/subr_disk.c to kern/subr_bufq.c. - rename method to strategy for consistency. - move some definitions which don't need to be exposed to the rest of kernel from sys/bufq.h to sys/bufq_impl.h. (is it better to move it to kern/ or somewhere?) - fix some obvious breakage in dev/qbus/ts.c. (not tested)
2005-10-15 21:29:10 +04:00
BUFQ_PUT(cd->buf_queue, nbp);
Add support for smaller sector sizes so we can mount ffs filesystems.
2001-08-16 02:21:01 +04:00
/*
* Tell the device to get going on the transfer if it's
* not doing anything, otherwise just wait for
* completion
*/
cdstart(cd->sc_periph);
splx(s);
return;
}
}
done:
obp->b_flags |= (bp->b_flags&(B_EINTR|B_ERROR));
obp->b_error = bp->b_error;
obp->b_resid = bp->b_resid;
free(bp->b_data, M_DEVBUF);
biodone(obp);
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cd_interpret_sense(struct scsipi_xfer *xs)
Retry requests if the CD-ROM drive returns the sense code "Device in Process of Becoming Ready".
2001-01-20 01:47:46 +03:00
{
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
struct scsipi_periph *periph = xs->xs_periph;
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
struct scsi_sense_data *sense = &xs->sense.scsi_sense;
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
int retval = EJUSTRETURN;
Retry requests if the CD-ROM drive returns the sense code "Device in Process of Becoming Ready".
2001-01-20 01:47:46 +03:00
/*
* If it isn't a extended or extended/deferred error, let
* the generic code handle it.
*/
fix more SSD_RCODE_VALID misuse introduced in cleanup
2005-04-25 21:52:30 +04:00
if (SSD_RCODE(sense->response_code) != SSD_RCODE_CURRENT &&
SSD_RCODE(sense->response_code) != SSD_RCODE_DEFERRED)
Retry requests if the CD-ROM drive returns the sense code "Device in Process of Becoming Ready".
2001-01-20 01:47:46 +03:00
return (retval);
Don't special case the retry for Media Changed... More later.
2001-08-20 19:45:10 +04:00
/*
* If we got a "Unit not ready" (SKEY_NOT_READY) and "Logical Unit
* Is In The Process of Becoming Ready" (Sense code 0x04,0x01), then
* wait a bit for the drive to spin up
*/
Retry requests if the CD-ROM drive returns the sense code "Device in Process of Becoming Ready".
2001-01-20 01:47:46 +03:00
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
if (SSD_SENSE_KEY(sense->flags) == SKEY_NOT_READY &&
sense->asc == 0x4 &&
sense->ascq == 0x01) {
Two changes submitted by Sergey Svishchev <svs@ropnet.ru>: - Don't send START UNIT upon open if the drive is playing audio (PR 11768). - If we receive sense indicating that the media changed, retry (PR 8326).
2001-08-20 15:31:10 +04:00
/*
Don't special case the retry for Media Changed... More later.
2001-08-20 19:45:10 +04:00
* Sleep for 5 seconds to wait for the drive to spin up
Two changes submitted by Sergey Svishchev <svs@ropnet.ru>: - Don't send START UNIT upon open if the drive is playing audio (PR 11768). - If we receive sense indicating that the media changed, retry (PR 8326).
2001-08-20 15:31:10 +04:00
*/
Don't special case the retry for Media Changed... More later.
2001-08-20 19:45:10 +04:00
SC_DEBUG(periph, SCSIPI_DB1, ("Waiting 5 sec for CD "
"spinup\n"));
Test callout_pending(), not callout_active(), and eliminate now-unnecessary callout_deactivate() calls.
2003-02-04 02:50:59 +03:00
if (!callout_pending(&periph->periph_callout))
If periph->periph_callout is already active, don't freeze the periph again: scispi_periph_timed_thaw() will be called only one time anyway.
2002-05-05 19:16:30 +04:00
scsipi_periph_freeze(periph, 1);
Don't special case the retry for Media Changed... More later.
2001-08-20 19:45:10 +04:00
callout_reset(&periph->periph_callout,
5 * hz, scsipi_periph_timed_thaw, periph);
retval = ERESTART;
}
Retry requests if the CD-ROM drive returns the sense code "Device in Process of Becoming Ready".
2001-01-20 01:47:46 +03:00
return (retval);
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static void
cdminphys(struct buf *bp)
Increase *OUTSTANDING. Remove old dk_establish() garbage. Support the `ancient' bit in both drivers.
1996-03-26 23:32:06 +03:00
{
SDUNIT -> CDUNIT in last.
1996-03-27 03:48:50 +03:00
struct cd_softc *cd = cd_cd.cd_devs[CDUNIT(bp->b_dev)];
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
long xmax;
Increase *OUTSTANDING. Remove old dk_establish() garbage. Support the `ancient' bit in both drivers.
1996-03-26 23:32:06 +03:00
/*
* If the device is ancient, we want to make sure that
* the transfer fits into a 6-byte cdb.
*
* XXX Note that the SCSI-I spec says that 256-block transfers
* are allowed in a 6-byte read/write, and are specified
* by settng the "length" to 0. However, we're conservative
* here, allowing only 255-block transfers in case an
* ancient device gets confused by length == 0. A length of 0
* in a 10-byte read/write actually means 0 blocks.
*/
if (cd->flags & CDF_ANCIENT) {
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
xmax = cd->sc_dk.dk_label->d_secsize * 0xff;
Increase *OUTSTANDING. Remove old dk_establish() garbage. Support the `ancient' bit in both drivers.
1996-03-26 23:32:06 +03:00
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
if (bp->b_bcount > xmax)
bp->b_bcount = xmax;
Increase *OUTSTANDING. Remove old dk_establish() garbage. Support the `ancient' bit in both drivers.
1996-03-26 23:32:06 +03:00
}
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
(*cd->sc_periph->periph_channel->chan_adapter->adapt_minphys)(bp);
Increase *OUTSTANDING. Remove old dk_establish() garbage. Support the `ancient' bit in both drivers.
1996-03-26 23:32:06 +03:00
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cdread(dev_t dev, struct uio *uio, int ioflag)
Make each disk and tape driver define its own read and write functions. Deprecate rawread() and rawwrite() completely. Remove d_strategy from cdevsw to force the abstraction barrier.
1995-07-04 11:15:28 +04:00
{
Increase *OUTSTANDING. Remove old dk_establish() garbage. Support the `ancient' bit in both drivers.
1996-03-26 23:32:06 +03:00
return (physio(cdstrategy, NULL, dev, B_READ, cdminphys, uio));
Make each disk and tape driver define its own read and write functions. Deprecate rawread() and rawwrite() completely. Remove d_strategy from cdevsw to force the abstraction barrier.
1995-07-04 11:15:28 +04:00
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cdwrite(dev_t dev, struct uio *uio, int ioflag)
Make each disk and tape driver define its own read and write functions. Deprecate rawread() and rawwrite() completely. Remove d_strategy from cdevsw to force the abstraction barrier.
1995-07-04 11:15:28 +04:00
{
Increase *OUTSTANDING. Remove old dk_establish() garbage. Support the `ancient' bit in both drivers.
1996-03-26 23:32:06 +03:00
return (physio(cdstrategy, NULL, dev, B_WRITE, cdminphys, uio));
Make each disk and tape driver define its own read and write functions. Deprecate rawread() and rawwrite() completely. Remove d_strategy from cdevsw to force the abstraction barrier.
1995-07-04 11:15:28 +04:00
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
#if 0 /* XXX Not used */
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
/*
Spell address with two d's. Inspired by similar changes in OpenBSD, originating from Jonathon Gray and forwarded by jmc@openbsd.
2003-11-10 11:51:51 +03:00
* conversion between minute-seconde-frame and logical block address
* addresses format
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
*/
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static void
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
lba2msf(u_long lba, u_char *m, u_char *s, u_char *f)
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
{
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
u_long tmp;
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
tmp = lba + CD_BLOCK_OFFSET; /* offset of first logical frame */
tmp &= 0xffffff; /* negative lbas use only 24 bits */
*m = tmp / (CD_SECS * CD_FRAMES);
tmp %= (CD_SECS * CD_FRAMES);
*s = tmp / CD_FRAMES;
*f = tmp % CD_FRAMES;
}
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
static u_long
msf2lba(u_char m, u_char s, u_char f)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
{
return ((((m * CD_SECS) + s) * CD_FRAMES + f) - CD_BLOCK_OFFSET);
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
#endif /* XXX Not used */
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cdreadmsaddr(struct cd_softc *cd, int *addr)
put multisession offset code into a separate function and initialize the p_cdsession field of partition 'a' in the default disklabel
2002-05-27 20:42:30 +04:00
{
struct scsipi_periph *periph = cd->sc_periph;
int error;
Second stage of read TOC modifications. Introducing the various forms of TOC reading. Note that the external interfaces haven't changed. Only the formatted toc is requestable. The read msinfo command hasn't been changed in this patch though it could become more clever using the CD_TOC_MSINFO format.
2005-12-11 00:17:21 +03:00
struct cd_formatted_toc toc;
put multisession offset code into a separate function and initialize the p_cdsession field of partition 'a' in the default disklabel
2002-05-27 20:42:30 +04:00
struct cd_toc_entry *cte;
Second stage of read TOC modifications. Introducing the various forms of TOC reading. Note that the external interfaces haven't changed. Only the formatted toc is requestable. The read msinfo command hasn't been changed in this patch though it could become more clever using the CD_TOC_MSINFO format.
2005-12-11 00:17:21 +03:00
error = cd_read_toc(cd, CD_TOC_FORM, 0, 0, &toc,
put multisession offset code into a separate function and initialize the p_cdsession field of partition 'a' in the default disklabel
2002-05-27 20:42:30 +04:00
sizeof(struct ioc_toc_header) + sizeof(struct cd_toc_entry),
XS_CTL_DATA_ONSTACK,
0x40 /* control word for "get MS info" */);
if (error)
return (error);
cte = &toc.entries[0];
if (periph->periph_quirks & PQUIRK_LITTLETOC) {
cte->addr.lba = le32toh(cte->addr.lba);
toc.header.len = le16toh(toc.header.len);
} else {
cte->addr.lba = be32toh(cte->addr.lba);
toc.header.len = be16toh(toc.header.len);
}
*addr = (toc.header.len >= 10 && cte->track > 1) ?
cte->addr.lba : 0;
return 0;
}
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
Implement DIOCCACHESYNC for SCSI/ATAPI cd/mmc recordable devices found by scsipi/cd.c by issueing the SYNCHRONISE CACHES scsi/atapi call as defined per MMC standard.
2005-09-06 01:16:24 +04:00
/* synchronise caches code from sd.c, move to scsipi_ioctl.c ? */
static int
cdcachesync(struct scsipi_periph *periph, int flags) {
struct scsi_synchronize_cache_10 cmd;
/*
* Issue a SYNCHRONIZE CACHE. MMC devices have to issue with address 0
* and length 0 as it can't synchronise parts of the disc per spec.
* We ignore ILLEGAL REQUEST in the event that the command is not
* supported by the device, and poll for completion so that we know
* that the cache has actually been flushed.
*
* XXX should we handle the PQUIRK_NOSYNCCACHE ?
*/
memset(&cmd, 0, sizeof(cmd));
cmd.opcode = SCSI_SYNCHRONIZE_CACHE_10;
return (scsipi_command(periph, (void *)&cmd, sizeof(cmd), 0, 0,
CDRETRIES, 30000, NULL, flags | XS_CTL_IGNORE_ILLEGAL_REQUEST));
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* Perform special action on behalf of the user.
* Knows about the internals of this device
*/
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
cdioctl(dev_t dev, u_long cmd, caddr_t addr, int flag, struct lwp *l)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
{
New device attachment scheme: - split softc size and match/attach out from cfdriver into a new struct cfattach. - new "attach" directive for files.*. May specify the name of the cfattach structure, so that devices may be easily attached to parents with different autoconfiguration semantics.
