Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
72315d3199
NetBSD/sys/compat/darwin/darwin_iohidsystem.c

724 lines
23 KiB
C
Raw Normal View History

Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
/* $NetBSD: darwin_iohidsystem.c,v 1.17 2003/10/25 10:43:45 manu Exp $ */
Match IOHIDSystem and IOFramebuffer request in the IOKit and handle them in Darwin specific files. The handler does nothing yet. This needs some improvement. Darwin attaches an IOHIDSystem device to each human-interface device in the tree: mouse, keyboard, and a few others. For now we only emulate one IOHIDSystem device.
2003-02-16 18:02:05 +03:00
/*-
* Copyright (c) 2003 The NetBSD Foundation, Inc.
* All rights reserved.
*
* This code is derived from software contributed to The NetBSD Foundation
* by Emmanuel Dreyfus.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the 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.
*
* 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.
*/
#include <sys/cdefs.h>
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
__KERNEL_RCSID(0, "$NetBSD: darwin_iohidsystem.c,v 1.17 2003/10/25 10:43:45 manu Exp $");
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
#include "ioconf.h"
#include "wsmux.h"
Match IOHIDSystem and IOFramebuffer request in the IOKit and handle them in Darwin specific files. The handler does nothing yet. This needs some improvement. Darwin attaches an IOHIDSystem device to each human-interface device in the tree: mouse, keyboard, and a few others. For now we only emulate one IOHIDSystem device.
2003-02-16 18:02:05 +03:00
#include <sys/types.h>
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/malloc.h>
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
#include <sys/fcntl.h>
#include <sys/uio.h>
#include <sys/select.h>
Match IOHIDSystem and IOFramebuffer request in the IOKit and handle them in Darwin specific files. The handler does nothing yet. This needs some improvement. Darwin attaches an IOHIDSystem device to each human-interface device in the tree: mouse, keyboard, and a few others. For now we only emulate one IOHIDSystem device.
2003-02-16 18:02:05 +03:00
#include <sys/signal.h>
#include <sys/mount.h>
#include <sys/proc.h>
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
#include <sys/conf.h>
Match IOHIDSystem and IOFramebuffer request in the IOKit and handle them in Darwin specific files. The handler does nothing yet. This needs some improvement. Darwin attaches an IOHIDSystem device to each human-interface device in the tree: mouse, keyboard, and a few others. For now we only emulate one IOHIDSystem device.
2003-02-16 18:02:05 +03:00
#include <sys/device.h>
Added the map_memory method for the IOKit, plus an implementation for the IOHIDSystem driver in Darwin. IOHIDSystem provides access to user interface devices. Communication between userspace and kernel is done through a shared memory page. Our plan is to have a kernel thread waiting on wscons events (for now it just does nothing), and have it feeding the shared page.
2003-04-30 02:16:38 +04:00
#include <sys/kthread.h>
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
#include <dev/wscons/wsconsio.h>
#include <dev/wscons/wseventvar.h>
#include <dev/wscons/wsmuxvar.h>
Added the map_memory method for the IOKit, plus an implementation for the IOHIDSystem driver in Darwin. IOHIDSystem provides access to user interface devices. Communication between userspace and kernel is done through a shared memory page. Our plan is to have a kernel thread waiting on wscons events (for now it just does nothing), and have it feeding the shared page.
2003-04-30 02:16:38 +04:00
#include <uvm/uvm_extern.h>
#include <uvm/uvm_map.h>
#include <uvm/uvm.h>
Match IOHIDSystem and IOFramebuffer request in the IOKit and handle them in Darwin specific files. The handler does nothing yet. This needs some improvement. Darwin attaches an IOHIDSystem device to each human-interface device in the tree: mouse, keyboard, and a few others. For now we only emulate one IOHIDSystem device.
2003-02-16 18:02:05 +03:00
#include <compat/mach/mach_types.h>
#include <compat/mach/mach_message.h>
#include <compat/mach/mach_port.h>
#include <compat/mach/mach_errno.h>
#include <compat/mach/mach_iokit.h>
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
#include <compat/darwin/darwin_exec.h>
Match IOHIDSystem and IOFramebuffer request in the IOKit and handle them in Darwin specific files. The handler does nothing yet. This needs some improvement. Darwin attaches an IOHIDSystem device to each human-interface device in