1996-03-17 03:57:14 +03:00
struct cd_softc *cd = cd_cd.cd_devs[CDUNIT(dev)];
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
struct scsipi_periph *periph = cd->sc_periph;
Change DIOCEJECT to do what's needed to eject a device before the eject command (unlock for sd and cd) if no other partitions are open, return EBUSY otherwise. DIOCEJECT will have the old semantic if its argument is not 0. The old ioctl has been renamed to ODIOCEJECT for binary compatibility.
1999-02-08 19:33:16 +03:00
int part = CDPART(dev);
Don't waste too much kernel stack space on (infrequent) ioctl operations, use malloc instead for temp space.
2002-12-15 04:55:44 +03:00
int error = 0;
Adapt all disk devices in MI directories to handle ODIOC* calls for ports that have bumped MAXPARTITIONS (and thus define __HAVE_OLD_DISKLABEL).
2001-01-07 21:09:01 +03:00
#ifdef __HAVE_OLD_DISKLABEL
Don't waste too much kernel stack space on (infrequent) ioctl operations, use malloc instead for temp space.
2002-12-15 04:55:44 +03:00
struct disklabel *newlabel = NULL;
Adapt all disk devices in MI directories to handle ODIOC* calls for ports that have bumped MAXPARTITIONS (and thus define __HAVE_OLD_DISKLABEL).
2001-01-07 21:09:01 +03:00
#endif
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
SC_DEBUG(cd->sc_periph, SCSIPI_DB2, ("cdioctl 0x%lx ", cmd));
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
Allows the raw partition to be open()'ed, even when scsipi_start() fails (no media or other ...) so that we can always send ioctl's to the device.
1999-01-26 16:59:44 +03:00
* If the device is not valid, some IOCTLs can still be
* handled on the raw partition. Check this here.
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
*/
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
if ((periph->periph_flags & PERIPH_MEDIA_LOADED) == 0) {
Allows the raw partition to be open()'ed, even when scsipi_start() fails (no media or other ...) so that we can always send ioctl's to the device.
1999-01-26 16:59:44 +03:00
switch (cmd) {
case DIOCWLABEL:
case DIOCLOCK:
Change DIOCEJECT to do what's needed to eject a device before the eject command (unlock for sd and cd) if no other partitions are open, return EBUSY otherwise. DIOCEJECT will have the old semantic if its argument is not 0. The old ioctl has been renamed to ODIOCEJECT for binary compatibility.
1999-02-08 19:33:16 +03:00
case ODIOCEJECT:
Allows the raw partition to be open()'ed, even when scsipi_start() fails (no media or other ...) so that we can always send ioctl's to the device.
1999-01-26 16:59:44 +03:00
case DIOCEJECT:
Implement DIOCCACHESYNC for SCSI/ATAPI cd/mmc recordable devices found by scsipi/cd.c by issueing the SYNCHRONISE CACHES scsi/atapi call as defined per MMC standard.
2005-09-06 01:16:24 +04:00
case DIOCCACHESYNC:
Allows the raw partition to be open()'ed, even when scsipi_start() fails (no media or other ...) so that we can always send ioctl's to the device.
1999-01-26 16:59:44 +03:00
case SCIOCIDENTIFY:
case OSCIOCIDENTIFY:
case SCIOCCOMMAND:
case SCIOCDEBUG:
case CDIOCGETVOL:
case CDIOCSETVOL:
case CDIOCSETMONO:
case CDIOCSETSTEREO:
case CDIOCSETMUTE:
case CDIOCSETLEFT:
case CDIOCSETRIGHT:
case CDIOCCLOSE:
Return ENODEV instead of EIO when we are trying to open a device without media in the drive. restrict "opening of empty drive" to character devices only (reading a block device returns a short read instead of ENODEV, which can lead to confusion).
1999-01-29 14:17:58 +03:00
case CDIOCEJECT:
Allows the raw partition to be open()'ed, even when scsipi_start() fails (no media or other ...) so that we can always send ioctl's to the device.
1999-01-26 16:59:44 +03:00
case CDIOCALLOW:
case CDIOCPREVENT:
case CDIOCSETDEBUG:
case CDIOCCLRDEBUG:
case CDIOCRESET:
case SCIOCRESET:
case CDIOCLOADUNLOAD:
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
case DVD_AUTH:
case DVD_READ_STRUCT:
Change DIOCEJECT to do what's needed to eject a device before the eject command (unlock for sd and cd) if no other partitions are open, return EBUSY otherwise. DIOCEJECT will have the old semantic if its argument is not 0. The old ioctl has been renamed to ODIOCEJECT for binary compatibility.
1999-02-08 19:33:16 +03:00
if (part == RAW_PART)
Allows the raw partition to be open()'ed, even when scsipi_start() fails (no media or other ...) so that we can always send ioctl's to the device.
1999-01-26 16:59:44 +03:00
break;
/* FALLTHROUGH */
default:
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
if ((periph->periph_flags & PERIPH_OPEN) == 0)
Return ENODEV instead of EIO when we are trying to open a device without media in the drive. restrict "opening of empty drive" to character devices only (reading a block device returns a short read instead of ENODEV, which can lead to confusion).
1999-01-29 14:17:58 +03:00
return (ENODEV);
else
return (EIO);
Allows the raw partition to be open()'ed, even when scsipi_start() fails (no media or other ...) so that we can always send ioctl's to the device.
1999-01-26 16:59:44 +03:00
}
}
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
switch (cmd) {
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
case DIOCGDINFO:
New generic disk framework. Highlights: - New metrics handling. Metrics are now kept in the new `struct disk'. Busy time is now stored as a timeval, and transfer count in bytes. - Storage for disklabels is now dynamically allocated, so that the size of the disk structure is not machine-dependent. - Several new functions for attaching and detaching disks, and handling metrics calculation. Old-style instrumentation is still supported in drivers that did it before. However, old-style instrumentation is being deprecated, and will go away once the userland utilities are updated for the new framework. For usage and architectural details, see the forthcoming disk(9) manual page.
1996-01-08 01:01:38 +03:00
*(struct disklabel *)addr = *(cd->sc_dk.dk_label);
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (0);
Adapt all disk devices in MI directories to handle ODIOC* calls for ports that have bumped MAXPARTITIONS (and thus define __HAVE_OLD_DISKLABEL).
2001-01-07 21:09:01 +03:00
#ifdef __HAVE_OLD_DISKLABEL
case ODIOCGDINFO:
Don't waste too much kernel stack space on (infrequent) ioctl operations, use malloc instead for temp space.
2002-12-15 04:55:44 +03:00
newlabel = malloc(sizeof (*newlabel), M_TEMP, M_WAITOK);
if (newlabel == NULL)
return (EIO);
memcpy(newlabel, cd->sc_dk.dk_label, sizeof (*newlabel));
if (newlabel->d_npartitions > OLDMAXPARTITIONS)
error = ENOTTY;
else
memcpy(addr, newlabel, sizeof (struct olddisklabel));
free(newlabel, M_TEMP);
return error;
Adapt all disk devices in MI directories to handle ODIOC* calls for ports that have bumped MAXPARTITIONS (and thus define __HAVE_OLD_DISKLABEL).
2001-01-07 21:09:01 +03:00
#endif
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
case DIOCGPART:
New generic disk framework. Highlights: - New metrics handling. Metrics are now kept in the new `struct disk'. Busy time is now stored as a timeval, and transfer count in bytes. - Storage for disklabels is now dynamically allocated, so that the size of the disk structure is not machine-dependent. - Several new functions for attaching and detaching disks, and handling metrics calculation. Old-style instrumentation is still supported in drivers that did it before. However, old-style instrumentation is being deprecated, and will go away once the userland utilities are updated for the new framework. For usage and architectural details, see the forthcoming disk(9) manual page.
1996-01-08 01:01:38 +03:00
((struct partinfo *)addr)->disklab = cd->sc_dk.dk_label;
Similar changes to sd.c.
1994-10-20 17:03:38 +03:00
((struct partinfo *)addr)->part =
Change DIOCEJECT to do what's needed to eject a device before the eject command (unlock for sd and cd) if no other partitions are open, return EBUSY otherwise. DIOCEJECT will have the old semantic if its argument is not 0. The old ioctl has been renamed to ODIOCEJECT for binary compatibility.
1999-02-08 19:33:16 +03:00
&cd->sc_dk.dk_label->d_partitions[part];
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (0);
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
case DIOCWDINFO:
case DIOCSDINFO:
Adapt all disk devices in MI directories to handle ODIOC* calls for ports that have bumped MAXPARTITIONS (and thus define __HAVE_OLD_DISKLABEL).
2001-01-07 21:09:01 +03:00
#ifdef __HAVE_OLD_DISKLABEL
case ODIOCWDINFO:
case ODIOCSDINFO:
#endif
{
struct disklabel *lp;
Don't waste too much kernel stack space on (infrequent) ioctl operations, use malloc instead for temp space.
2002-12-15 04:55:44 +03:00
if ((flag & FWRITE) == 0)
return (EBADF);
Adapt all disk devices in MI directories to handle ODIOC* calls for ports that have bumped MAXPARTITIONS (and thus define __HAVE_OLD_DISKLABEL).
2001-01-07 21:09:01 +03:00
#ifdef __HAVE_OLD_DISKLABEL
if (cmd == ODIOCSDINFO || cmd == ODIOCWDINFO) {
Don't waste too much kernel stack space on (infrequent) ioctl operations, use malloc instead for temp space.
2002-12-15 04:55:44 +03:00
newlabel = malloc(sizeof (*newlabel), M_TEMP, M_WAITOK);
if (newlabel == NULL)
return (EIO);
memset(newlabel, 0, sizeof newlabel);
memcpy(newlabel, addr, sizeof (struct olddisklabel));
lp = newlabel;
Adapt all disk devices in MI directories to handle ODIOC* calls for ports that have bumped MAXPARTITIONS (and thus define __HAVE_OLD_DISKLABEL).
2001-01-07 21:09:01 +03:00
} else
#endif
lp = (struct disklabel *)addr;
Do not manipulate xs->bp in "generic" code -- do it only in the psw_done routine. As part of this, pass down our pre-parsed error code -- though this interface will probably change later to accomodate better error handling.
2004-09-18 03:10:50 +04:00
if ((error = lockmgr(&cd->sc_lock, LK_EXCLUSIVE, NULL)) != 0)
Don't waste too much kernel stack space on (infrequent) ioctl operations, use malloc instead for temp space.
2002-12-15 04:55:44 +03:00
goto bad;
Revamp the locking mechanism slightly.
1995-03-30 03:04:39 +04:00
cd->flags |= CDF_LABELLING;
Allow DIOCWDINFO even if label is not `writable'.
1995-03-23 14:33:18 +03:00
New generic disk framework. Highlights: - New metrics handling. Metrics are now kept in the new `struct disk'. Busy time is now stored as a timeval, and transfer count in bytes. - Storage for disklabels is now dynamically allocated, so that the size of the disk structure is not machine-dependent. - Several new functions for attaching and detaching disks, and handling metrics calculation. Old-style instrumentation is still supported in drivers that did it before. However, old-style instrumentation is being deprecated, and will go away once the userland utilities are updated for the new framework. For usage and architectural details, see the forthcoming disk(9) manual page.
1996-01-08 01:01:38 +03:00
error = setdisklabel(cd->sc_dk.dk_label,
Adapt all disk devices in MI directories to handle ODIOC* calls for ports that have bumped MAXPARTITIONS (and thus define __HAVE_OLD_DISKLABEL).
2001-01-07 21:09:01 +03:00
lp, /*cd->sc_dk.dk_openmask : */0,
New generic disk framework. Highlights: - New metrics handling. Metrics are now kept in the new `struct disk'. Busy time is now stored as a timeval, and transfer count in bytes. - Storage for disklabels is now dynamically allocated, so that the size of the disk structure is not machine-dependent. - Several new functions for attaching and detaching disks, and handling metrics calculation. Old-style instrumentation is still supported in drivers that did it before. However, old-style instrumentation is being deprecated, and will go away once the userland utilities are updated for the new framework. For usage and architectural details, see the forthcoming disk(9) manual page.
1996-01-08 01:01:38 +03:00
cd->sc_dk.dk_cpulabel);
Rearrange DIOCWDINFO a little; closer to wd.c.
1995-03-23 14:43:09 +03:00
if (error == 0) {
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
/* XXX ? */
Rearrange DIOCWDINFO a little; closer to wd.c.
1995-03-23 14:43:09 +03:00
}
Allow DIOCWDINFO even if label is not `writable'.
1995-03-23 14:33:18 +03:00
cd->flags &= ~CDF_LABELLING;
Do not manipulate xs->bp in "generic" code -- do it only in the psw_done routine. As part of this, pass down our pre-parsed error code -- though this interface will probably change later to accomodate better error handling.
2004-09-18 03:10:50 +04:00
lockmgr(&cd->sc_lock, LK_RELEASE, NULL);
Don't waste too much kernel stack space on (infrequent) ioctl operations, use malloc instead for temp space.
2002-12-15 04:55:44 +03:00
bad:
#ifdef __HAVE_OLD_DISKLABEL
if (newlabel != NULL)
free(newlabel, M_TEMP);
#endif
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (error);
Adapt all disk devices in MI directories to handle ODIOC* calls for ports that have bumped MAXPARTITIONS (and thus define __HAVE_OLD_DISKLABEL).
2001-01-07 21:09:01 +03:00
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
case DIOCWLABEL:
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (EBADF);
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
Implement DIOCGDEFLABEL.
1997-10-09 03:05:22 +04:00
case DIOCGDEFLABEL:
cdgetdefaultlabel(cd, (struct disklabel *)addr);
return (0);
Adapt all disk devices in MI directories to handle ODIOC* calls for ports that have bumped MAXPARTITIONS (and thus define __HAVE_OLD_DISKLABEL).
2001-01-07 21:09:01 +03:00
#ifdef __HAVE_OLD_DISKLABEL
case ODIOCGDEFLABEL:
Don't waste too much kernel stack space on (infrequent) ioctl operations, use malloc instead for temp space.
2002-12-15 04:55:44 +03:00
newlabel = malloc(sizeof (*newlabel), M_TEMP, M_WAITOK);
if (newlabel == NULL)
return (EIO);
cdgetdefaultlabel(cd, newlabel);
if (newlabel->d_npartitions > OLDMAXPARTITIONS)
error = ENOTTY;
else
memcpy(addr, newlabel, sizeof (struct olddisklabel));
free(newlabel, M_TEMP);
return error;
Adapt all disk devices in MI directories to handle ODIOC* calls for ports that have bumped MAXPARTITIONS (and thus define __HAVE_OLD_DISKLABEL).
2001-01-07 21:09:01 +03:00
#endif
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
case CDIOCPLAYTRACKS: {
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* PLAY_MSF command */
Similar changes to sd.c.
1994-10-20 17:03:38 +03:00
struct ioc_play_track *args = (struct ioc_play_track *)addr;
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
if ((error = cd_set_pa_immed(cd, 0)) != 0)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (error);
return (cd_play_tracks(cd, args->start_track,
args->start_index, args->end_track, args->end_index));
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
}
case CDIOCPLAYMSF: {
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* PLAY_MSF command */
scsi prototypes
1996-02-15 00:46:52 +03:00
struct ioc_play_msf *args = (struct ioc_play_msf *)addr;
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
if ((error = cd_set_pa_immed(cd, 0)) != 0)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (error);
return (cd_play_msf(cd, args->start_m, args->start_s,
args->start_f, args->end_m, args->end_s, args->end_f));
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
}
case CDIOCPLAYBLOCKS: {
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* PLAY command */
scsi prototypes
1996-02-15 00:46:52 +03:00
struct ioc_play_blocks *args = (struct ioc_play_blocks *)addr;
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
if ((error = cd_set_pa_immed(cd, 0)) != 0)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (error);
return (cd_play(cd, args->blk, args->len));
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
}
case CDIOCREADSUBCHANNEL: {
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* READ_SUBCHANNEL command */
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
struct ioc_read_subchannel *args =
(struct ioc_read_subchannel *)addr;
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
struct cd_sub_channel_info data;
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
u_int len = args->data_len;
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
if (len > sizeof(data) ||
len < sizeof(struct cd_sub_channel_header))
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (EINVAL);
scsi prototypes
1996-02-15 00:46:52 +03:00
error = cd_read_subchannel(cd, args->address_format,
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
args->data_format, args->track, &data, len,
XS_CTL_DATA_ONSTACK);
scsi prototypes
1996-02-15 00:46:52 +03:00
if (error)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (error);
Define a full set of [234][bl]tol() and lto[234][bl]() conversion functions, inlined. Use sized types in protocol structures. Make the definition of scsi_sense_data less ugly.
1996-03-19 06:05:15 +03:00
len = min(len, _2btol(data.header.data_len) +
sizeof(struct cd_sub_channel_header));
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (copyout(&data, args->data, len));
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
}
case CDIOREADTOCHEADER: {
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* READ TOC format 0 command, static header */
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
struct ioc_toc_header th;
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
Second stage of read TOC modifications. Introducing the various forms of TOC reading. Note that the external interfaces haven't changed. Only the formatted toc is requestable. The read msinfo command hasn't been changed in this patch though it could become more clever using the CD_TOC_MSINFO format.
2005-12-11 00:17:21 +03:00
if ((error = cd_read_toc(cd, CD_TOC_FORM, 0, 0, &th, sizeof(th),
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
XS_CTL_DATA_ONSTACK, 0)) != 0)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (error);
Use be16toh(), be32toh(), le16toh() and le32toh() rather than ntohs(), ntohl() or homegrown bswap() functions. XXX Does anyone use drives with PQUIRK_LITTLETOC on big endian machines?
2001-09-02 17:11:53 +04:00
if (cd->sc_periph->periph_quirks & PQUIRK_LITTLETOC)
th.len = le16toh(th.len);
else
th.len = be16toh(th.len);
bcopy -> memcpy
2001-07-18 22:25:41 +04:00
memcpy(addr, &th, sizeof(th));
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (0);
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
}
case CDIOREADTOCENTRYS: {
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* READ TOC format 0 command, entries */
Second stage of read TOC modifications. Introducing the various forms of TOC reading. Note that the external interfaces haven't changed. Only the formatted toc is requestable. The read msinfo command hasn't been changed in this patch though it could become more clever using the CD_TOC_MSINFO format.
2005-12-11 00:17:21 +03:00
struct cd_formatted_toc toc;
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
struct ioc_read_toc_entry *te =
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
(struct ioc_read_toc_entry *)addr;
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
struct ioc_toc_header *th;
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
struct cd_toc_entry *cte;
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
u_int len = te->data_len;
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
int ntracks;
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
th = &toc.header;
if (len > sizeof(toc.entries) ||
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
len < sizeof(struct cd_toc_entry))
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (EINVAL);
Second stage of read TOC modifications. Introducing the various forms of TOC reading. Note that the external interfaces haven't changed. Only the formatted toc is requestable. The read msinfo command hasn't been changed in this patch though it could become more clever using the CD_TOC_MSINFO format.
2005-12-11 00:17:21 +03:00
error = cd_read_toc(cd, CD_TOC_FORM, te->address_format,
te->starting_track, &toc,
len + sizeof(struct ioc_toc_header),
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
XS_CTL_DATA_ONSTACK, 0);
scsi prototypes
1996-02-15 00:46:52 +03:00
if (error)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (error);
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
if (te->address_format == CD_LBA_FORMAT)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
for (ntracks =
th->ending_track - th->starting_track + 1;
ntracks >= 0; ntracks--) {
cte = &toc.entries[ntracks];
cte->addr_type = CD_LBA_FORMAT;
Use be16toh(), be32toh(), le16toh() and le32toh() rather than ntohs(), ntohl() or homegrown bswap() functions. XXX Does anyone use drives with PQUIRK_LITTLETOC on big endian machines?
2001-09-02 17:11:53 +04:00
if (periph->periph_quirks & PQUIRK_LITTLETOC)
cte->addr.lba = le32toh(cte->addr.lba);
else
cte->addr.lba = be32toh(cte->addr.lba);
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
}
Use be16toh(), be32toh(), le16toh() and le32toh() rather than ntohs(), ntohl() or homegrown bswap() functions. XXX Does anyone use drives with PQUIRK_LITTLETOC on big endian machines?
2001-09-02 17:11:53 +04:00
if (periph->periph_quirks & PQUIRK_LITTLETOC)
th->len = le16toh(th->len);
else
th->len = be16toh(th->len);
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
len = min(len, th->len - (sizeof(th->starting_track) +
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
sizeof(th->ending_track)));
return (copyout(toc.entries, te->data, len));
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
}
Multisession-CD-Support, by Torsten Duwe <duwe@ns.lst.de>.
1999-01-04 18:32:08 +03:00
case CDIOREADMSADDR: {
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* READ TOC format 0 command, length of first track only */
Multisession-CD-Support, by Torsten Duwe <duwe@ns.lst.de>.
1999-01-04 18:32:08 +03:00
int sessno = *(int*)addr;
if (sessno != 0)
return (EINVAL);
put multisession offset code into a separate function and initialize the p_cdsession field of partition 'a' in the default disklabel
2002-05-27 20:42:30 +04:00
return (cdreadmsaddr(cd, (int*)addr));
Multisession-CD-Support, by Torsten Duwe <duwe@ns.lst.de>.
1999-01-04 18:32:08 +03:00
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
case CDIOCSETPATCH: {
Similar changes to sd.c.
1994-10-20 17:03:38 +03:00
struct ioc_patch *arg = (struct ioc_patch *)addr;
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (cd_setchan(cd, arg->patch[0], arg->patch[1],
arg->patch[2], arg->patch[3], 0));
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
}
case CDIOCGETVOL: {
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* MODE SENSE command (AUDIO page) */
Similar changes to sd.c.
1994-10-20 17:03:38 +03:00
struct ioc_vol *arg = (struct ioc_vol *)addr;
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (cd_getvol(cd, arg, 0));
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
}
case CDIOCSETVOL: {
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* MODE SENSE/MODE SELECT commands (AUDIO page) */
Similar changes to sd.c.
1994-10-20 17:03:38 +03:00
struct ioc_vol *arg = (struct ioc_vol *)addr;
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (cd_setvol(cd, arg, 0));
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
}
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
case CDIOCSETMONO:
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* MODE SENSE/MODE SELECT commands (AUDIO page) */
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (cd_setchan(cd, BOTH_CHANNEL, BOTH_CHANNEL,
MUTE_CHANNEL, MUTE_CHANNEL, 0));
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
case CDIOCSETSTEREO:
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* MODE SENSE/MODE SELECT commands (AUDIO page) */
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (cd_setchan(cd, LEFT_CHANNEL, RIGHT_CHANNEL,
MUTE_CHANNEL, MUTE_CHANNEL, 0));
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
case CDIOCSETMUTE:
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* MODE SENSE/MODE SELECT commands (AUDIO page) */
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (cd_setchan(cd, MUTE_CHANNEL, MUTE_CHANNEL,
MUTE_CHANNEL, MUTE_CHANNEL, 0));
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
case CDIOCSETLEFT:
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* MODE SENSE/MODE SELECT commands (AUDIO page) */
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (cd_setchan(cd, LEFT_CHANNEL, LEFT_CHANNEL,
MUTE_CHANNEL, MUTE_CHANNEL, 0));
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
case CDIOCSETRIGHT:
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* MODE SENSE/MODE SELECT commands (AUDIO page) */
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (cd_setchan(cd, RIGHT_CHANNEL, RIGHT_CHANNEL,
MUTE_CHANNEL, MUTE_CHANNEL, 0));
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
case CDIOCRESUME:
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* PAUSE command */
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (cd_pause(cd, PA_RESUME));
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
case CDIOCPAUSE:
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* PAUSE command */
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (cd_pause(cd, PA_PAUSE));
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
case CDIOCSTART:
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
return (scsipi_start(periph, SSS_START, 0));
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
case CDIOCSTOP:
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
return (scsipi_start(periph, SSS_STOP, 0));
Implement CDIOCCLOSE.
1998-03-16 07:17:53 +03:00
case CDIOCCLOSE:
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
return (scsipi_start(periph, SSS_START|SSS_LOEJ,
Cleanup the scsipi_xfer flags: - `flags' is now gone, replaced with `xs_control' and `xs_status'. - Massive cleanup of the control flags. Now we explicitly say that a job is to complete asynchronously, rather than relying on side-effects, and use a new flag to now that device discovery is being performed. - Do SCSI device discovery interrupt-driven.
1999-10-01 02:57:52 +04:00
XS_CTL_IGNORE_NOT_READY | XS_CTL_IGNORE_MEDIA_CHANGE));
Accept DIOCEJECT as a synonym for CDIOCEJECT. Implement DIOCLOCK separately from CDIOCALLOW and CDIOCPREVENT, even though they perform basically the same function (with a different interface XXX).
1996-01-30 21:28:02 +03:00
case DIOCEJECT:
Change DIOCEJECT to do what's needed to eject a device before the eject command (unlock for sd and cd) if no other partitions are open, return EBUSY otherwise. DIOCEJECT will have the old semantic if its argument is not 0. The old ioctl has been renamed to ODIOCEJECT for binary compatibility.
1999-02-08 19:33:16 +03:00
if (*(int *)addr == 0) {
/*
* Don't force eject: check that we are the only
* partition open. If so, unlock it.
*/
if ((cd->sc_dk.dk_openmask & ~(1 << part)) == 0 &&
cd->sc_dk.dk_bopenmask + cd->sc_dk.dk_copenmask ==
cd->sc_dk.dk_openmask) {
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
error = scsipi_prevent(periph, SPAMR_ALLOW,
Cleanup the scsipi_xfer flags: - `flags' is now gone, replaced with `xs_control' and `xs_status'. - Massive cleanup of the control flags. Now we explicitly say that a job is to complete asynchronously, rather than relying on side-effects, and use a new flag to now that device discovery is being performed. - Do SCSI device discovery interrupt-driven.
1999-10-01 02:57:52 +04:00
XS_CTL_IGNORE_NOT_READY);
Change DIOCEJECT to do what's needed to eject a device before the eject command (unlock for sd and cd) if no other partitions are open, return EBUSY otherwise. DIOCEJECT will have the old semantic if its argument is not 0. The old ioctl has been renamed to ODIOCEJECT for binary compatibility.
1999-02-08 19:33:16 +03:00
if (error)
return (error);
} else {
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
return (EBUSY);
Change DIOCEJECT to do what's needed to eject a device before the eject command (unlock for sd and cd) if no other partitions are open, return EBUSY otherwise. DIOCEJECT will have the old semantic if its argument is not 0. The old ioctl has been renamed to ODIOCEJECT for binary compatibility.
1999-02-08 19:33:16 +03:00
}
}
/* FALLTHROUGH */
case CDIOCEJECT: /* FALLTHROUGH */
case ODIOCEJECT:
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
return (scsipi_start(periph, SSS_STOP|SSS_LOEJ, 0));
Implement DIOCCACHESYNC for SCSI/ATAPI cd/mmc recordable devices found by scsipi/cd.c by issueing the SYNCHRONISE CACHES scsi/atapi call as defined per MMC standard.
2005-09-06 01:16:24 +04:00
case DIOCCACHESYNC:
/* SYNCHRONISE CACHES command */
return (cdcachesync(periph, 0));
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
case CDIOCALLOW:
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
return (scsipi_prevent(periph, SPAMR_ALLOW, 0));
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
case CDIOCPREVENT:
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
return (scsipi_prevent(periph, SPAMR_PREVENT_DT, 0));
Accept DIOCEJECT as a synonym for CDIOCEJECT. Implement DIOCLOCK separately from CDIOCALLOW and CDIOCPREVENT, even though they perform basically the same function (with a different interface XXX).
1996-01-30 21:28:02 +03:00
case DIOCLOCK:
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
return (scsipi_prevent(periph,
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
(*(int *)addr) ? SPAMR_PREVENT_DT : SPAMR_ALLOW, 0));
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
case CDIOCSETDEBUG:
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
cd->sc_periph->periph_dbflags |= (SCSIPI_DB1 | SCSIPI_DB2);
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (0);
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
case CDIOCCLRDEBUG:
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
cd->sc_periph->periph_dbflags &= ~(SCSIPI_DB1 | SCSIPI_DB2);
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (0);
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
case CDIOCRESET:
Allows the raw partition to be open()'ed, even when scsipi_start() fails (no media or other ...) so that we can always send ioctl's to the device.
1999-01-26 16:59:44 +03:00
case SCIOCRESET:
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (cd_reset(cd));
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
case CDIOCLOADUNLOAD:
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* LOAD_UNLOAD command */
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (cd_load_unload(cd, (struct ioc_load_unload *)addr));
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
case DVD_AUTH:
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* GPCMD_REPORT_KEY or GPCMD_SEND_KEY command */
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
return (dvd_auth(cd, (dvd_authinfo *)addr));
case DVD_READ_STRUCT:
Document executed SCSI commands with the IOCTL's to prevent confusion.
2005-08-29 02:51:01 +04:00
/* GPCMD_READ_DVD_STRUCTURE command */
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
return (dvd_read_struct(cd, (dvd_struct *)addr));
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
default:
Change DIOCEJECT to do what's needed to eject a device before the eject command (unlock for sd and cd) if no other partitions are open, return EBUSY otherwise. DIOCEJECT will have the old semantic if its argument is not 0. The old ioctl has been renamed to ODIOCEJECT for binary compatibility.
1999-02-08 19:33:16 +03:00
if (part != RAW_PART)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (ENOTTY);
merge ktrace-lwp.
2005-12-11 15:16:03 +03:00
return (scsipi_do_ioctl(periph, dev, cmd, addr, flag, l));
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
Add some missing locking, and some general cleanup.
1994-11-21 01:36:43 +03:00
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
#ifdef DIAGNOSTIC
panic("cdioctl: impossible");
#endif
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static void
cdgetdefaultlabel(struct cd_softc *cd, struct disklabel *lp)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
{
put multisession offset code into a separate function and initialize the p_cdsession field of partition 'a' in the default disklabel
2002-05-27 20:42:30 +04:00
int lastsession;
Add some missing locking, and some general cleanup.
1994-11-21 01:36:43 +03:00
bzero -> memset
2001-07-18 22:21:04 +04:00
memset(lp, 0, sizeof(struct disklabel));
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
Use an intermediate variable to shorten label initialization code.
1995-08-06 03:47:52 +04:00
lp->d_secsize = cd->params.blksize;
lp->d_ntracks = 1;
lp->d_nsectors = 100;
lp->d_ncylinders = (cd->params.disksize / 100) + 1;
lp->d_secpercyl = lp->d_ntracks * lp->d_nsectors;
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
switch (scsipi_periph_bustype(cd->sc_periph)) {
case SCSIPI_BUSTYPE_SCSI:
Improve generation of default disklabels: -store printable product ID in cd's and sd's softc, use it as "typename" -for this, add a "destination buffer length" argument to scsipi_strvis() -return ATAPI device type for ATAPI devices
1998-08-05 20:29:04 +04:00
lp->d_type = DTYPE_SCSI;
break;
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
case SCSIPI_BUSTYPE_ATAPI:
Improve generation of default disklabels: -store printable product ID in cd's and sd's softc, use it as "typename" -for this, add a "destination buffer length" argument to scsipi_strvis() -return ATAPI device type for ATAPI devices
1998-08-05 20:29:04 +04:00
lp->d_type = DTYPE_ATAPI;
break;
}
Eliminate the separate ATAPI and SCSI attachments for "cd".
2003-09-08 05:56:33 +04:00
/*
* XXX
* We could probe the mode pages to figure out what kind of disc it is.
* Is this worthwhile?
*/
Merge the geometry and cache handling code for all direct access and optical devices, as it's general to all SCSI MMC devices. In the process, remove PQUIRK_NO_FLEX_PAGE.
2003-09-18 04:06:31 +04:00
strncpy(lp->d_typename, "mydisc", 16);
Use an intermediate variable to shorten label initialization code.
1995-08-06 03:47:52 +04:00
strncpy(lp->d_packname, "fictitious", 16);
lp->d_secperunit = cd->params.disksize;
lp->d_rpm = 300;
lp->d_interleave = 1;
lp->d_flags = D_REMOVABLE;
put multisession offset code into a separate function and initialize the p_cdsession field of partition 'a' in the default disklabel
2002-05-27 20:42:30 +04:00
if (cdreadmsaddr(cd, &lastsession) != 0)
lastsession = 0;
Use an intermediate variable to shorten label initialization code.
1995-08-06 03:47:52 +04:00
lp->d_partitions[0].p_offset = 0;
put multisession offset code into a separate function and initialize the p_cdsession field of partition 'a' in the default disklabel
2002-05-27 20:42:30 +04:00
#ifdef notyet /* have to fix bounds_check_with_label() first */
lp->d_partitions[0].p_size = lp->d_secperunit;
#else
Use an intermediate variable to shorten label initialization code.
1995-08-06 03:47:52 +04:00
lp->d_partitions[0].p_size =
lp->d_secperunit * (lp->d_secsize / DEV_BSIZE);
put multisession offset code into a separate function and initialize the p_cdsession field of partition 'a' in the default disklabel
2002-05-27 20:42:30 +04:00
#endif
lp->d_partitions[0].p_cdsession = lastsession;
Use an intermediate variable to shorten label initialization code.
1995-08-06 03:47:52 +04:00
lp->d_partitions[0].p_fstype = FS_ISO9660;
lp->d_partitions[RAW_PART].p_offset = 0;
put multisession offset code into a separate function and initialize the p_cdsession field of partition 'a' in the default disklabel
2002-05-27 20:42:30 +04:00
#ifdef notyet
lp->d_partitions[RAW_PART].p_size = lp->d_secperunit;
#else
Use an intermediate variable to shorten label initialization code.
1995-08-06 03:47:52 +04:00
lp->d_partitions[RAW_PART].p_size =
lp->d_secperunit * (lp->d_secsize / DEV_BSIZE);
put multisession offset code into a separate function and initialize the p_cdsession field of partition 'a' in the default disklabel
2002-05-27 20:42:30 +04:00
#endif
Use an intermediate variable to shorten label initialization code.
1995-08-06 03:47:52 +04:00
lp->d_partitions[RAW_PART].p_fstype = FS_ISO9660;
lp->d_npartitions = RAW_PART + 1;
lp->d_magic = DISKMAGIC;
lp->d_magic2 = DISKMAGIC;
lp->d_checksum = dkcksum(lp);
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
Implement DIOCGDEFLABEL.
1997-10-09 03:05:22 +04:00
/*
* Load the label information on the named device
* Actually fabricate a disklabel
*
* EVENTUALLY take information about different
* data tracks from the TOC and put it in the disklabel
*/
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static void
cdgetdisklabel(struct cd_softc *cd)
Implement DIOCGDEFLABEL.
1997-10-09 03:05:22 +04:00
{
struct disklabel *lp = cd->sc_dk.dk_label;
Change return type of readdisklabel() to const char * I hope I've found all the correct places!
2003-05-02 12:45:10 +04:00
const char *errstring;
Add support for smaller sector sizes so we can mount ffs filesystems.
2001-08-16 02:21:01 +04:00
memset(cd->sc_dk.dk_cpulabel, 0, sizeof(struct cpu_disklabel));
cdgetdefaultlabel(cd, lp);
/*
* Call the generic disklabel extraction routine
*/
errstring = readdisklabel(MAKECDDEV(0, cd->sc_dev.dv_unit, RAW_PART),
cdstrategy, lp, cd->sc_dk.dk_cpulabel);
Implement DIOCGDEFLABEL.
1997-10-09 03:05:22 +04:00
Clean up cdgetdisklabel. It was jumping over a return...
2005-12-15 20:56:32 +03:00
/* if all went OK, we are passed a NULL error string */
if (errstring == NULL)
return;
Implement DIOCGDEFLABEL.
1997-10-09 03:05:22 +04:00
Clean up cdgetdisklabel. It was jumping over a return...
2005-12-15 20:56:32 +03:00
/* Reset to default label -- after printing error and the warning */
printf("%s: %s\n", cd->sc_dev.dv_xname, errstring);
memset(cd->sc_dk.dk_cpulabel, 0, sizeof(struct cpu_disklabel));
Implement DIOCGDEFLABEL.
1997-10-09 03:05:22 +04:00
cdgetdefaultlabel(cd, lp);
}
Cleanup and fix cd_size() function; it was returning wrong results. Also factor out read_cd_capacity()
2005-09-07 02:19:14 +04:00
static int
read_cd_capacity(struct scsipi_periph *periph, int *blksize, u_long *size)
{
struct scsipi_read_cd_capacity cmd;
struct scsipi_read_cd_cap_data data;
int error, flags;
/* if the device doesn't grog capacity, return the dummies */
if (periph->periph_quirks & PQUIRK_NOCAPACITY)
return 0;
/* issue the request */
flags = XS_CTL_DATA_IN | XS_CTL_DATA_ONSTACK;
memset(&cmd, 0, sizeof(cmd));
cmd.opcode = READ_CD_CAPACITY;
error = scsipi_command(periph,
(void *) &cmd, sizeof(cmd),
(void *) &data, sizeof(data),
CDRETRIES, 30000, NULL, flags);
if (error)
return error;
/* retrieve values and sanity check them */
*blksize = _4btol(data.length);
/* blksize is 2048 for CD, but some drives give gibberish */
if ((*blksize < 512) || ((*blksize & 511) != 0))
*blksize = 2048; /* some drives lie ! */
*size = _4btol(data.addr);
if (*size < 100)
*size = 400000;
return 0;
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* Find out from the device what it's capacity is
*/
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
static u_long
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
cd_size(struct cd_softc *cd, int flags)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
{
Numerous changes. Many bugs fixed, better autoconfig, a few new features.
1994-12-28 22:42:47 +03:00
int blksize;
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
u_long size;
Cleanup and fix cd_size() function; it was returning wrong results. Also factor out read_cd_capacity()
2005-09-07 02:19:14 +04:00
int error;
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
Cleanup and fix cd_size() function; it was returning wrong results. Also factor out read_cd_capacity()
2005-09-07 02:19:14 +04:00
/* set up fake values */
blksize = 2048;
size = 400000;
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
Cleanup and fix cd_size() function; it was returning wrong results. Also factor out read_cd_capacity()
2005-09-07 02:19:14 +04:00
/* if this function bounces with an error return fake value */
error = read_cd_capacity(cd->sc_periph, &blksize, &size);
if (error)
return size;
On probe or open, retry the READ CAPACITY once if the drive reports NOT READY.
1994-08-12 03:51:26 +04:00
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
if (blksize != 2048) {
if (cd_setblksize(cd) == 0)
blksize = 2048;
If a CD drive reports != 2048 byte block size, try to change it into 2048 byte mode. Fixes PR kern/22090. Reviewed by Manuel Bouyer.
2003-07-10 22:18:40 +04:00
}
Cleanup and fix cd_size() function; it was returning wrong results. Also factor out read_cd_capacity()
2005-09-07 02:19:14 +04:00
cd->params.blksize = blksize;
cd->params.disksize = size-1; /* disklabel is exclusive */
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
cd->params.disksize512 = ((u_int64_t)cd->params.disksize * blksize) / DEV_BSIZE;
On probe or open, retry the READ CAPACITY once if the drive reports NOT READY.
1994-08-12 03:51:26 +04:00
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
SC_DEBUG(cd->sc_periph, SCSIPI_DB2,
("cd_size: %d %ld\n", blksize, size));
Cleanup and fix cd_size() function; it was returning wrong results. Also factor out read_cd_capacity()
2005-09-07 02:19:14 +04:00
return size;
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* Get scsi driver to send a "start playing" command
*/
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cd_play(struct cd_softc *cd, int blkno, int nblks)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
{
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
struct scsipi_play cmd;
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
memset(&cmd, 0, sizeof(cmd));
cmd.opcode = PLAY;
_lto4b(blkno, cmd.blk_addr);
_lto2b(nblks, cmd.xfer_len);
return (scsipi_command(cd->sc_periph, (void *)&cmd, sizeof(cmd), 0, 0,
CDRETRIES, 30000, NULL, 0));
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* Get scsi driver to send a "start playing" command
*/
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cd_play_tracks(struct cd_softc *cd, int strack, int sindex, int etrack,
int eindex)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
{
Second stage of read TOC modifications. Introducing the various forms of TOC reading. Note that the external interfaces haven't changed. Only the formatted toc is requestable. The read msinfo command hasn't been changed in this patch though it could become more clever using the CD_TOC_MSINFO format.
2005-12-11 00:17:21 +03:00
struct cd_formatted_toc toc;
Fix illogical use of extra variable. It was mainly to get a better readability in KNF. This solution gives better readability without the extra variable.
2005-12-11 03:01:54 +03:00
int error;
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
if (!etrack)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (EIO);
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
if (strack > etrack)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (EINVAL);
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
Fix illogical use of extra variable. It was mainly to get a better readability in KNF. This solution gives better readability without the extra variable.
2005-12-11 03:01:54 +03:00
error = cd_load_toc(cd, CD_TOC_FORM, &toc, XS_CTL_DATA_ONSTACK);
if (error)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (error);
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
if (++etrack > (toc.header.ending_track+1))
etrack = toc.header.ending_track+1;
strack -= toc.header.starting_track;
etrack -= toc.header.starting_track;
if (strack < 0)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (EINVAL);
return (cd_play_msf(cd, toc.entries[strack].addr.msf.minute,
toc.entries[strack].addr.msf.second,
toc.entries[strack].addr.msf.frame,
toc.entries[etrack].addr.msf.minute,
toc.entries[etrack].addr.msf.second,
toc.entries[etrack].addr.msf.frame));
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* Get scsi driver to send a "play msf" command
*/
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cd_play_msf(struct cd_softc *cd, int startm, int starts, int startf, int endm,
int ends, int endf)
Add play_msf command and MSF variants. This is SCSI2 stuff and is needed for many flavors of player utilities.
1993-08-04 21:26:20 +04:00
{
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
struct scsipi_play_msf cmd;
memset(&cmd, 0, sizeof(cmd));
cmd.opcode = PLAY_MSF;
cmd.start_m = startm;
cmd.start_s = starts;
cmd.start_f = startf;
cmd.end_m = endm;
cmd.end_s = ends;
cmd.end_f = endf;
return (scsipi_command(cd->sc_periph, (void *)&cmd, sizeof(cmd), 0, 0,
CDRETRIES, 30000, NULL, 0));
Add play_msf command and MSF variants. This is SCSI2 stuff and is needed for many flavors of player utilities.
1993-08-04 21:26:20 +04:00
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* Get scsi driver to send a "start up" command
*/
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cd_pause(struct cd_softc *cd, int go)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
{
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
struct scsipi_pause cmd;
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
memset(&cmd, 0, sizeof(cmd));
cmd.opcode = PAUSE;
cmd.resume = go & 0xff;
return (scsipi_command(cd->sc_periph, (void *)&cmd, sizeof(cmd), 0, 0,
CDRETRIES, 30000, NULL, 0));
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* Get scsi driver to send a "RESET" command
*/
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cd_reset(struct cd_softc *cd)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
{
Remove the "xfer" argument to scsipi_command().
2004-09-18 03:43:17 +04:00
return (scsipi_command(cd->sc_periph, 0, 0, 0, 0,
Cleanup the scsipi_xfer flags: - `flags' is now gone, replaced with `xs_control' and `xs_status'. - Massive cleanup of the control flags. Now we explicitly say that a job is to complete asynchronously, rather than relying on side-effects, and use a new flag to now that device discovery is being performed. - Do SCSI device discovery interrupt-driven.
1999-10-01 02:57:52 +04:00
CDRETRIES, 30000, NULL, XS_CTL_RESET));
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
}
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* Read subchannel
*/
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cd_read_subchannel(struct cd_softc *cd, int mode, int format, int track,
struct cd_sub_channel_info *data, int len, int flags)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
{
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
struct scsipi_read_subchannel cmd;
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
memset(&cmd, 0, sizeof(cmd));
cmd.opcode = READ_SUBCHANNEL;
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
if (mode == CD_MSF_FORMAT)
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
cmd.byte2 |= CD_MSF;
cmd.byte3 = SRS_SUBQ;
cmd.subchan_format = format;
cmd.track = track;
_lto2b(len, cmd.data_len);
Remove the "xfer" argument to scsipi_command().
2004-09-18 03:43:17 +04:00
return (scsipi_command(cd->sc_periph,
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
(void *)&cmd, sizeof(struct scsipi_read_subchannel),
(void *)data, len,
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
CDRETRIES, 30000, NULL, flags | XS_CTL_DATA_IN | XS_CTL_SILENT));
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* Read table of contents
*/
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
Second stage of read TOC modifications. Introducing the various forms of TOC reading. Note that the external interfaces haven't changed. Only the formatted toc is requestable. The read msinfo command hasn't been changed in this patch though it could become more clever using the CD_TOC_MSINFO format.
2005-12-11 00:17:21 +03:00
cd_read_toc(struct cd_softc *cd, int respf, int mode, int start, void *data, int len,
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
int flags, int control)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
{
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
struct scsipi_read_toc cmd;
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
int ntoc;
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
memset(&cmd, 0, sizeof(cmd));
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
#if 0
if (len != sizeof(struct ioc_toc_header))
ntoc = ((len) - sizeof(struct ioc_toc_header)) /
sizeof(struct cd_toc_entry);
else
#endif
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
ntoc = len;
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
cmd.opcode = READ_TOC;
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
if (mode == CD_MSF_FORMAT)
The code asumes all programs want to know is TOC response format 0. Support for TOC response format 1 and 2 are mandatory on CD/DVD too and provide more information. Next an IOCTL needs to be implemented that can read all TOC formats in a generic way. This is pending.
2005-01-31 02:39:42 +03:00
cmd.addr_mode |= CD_MSF;
Second stage of read TOC modifications. Introducing the various forms of TOC reading. Note that the external interfaces haven't changed. Only the formatted toc is requestable. The read msinfo command hasn't been changed in this patch though it could become more clever using the CD_TOC_MSINFO format.
2005-12-11 00:17:21 +03:00
cmd.resp_format = respf;
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
cmd.from_track = start;
_lto2b(ntoc, cmd.data_len);
cmd.control = control;
Remove the "xfer" argument to scsipi_command().
2004-09-18 03:43:17 +04:00
return (scsipi_command(cd->sc_periph,
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
(void *)&cmd, sizeof(cmd), (void *)data, len, CDRETRIES,
Pull up the thorpej_scsipi branch to main branch. This is a completely rewritten scsipi_xfer execution engine, and the associated changes to HBA drivers. Overview of changes & features: - All xfers are queued in the mid-layer, rather than doing so in an ad-hoc fashion in individual adapter drivers. - Adapter/channel resource management in the mid-layer, avoids even trying to start running an xfer if the adapter/channel doesn't have the resources. - Better communication between the mid-layer and the adapters. - Asynchronous event notification mechanism from adapter to mid-layer and peripherals. - Better peripheral queue management: freeze/thaw, sorted requeueing during recovery, etc. - Clean separation of peripherals, adapters, and adapter channels (no more scsipi_link). - Kernel thread for each scsipi_channel makes error recovery much easier (no more dealing with interrupt context when recovering from an error). - Mid-layer support for tagged queueing: commands can have the tag type set explicitly, tag IDs are allocated in the mid-layer (thus eliminating the need to use buggy tag ID allocation schemes in many adapter drivers). - support for QUEUE FULL and CHECK CONDITION status in mid-layer; the command will be requeued, or a REQUEST SENSE will be sent as appropriate. Just before the merge syssrc has been tagged with thorpej_scsipi_beforemerge
2001-04-25 21:53:04 +04:00
30000, NULL, flags | XS_CTL_DATA_IN));
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
}
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
Second stage of read TOC modifications. Introducing the various forms of TOC reading. Note that the external interfaces haven't changed. Only the formatted toc is requestable. The read msinfo command hasn't been changed in this patch though it could become more clever using the CD_TOC_MSINFO format.
2005-12-11 00:17:21 +03:00
cd_load_toc(struct cd_softc *cd, int respf, struct cd_formatted_toc *toc, int flags)
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
{
Uses PLAY_MSF instead of PLAY for the CDIOCPLAYTRACKS ioctl. This avoid a int16 overflow in the PLAY cbd when the resquested track(s) are more than 65535 blocks long. Submitted by t-nkyma@tcp-ip.or.jp in PR kern/4092.
1997-09-09 13:07:43 +04:00
int ntracks, len, error;
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
Second stage of read TOC modifications. Introducing the various forms of TOC reading. Note that the external interfaces haven't changed. Only the formatted toc is requestable. The read msinfo command hasn't been changed in this patch though it could become more clever using the CD_TOC_MSINFO format.
2005-12-11 00:17:21 +03:00
if ((error = cd_read_toc(cd, respf, 0, 0, toc, sizeof(toc->header),
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
flags, 0)) != 0)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (error);
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
ntracks = toc->header.ending_track - toc->header.starting_track + 1;
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
len = (ntracks + 1) * sizeof(struct cd_toc_entry) +
sizeof(toc->header);
Second stage of read TOC modifications. Introducing the various forms of TOC reading. Note that the external interfaces haven't changed. Only the formatted toc is requestable. The read msinfo command hasn't been changed in this patch though it could become more clever using the CD_TOC_MSINFO format.
2005-12-11 00:17:21 +03:00
if ((error = cd_read_toc(cd, respf, CD_MSF_FORMAT, 0, toc, len,
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
flags, 0)) != 0)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (error);
return (0);
Merge scsipi branch in the mainline. This add support for ATAPI devices (currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2 busses to which devices can attach (scsibus and atapibus). This needed to change some include files and structure names in the low level scsi drivers.
1997-08-27 15:22:52 +04:00
}
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* Get the scsi driver to send a full inquiry to the device and use the
* results to fill out the disk parameter structure.
*/
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cd_get_parms(struct cd_softc *cd, int flags)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
{
fixed various bugs like cdattach() returning garbage.
1993-04-12 16:10:28 +04:00
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
/*
* give a number of sectors so that sec * trks * cyls
minphys() functions really should return void.
1995-08-13 00:30:45 +04:00
* is <= disk_size
New SCSI system, based on Julian's more recent work.
1994-03-29 08:29:20 +04:00
*/
Add some missing locking, and some general cleanup.
1994-11-21 01:36:43 +03:00
if (cd_size(cd, flags) == 0)
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (ENXIO);
return (0);
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cdsize(dev_t dev)
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
{
Revert cdsize() to do nothing.
1995-03-23 14:51:22 +03:00
/* CD-ROMs are read-only. */
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (-1);
after 0.2.2 "stable" patches applied
1993-03-21 21:04:42 +03:00
}
Add dummy *dump() routines.
1994-05-11 13:49:17 +04:00
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cddump(dev_t dev, daddr_t blkno, caddr_t va, size_t size)
Add dummy *dump() routines.
1994-05-11 13:49:17 +04:00
{
/* Not implemented. */
Cosmetic changes to keep coding style consistency in this directory; - Indent with tab of width 8. - Use four column to indent continuation line. - Fold long line if possible. - Use return (xx) instead of return xx. - Compare pointer against NULL instead of testing like boolean. - Delete whitespace at the end of line. - Delete whitespace in front of function call operator. - Delete whitespace after cast. - Dereference a pointer to function explicitly. - Add an empty line after local variable declaration. - Use NULL instead of (char *)0. - Dont use block for single statement.
1997-10-01 05:18:38 +04:00
return (ENXIO);
Add dummy *dump() routines.
1994-05-11 13:49:17 +04:00
}
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
#define dvd_copy_key(dst, src) memcpy((dst), (src), sizeof(dvd_key))
#define dvd_copy_challenge(dst, src) memcpy((dst), (src), sizeof(dvd_challenge))
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
dvd_auth(struct cd_softc *cd, dvd_authinfo *a)
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
{
struct scsipi_generic cmd;
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
u_int8_t bf[20];
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
int error;
memset(cmd.bytes, 0, 15);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
memset(bf, 0, sizeof(bf));
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
switch (a->type) {
case DVD_LU_SEND_AGID:
cmd.opcode = GPCMD_REPORT_KEY;
cmd.bytes[8] = 8;
cmd.bytes[9] = 0 | (0 << 6);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
error = scsipi_command(cd->sc_periph, &cmd, 12, bf, 8,
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
CDRETRIES, 30000, NULL,
XS_CTL_DATA_IN|XS_CTL_DATA_ONSTACK);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
if (error)
return (error);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
a->lsa.agid = bf[7] >> 6;
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
return (0);
case DVD_LU_SEND_CHALLENGE:
cmd.opcode = GPCMD_REPORT_KEY;
cmd.bytes[8] = 16;
cmd.bytes[9] = 1 | (a->lsc.agid << 6);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
error = scsipi_command(cd->sc_periph, &cmd, 12, bf, 16,
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
CDRETRIES, 30000, NULL,
XS_CTL_DATA_IN|XS_CTL_DATA_ONSTACK);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
if (error)
return (error);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
dvd_copy_challenge(a->lsc.chal, &bf[4]);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
return (0);
case DVD_LU_SEND_KEY1:
cmd.opcode = GPCMD_REPORT_KEY;
cmd.bytes[8] = 12;
cmd.bytes[9] = 2 | (a->lsk.agid << 6);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
error = scsipi_command(cd->sc_periph, &cmd, 12, bf, 12,
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
CDRETRIES, 30000, NULL,
XS_CTL_DATA_IN|XS_CTL_DATA_ONSTACK);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
if (error)
return (error);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
dvd_copy_key(a->lsk.key, &bf[4]);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
return (0);
case DVD_LU_SEND_TITLE_KEY:
cmd.opcode = GPCMD_REPORT_KEY;
_lto4b(a->lstk.lba, &cmd.bytes[1]);
cmd.bytes[8] = 12;
cmd.bytes[9] = 4 | (a->lstk.agid << 6);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
error = scsipi_command(cd->sc_periph, &cmd, 12, bf, 12,
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
CDRETRIES, 30000, NULL,
XS_CTL_DATA_IN|XS_CTL_DATA_ONSTACK);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
if (error)
return (error);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
a->lstk.cpm = (bf[4] >> 7) & 1;
a->lstk.cp_sec = (bf[4] >> 6) & 1;
a->lstk.cgms = (bf[4] >> 4) & 3;
dvd_copy_key(a->lstk.title_key, &bf[5]);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
return (0);
case DVD_LU_SEND_ASF:
cmd.opcode = GPCMD_REPORT_KEY;
cmd.bytes[8] = 8;
cmd.bytes[9] = 5 | (a->lsasf.agid << 6);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
error = scsipi_command(cd->sc_periph, &cmd, 12, bf, 8,
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
CDRETRIES, 30000, NULL,
XS_CTL_DATA_IN|XS_CTL_DATA_ONSTACK);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
if (error)
return (error);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
a->lsasf.asf = bf[7] & 1;
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
return (0);
case DVD_HOST_SEND_CHALLENGE:
cmd.opcode = GPCMD_SEND_KEY;
cmd.bytes[8] = 16;
cmd.bytes[9] = 1 | (a->hsc.agid << 6);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
bf[1] = 14;
dvd_copy_challenge(&bf[4], a->hsc.chal);
error = scsipi_command(cd->sc_periph, &cmd, 12, bf, 16,
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
CDRETRIES, 30000, NULL,
We can't have XS_CTL_DATA_OUT and XS_CTL_DATA_IN at the same time. Pointed out by Bernd Ernesti.
2001-10-22 20:16:00 +04:00
XS_CTL_DATA_OUT|XS_CTL_DATA_ONSTACK);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
if (error)
return (error);
a->type = DVD_LU_SEND_KEY1;
return (0);
case DVD_HOST_SEND_KEY2:
cmd.opcode = GPCMD_SEND_KEY;
cmd.bytes[8] = 12;
cmd.bytes[9] = 3 | (a->hsk.agid << 6);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
bf[1] = 10;
dvd_copy_key(&bf[4], a->hsk.key);
error = scsipi_command(cd->sc_periph, &cmd, 12, bf, 12,
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
CDRETRIES, 30000, NULL,
We can't have XS_CTL_DATA_OUT and XS_CTL_DATA_IN at the same time. Pointed out by Bernd Ernesti.
2001-10-22 20:16:00 +04:00
XS_CTL_DATA_OUT|XS_CTL_DATA_ONSTACK);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
if (error) {
a->type = DVD_AUTH_FAILURE;
return (error);
}
a->type = DVD_AUTH_ESTABLISHED;
return (0);
case DVD_INVALIDATE_AGID:
cmd.opcode = GPCMD_REPORT_KEY;
cmd.bytes[9] = 0x3f | (a->lsa.agid << 6);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
error = scsipi_command(cd->sc_periph, &cmd, 12, bf, 16,
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
CDRETRIES, 30000, NULL, 0);
if (error)
return (error);
return (0);
Support for dvd region code (RPC).
2001-12-10 01:56:10 +03:00
case DVD_LU_SEND_RPC_STATE:
cmd.opcode = GPCMD_REPORT_KEY;
cmd.bytes[8] = 8;
cmd.bytes[9] = 8 | (0 << 6);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
error = scsipi_command(cd->sc_periph, &cmd, 12, bf, 8,
Support for dvd region code (RPC).
2001-12-10 01:56:10 +03:00
CDRETRIES, 30000, NULL,
XS_CTL_DATA_IN|XS_CTL_DATA_ONSTACK);
if (error)
return (error);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
a->lrpcs.type = (bf[4] >> 6) & 3;
a->lrpcs.vra = (bf[4] >> 3) & 7;
a->lrpcs.ucca = (bf[4]) & 7;
a->lrpcs.region_mask = bf[5];
a->lrpcs.rpc_scheme = bf[6];
Support for dvd region code (RPC).
2001-12-10 01:56:10 +03:00
return (0);
case DVD_HOST_SEND_RPC_STATE:
cmd.opcode = GPCMD_SEND_KEY;
cmd.bytes[8] = 8;
cmd.bytes[9] = 6 | (0 << 6);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
bf[1] = 6;
bf[4] = a->hrpcs.pdrc;
error = scsipi_command(cd->sc_periph, &cmd, 12, bf, 8,
Support for dvd region code (RPC).
2001-12-10 01:56:10 +03:00
CDRETRIES, 30000, NULL,
XS_CTL_DATA_OUT|XS_CTL_DATA_ONSTACK);
if (error)
return (error);
return (0);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
default:
return (ENOTTY);
}
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
dvd_read_physical(struct cd_softc *cd, dvd_struct *s)
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
{
struct scsipi_generic cmd;
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
u_int8_t bf[4 + 4 * 20], *bufp;
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
int error;
struct dvd_layer *layer;
int i;
memset(cmd.bytes, 0, 15);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
memset(bf, 0, sizeof(bf));
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
cmd.opcode = GPCMD_READ_DVD_STRUCTURE;
cmd.bytes[6] = s->type;
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
_lto2b(sizeof(bf), &cmd.bytes[7]);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
cmd.bytes[5] = s->physical.layer_num;
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
error = scsipi_command(cd->sc_periph, &cmd, 12, bf, sizeof(bf),
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
CDRETRIES, 30000, NULL, XS_CTL_DATA_IN|XS_CTL_DATA_ONSTACK);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
if (error)
return (error);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
for (i = 0, bufp = &bf[4], layer = &s->physical.layer[0]; i < 4;
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
i++, bufp += 20, layer++) {
memset(layer, 0, sizeof(*layer));
layer->book_version = bufp[0] & 0xf;
layer->book_type = bufp[0] >> 4;
layer->min_rate = bufp[1] & 0xf;
layer->disc_size = bufp[1] >> 4;
layer->layer_type = bufp[2] & 0xf;
layer->track_path = (bufp[2] >> 4) & 1;
layer->nlayers = (bufp[2] >> 5) & 3;
layer->track_density = bufp[3] & 0xf;
layer->linear_density = bufp[3] >> 4;
Fix the size of start_sector, end_sector and end_sector_l0. (I'm just going to assume the Linux people will also fix this, since I reported it to them.)
1999-10-31 17:03:16 +03:00
layer->start_sector = _4btol(&bufp[4]);
layer->end_sector = _4btol(&bufp[8]);
layer->end_sector_l0 = _4btol(&bufp[12]);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
layer->bca = bufp[16] >> 7;
}
return (0);
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
dvd_read_copyright(struct cd_softc *cd, dvd_struct *s)
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
{
struct scsipi_generic cmd;
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
u_int8_t bf[8];
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
int error;
memset(cmd.bytes, 0, 15);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
memset(bf, 0, sizeof(bf));
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
cmd.opcode = GPCMD_READ_DVD_STRUCTURE;
cmd.bytes[6] = s->type;
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
_lto2b(sizeof(bf), &cmd.bytes[7]);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
cmd.bytes[5] = s->copyright.layer_num;
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
error = scsipi_command(cd->sc_periph, &cmd, 12, bf, sizeof(bf),
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
CDRETRIES, 30000, NULL, XS_CTL_DATA_IN|XS_CTL_DATA_ONSTACK);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
if (error)
return (error);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
s->copyright.cpst = bf[4];
s->copyright.rmi = bf[5];
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
return (0);
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
dvd_read_disckey(struct cd_softc *cd, dvd_struct *s)
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
{
struct scsipi_generic cmd;
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
u_int8_t *bf;
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
int error;
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
bf = malloc(4 + 2048, M_TEMP, M_WAITOK|M_ZERO);
if (bf == NULL)
Don't waste too much kernel stack space on (infrequent) ioctl operations, use malloc instead for temp space.
2002-12-15 04:55:44 +03:00
return EIO;
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
memset(cmd.bytes, 0, 15);
cmd.opcode = GPCMD_READ_DVD_STRUCTURE;
cmd.bytes[6] = s->type;
In dvd_read_disckey, sizeof(buf) is no longer '4 + 2048' after the malloc change. Replace 'sizeof(buf)' with '4 + 2048' -- this makes dvd_read_disckey work properly again.
2002-12-15 20:43:50 +03:00
_lto2b(4 + 2048, &cmd.bytes[7]);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
cmd.bytes[9] = s->disckey.agid << 6;
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
error = scsipi_command(cd->sc_periph, &cmd, 12, bf, 4 + 2048,
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
CDRETRIES, 30000, NULL, XS_CTL_DATA_IN|XS_CTL_DATA_ONSTACK);
Don't waste too much kernel stack space on (infrequent) ioctl operations, use malloc instead for temp space.
2002-12-15 04:55:44 +03:00
if (error == 0)
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
memcpy(s->disckey.value, &bf[4], 2048);
free(bf, M_TEMP);
Don't waste too much kernel stack space on (infrequent) ioctl operations, use malloc instead for temp space.
2002-12-15 04:55:44 +03:00
return error;
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
dvd_read_bca(struct cd_softc *cd, dvd_struct *s)
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
{
struct scsipi_generic cmd;
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
u_int8_t bf[4 + 188];
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
int error;
memset(cmd.bytes, 0, 15);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
memset(bf, 0, sizeof(bf));
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
cmd.opcode = GPCMD_READ_DVD_STRUCTURE;
cmd.bytes[6] = s->type;
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
_lto2b(sizeof(bf), &cmd.bytes[7]);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
error = scsipi_command(cd->sc_periph, &cmd, 12, bf, sizeof(bf),
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
CDRETRIES, 30000, NULL, XS_CTL_DATA_IN|XS_CTL_DATA_ONSTACK);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
if (error)
return (error);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
s->bca.len = _2btol(&bf[0]);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
if (s->bca.len < 12 || s->bca.len > 188)
return (EIO);
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
memcpy(s->bca.value, &bf[4], s->bca.len);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
return (0);
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
dvd_read_manufact(struct cd_softc *cd, dvd_struct *s)
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
{
struct scsipi_generic cmd;
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
u_int8_t *bf;
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
int error;
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
bf = malloc(4 + 2048, M_TEMP, M_WAITOK|M_ZERO);
if (bf == NULL)
Don't waste too much kernel stack space on (infrequent) ioctl operations, use malloc instead for temp space.
2002-12-15 04:55:44 +03:00
return (EIO);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
memset(cmd.bytes, 0, 15);
cmd.opcode = GPCMD_READ_DVD_STRUCTURE;
cmd.bytes[6] = s->type;
Another sizeof -> 4 + 2048 fix for dvd_read_manufact after malloc changes
2002-12-15 20:48:14 +03:00
_lto2b(4 + 2048, &cmd.bytes[7]);
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
error = scsipi_command(cd->sc_periph, &cmd, 12, bf, 4 + 2048,
Prevent a process being swapped out during I/O if the data buffer is allocated on stack. This potential problem is noticed by Noriyuki Soda and the idea and sample code to fix is given by Jason R. Thorpe.
2000-06-09 12:54:19 +04:00
CDRETRIES, 30000, NULL, XS_CTL_DATA_IN|XS_CTL_DATA_ONSTACK);
Don't waste too much kernel stack space on (infrequent) ioctl operations, use malloc instead for temp space.
2002-12-15 04:55:44 +03:00
if (error == 0) {
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
s->manufact.len = _2btol(&bf[0]);
Don't waste too much kernel stack space on (infrequent) ioctl operations, use malloc instead for temp space.
2002-12-15 04:55:44 +03:00
if (s->manufact.len >= 0 && s->manufact.len <= 2048)
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
memcpy(s->manufact.value, &bf[4], s->manufact.len);
Don't waste too much kernel stack space on (infrequent) ioctl operations, use malloc instead for temp space.
2002-12-15 04:55:44 +03:00
else
error = EIO;
}
- Sprinkle const - Avoid variable shadowing. - Eliminate some caddr_t abuse.
2005-05-30 02:00:50 +04:00
free(bf, M_TEMP);
Don't waste too much kernel stack space on (infrequent) ioctl operations, use malloc instead for temp space.
2002-12-15 04:55:44 +03:00
return error;
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
dvd_read_struct(struct cd_softc *cd, dvd_struct *s)
Emulate the Linux DVD_* ioctls(2). This gets us 90% of the way to running the LiViD DVD player. (See forthcoming mail to current-users.) XXX NOTE: We should do something to probe capabilities, rather than allowing these ioctls on any device.
1999-10-29 19:02:56 +04:00
{
switch (s->type) {
case DVD_STRUCT_PHYSICAL:
return (dvd_read_physical(cd, s));
case DVD_STRUCT_COPYRIGHT:
return (dvd_read_copyright(cd, s));
case DVD_STRUCT_DISCKEY:
return (dvd_read_disckey(cd, s));
case DVD_STRUCT_BCA:
return (dvd_read_bca(cd, s));
case DVD_STRUCT_MANUFACT:
return (dvd_read_manufact(cd, s));
default:
return (EINVAL);
}
}
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
static int
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
cd_mode_sense(struct cd_softc *cd, u_int8_t byte2, void *sense, size_t size,
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
int page, int flags, int *big)
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
{
if (cd->sc_periph->periph_quirks & PQUIRK_ONLYBIG) {
*big = 1;
return scsipi_mode_sense_big(cd->sc_periph, byte2, page, sense,
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
size + sizeof(struct scsi_mode_parameter_header_10),
Attempt to deal with Martin's weirdass Sun drive by setting the mode sense allocation length a little more precisely -- add the space for the header in cd_mode_sense(). Also delete the XS_CTL_SILENT, since we really do want to see errors. Lastly, add a similar wrapper for mode select, simplifying the callers slightly.
2003-09-08 20:16:43 +04:00
flags | XS_CTL_DATA_ONSTACK, CDRETRIES, 20000);
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
} else {
*big = 0;
return scsipi_mode_sense(cd->sc_periph, byte2, page, sense,
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
size + sizeof(struct scsi_mode_parameter_header_6),
Attempt to deal with Martin's weirdass Sun drive by setting the mode sense allocation length a little more precisely -- add the space for the header in cd_mode_sense(). Also delete the XS_CTL_SILENT, since we really do want to see errors. Lastly, add a similar wrapper for mode select, simplifying the callers slightly.
2003-09-08 20:16:43 +04:00
flags | XS_CTL_DATA_ONSTACK, CDRETRIES, 20000);
}
}
static int
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
cd_mode_select(struct cd_softc *cd, u_int8_t byte2, void *sense, size_t size,
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
int flags, int big)
Attempt to deal with Martin's weirdass Sun drive by setting the mode sense allocation length a little more precisely -- add the space for the header in cd_mode_sense(). Also delete the XS_CTL_SILENT, since we really do want to see errors. Lastly, add a similar wrapper for mode select, simplifying the callers slightly.
2003-09-08 20:16:43 +04:00
{
if (big) {
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
struct scsi_mode_parameter_header_10 *header = sense;
Attempt to deal with Martin's weirdass Sun drive by setting the mode sense allocation length a little more precisely -- add the space for the header in cd_mode_sense(). Also delete the XS_CTL_SILENT, since we really do want to see errors. Lastly, add a similar wrapper for mode select, simplifying the callers slightly.
2003-09-08 20:16:43 +04:00
_lto2b(0, header->data_length);
return scsipi_mode_select_big(cd->sc_periph, byte2, sense,
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
size + sizeof(struct scsi_mode_parameter_header_10),
Attempt to deal with Martin's weirdass Sun drive by setting the mode sense allocation length a little more precisely -- add the space for the header in cd_mode_sense(). Also delete the XS_CTL_SILENT, since we really do want to see errors. Lastly, add a similar wrapper for mode select, simplifying the callers slightly.
2003-09-08 20:16:43 +04:00
flags | XS_CTL_DATA_ONSTACK, CDRETRIES, 20000);
} else {
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
struct scsi_mode_parameter_header_6 *header = sense;
Attempt to deal with Martin's weirdass Sun drive by setting the mode sense allocation length a little more precisely -- add the space for the header in cd_mode_sense(). Also delete the XS_CTL_SILENT, since we really do want to see errors. Lastly, add a similar wrapper for mode select, simplifying the callers slightly.
2003-09-08 20:16:43 +04:00
header->data_length = 0;
return scsipi_mode_select(cd->sc_periph, byte2, sense,
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
size + sizeof(struct scsi_mode_parameter_header_6),
Attempt to deal with Martin's weirdass Sun drive by setting the mode sense allocation length a little more precisely -- add the space for the header in cd_mode_sense(). Also delete the XS_CTL_SILENT, since we really do want to see errors. Lastly, add a similar wrapper for mode select, simplifying the callers slightly.
2003-09-08 20:16:43 +04:00
flags | XS_CTL_DATA_ONSTACK, CDRETRIES, 20000);
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
}
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cd_set_pa_immed(struct cd_softc *cd, int flags)
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
{
struct {
union {
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
struct scsi_mode_parameter_header_6 small;
struct scsi_mode_parameter_header_10 big;
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
} header;
struct cd_audio_page page;
} data;
int error;
uint8_t oflags;
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
int big, byte2;
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
struct cd_audio_page *page;
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
byte2 = SMS_DBD;
try_again:
if ((error = cd_mode_sense(cd, byte2, &data, sizeof(data.page),
AUDIO_PAGE, flags, &big)) != 0) {
if (byte2 == SMS_DBD) {
/* Device may not understand DBD; retry without */
byte2 = 0;
goto try_again;
}
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (error);
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
}
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
if (big)
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
page = (void *)((u_long)&data.header.big +
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
sizeof data.header.big +
_2btol(data.header.big.blk_desc_len));
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
else
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
page = (void *)((u_long)&data.header.small +
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
sizeof data.header.small +
data.header.small.blk_desc_len);
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
oflags = page->flags;
page->flags &= ~CD_PA_SOTC;
page->flags |= CD_PA_IMMED;
Nuke some printf()s.
2003-09-08 05:13:04 +04:00
if (oflags == page->flags)
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (0);
Attempt to deal with Martin's weirdass Sun drive by setting the mode sense allocation length a little more precisely -- add the space for the header in cd_mode_sense(). Also delete the XS_CTL_SILENT, since we really do want to see errors. Lastly, add a similar wrapper for mode select, simplifying the callers slightly.
2003-09-08 20:16:43 +04:00
return (cd_mode_select(cd, SMS_PF, &data,
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
sizeof(struct scsi_mode_page_header) + page->pg_length,
Attempt to deal with Martin's weirdass Sun drive by setting the mode sense allocation length a little more precisely -- add the space for the header in cd_mode_sense(). Also delete the XS_CTL_SILENT, since we really do want to see errors. Lastly, add a similar wrapper for mode select, simplifying the callers slightly.
2003-09-08 20:16:43 +04:00
flags, big));
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cd_setchan(struct cd_softc *cd, int p0, int p1, int p2, int p3, int flags)
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
{
struct {
union {
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
struct scsi_mode_parameter_header_6 small;
struct scsi_mode_parameter_header_10 big;
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
} header;
struct cd_audio_page page;
} data;
int error;
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
int big, byte2;
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
struct cd_audio_page *page;
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
byte2 = SMS_DBD;
try_again:
if ((error = cd_mode_sense(cd, byte2, &data, sizeof(data.page),
AUDIO_PAGE, flags, &big)) != 0) {
if (byte2 == SMS_DBD) {
/* Device may not understand DBD; retry without */
byte2 = 0;
goto try_again;
}
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (error);
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
}
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
if (big)
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
page = (void *)((u_long)&data.header.big +
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
sizeof data.header.big +
_2btol(data.header.big.blk_desc_len));
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
else
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
page = (void *)((u_long)&data.header.small +
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
sizeof data.header.small +
data.header.small.blk_desc_len);
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
page->port[0].channels = p0;
page->port[1].channels = p1;
page->port[2].channels = p2;
page->port[3].channels = p3;
Attempt to deal with Martin's weirdass Sun drive by setting the mode sense allocation length a little more precisely -- add the space for the header in cd_mode_sense(). Also delete the XS_CTL_SILENT, since we really do want to see errors. Lastly, add a similar wrapper for mode select, simplifying the callers slightly.
2003-09-08 20:16:43 +04:00
return (cd_mode_select(cd, SMS_PF, &data,
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
sizeof(struct scsi_mode_page_header) + page->pg_length,
Attempt to deal with Martin's weirdass Sun drive by setting the mode sense allocation length a little more precisely -- add the space for the header in cd_mode_sense(). Also delete the XS_CTL_SILENT, since we really do want to see errors. Lastly, add a similar wrapper for mode select, simplifying the callers slightly.
2003-09-08 20:16:43 +04:00
flags, big));
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cd_getvol(struct cd_softc *cd, struct ioc_vol *arg, int flags)
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
{
struct {
union {
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
struct scsi_mode_parameter_header_6 small;
struct scsi_mode_parameter_header_10 big;
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
} header;
struct cd_audio_page page;
} data;
int error;
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
int big, byte2;
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
struct cd_audio_page *page;
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
byte2 = SMS_DBD;
try_again:
if ((error = cd_mode_sense(cd, byte2, &data, sizeof(data.page),
AUDIO_PAGE, flags, &big)) != 0) {
if (byte2 == SMS_DBD) {
/* Device may not understand DBD; retry without */
byte2 = 0;
goto try_again;
}
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (error);
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
}
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
if (big)
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
page = (void *)((u_long)&data.header.big +
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
sizeof data.header.big +
_2btol(data.header.big.blk_desc_len));
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
else
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
page = (void *)((u_long)&data.header.small +
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
sizeof data.header.small +
data.header.small.blk_desc_len);
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
arg->vol[0] = page->port[0].volume;
arg->vol[1] = page->port[1].volume;
arg->vol[2] = page->port[2].volume;
arg->vol[3] = page->port[3].volume;
return (0);
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cd_setvol(struct cd_softc *cd, const struct ioc_vol *arg, int flags)
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
{
struct {
union {
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
struct scsi_mode_parameter_header_6 small;
struct scsi_mode_parameter_header_10 big;
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
} header;
struct cd_audio_page page;
} data, mask;
int error;
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
int big, byte2;
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
struct cd_audio_page *page, *page2;
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
byte2 = SMS_DBD;
try_again:
if ((error = cd_mode_sense(cd, byte2, &data, sizeof(data.page),
AUDIO_PAGE, flags, &big)) != 0) {
if (byte2 == SMS_DBD) {
/* Device may not understand DBD; retry without */
byte2 = 0;
goto try_again;
}
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (error);
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
}
if ((error = cd_mode_sense(cd, byte2, &mask, sizeof(mask.page),
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
AUDIO_PAGE|SMS_PCTRL_CHANGEABLE, flags, &big)) != 0)
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (error);
if (big) {
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
page = (void *)((u_long)&data.header.big +
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
sizeof data.header.big +
_2btol(data.header.big.blk_desc_len));
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
page2 = (void *)((u_long)&mask.header.big +
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
sizeof mask.header.big +
_2btol(mask.header.big.blk_desc_len));
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
} else {
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
page = (void *)((u_long)&data.header.small +
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
sizeof data.header.small +
data.header.small.blk_desc_len);
Backing out changes to clean up scsipi. I was pointed out there were problems i hadn't seen. To prevent lossage i'd decided to back off all changes and let them be reviewed on tech-kern.
2005-02-01 03:19:34 +03:00
page2 = (void *)((u_long)&mask.header.small +
Some older devices do not understand the `disable block descriptor' bit in the mode sense request. So fall back on mode sense data including a block descriptor section. See also sd.c rev. 1.214. Again: should we bother trying DBD at all?
2004-04-24 13:26:14 +04:00
sizeof mask.header.small +
mask.header.small.blk_desc_len);
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
}
page->port[0].volume = arg->vol[0] & page2->port[0].volume;
page->port[1].volume = arg->vol[1] & page2->port[1].volume;
page->port[2].volume = arg->vol[2] & page2->port[2].volume;
page->port[3].volume = arg->vol[3] & page2->port[3].volume;
page->port[0].channels = CHANNEL_0;
page->port[1].channels = CHANNEL_1;
Attempt to deal with Martin's weirdass Sun drive by setting the mode sense allocation length a little more precisely -- add the space for the header in cd_mode_sense(). Also delete the XS_CTL_SILENT, since we really do want to see errors. Lastly, add a similar wrapper for mode select, simplifying the callers slightly.
2003-09-08 20:16:43 +04:00
return (cd_mode_select(cd, SMS_PF, &data,
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
sizeof(struct scsi_mode_page_header) + page->pg_length,
Attempt to deal with Martin's weirdass Sun drive by setting the mode sense allocation length a little more precisely -- add the space for the header in cd_mode_sense(). Also delete the XS_CTL_SILENT, since we really do want to see errors. Lastly, add a similar wrapper for mode select, simplifying the callers slightly.
2003-09-08 20:16:43 +04:00
flags, big));
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cd_load_unload(struct cd_softc *cd, struct ioc_load_unload *args)
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
{
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
struct scsipi_load_unload cmd;
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
memset(&cmd, 0, sizeof(cmd));
cmd.opcode = LOAD_UNLOAD;
cmd.options = args->options; /* ioctl uses MMC values */
cmd.slot = args->slot;
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
Standardize some variable names and the calling pattern for scsipi_command(). Use void pointer casts.
2004-09-18 04:08:16 +04:00
return (scsipi_command(cd->sc_periph, (void *)&cmd, sizeof(cmd), 0, 0,
CDRETRIES, 200000, NULL, 0));
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
}
Use ANSI function decls and make use of static.
2004-08-21 21:40:25 +04:00
static int
cd_setblksize(struct cd_softc *cd)
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
{
struct {
union {
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
struct scsi_mode_parameter_header_6 small;
struct scsi_mode_parameter_header_10 big;
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
} header;
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
struct scsi_general_block_descriptor blk_desc;
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
} data;
int error;
int big, bsize;
Part 1 of a cleanup pass over the SCSI subsystem. The aim is to name everything "scsi_*", since we really are talking about the SCSI command set, ATAPI transport not withstanding. Improve the names of many structures, and prepend "SCSI_" onto all SCSI command opcodes. Place items described by the SCSI Primary Commands document into scsi_spc.h.
2005-02-21 03:29:06 +03:00
struct scsi_general_block_descriptor *bdesc;
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
Attempt to deal with Martin's weirdass Sun drive by setting the mode sense allocation length a little more precisely -- add the space for the header in cd_mode_sense(). Also delete the XS_CTL_SILENT, since we really do want to see errors. Lastly, add a similar wrapper for mode select, simplifying the callers slightly.
2003-09-08 20:16:43 +04:00
if ((error = cd_mode_sense(cd, 0, &data, sizeof(data.blk_desc), 0, 0,
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
&big)) != 0)
return (error);
if (big) {
bdesc = (void *)(&data.header.big + 1);
bsize = _2btol(data.header.big.blk_desc_len);
} else {
bdesc = (void *)(&data.header.small + 1);
bsize = data.header.small.blk_desc_len;
}
if (bsize == 0) {
Nuke some printf()s.
2003-09-08 05:13:04 +04:00
printf("cd_setblksize: trying to change bsize, but no blk_desc\n");
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (EINVAL);
}
if (_3btol(bdesc->blklen) == 2048) {
Nuke some printf()s.
2003-09-08 05:13:04 +04:00
printf("cd_setblksize: trying to change bsize, but blk_desc is correct\n");
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
return (EINVAL);
}
nuke trailing whitespace
2005-02-27 03:26:58 +03:00
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
_lto3b(2048, bdesc->blklen);
Oops. Fix a typo in cd_mode_select().
2003-09-09 03:44:29 +04:00
return (cd_mode_select(cd, SMS_PF, &data, sizeof(data.blk_desc), 0,
big));
Set PQUIRK_ONLYBIG in the wdc-atapi frontend, obviating the need to ever test the "bus type" for this. Merge all the code in the SCSI and ATAPI backends for "cd" devices. All of the mode page handling and whatnot is general to SCSI MMC devices, and should never have been separated to begin with. This fixes a variety of problems, and adds load/unload support for SCSI-attached devices.
2003-09-08 02:11:22 +04:00
}
Cleanup and fix cd_size() function; it was returning wrong results. Also factor out read_cd_capacity()
2005-09-07 02:19:14 +04:00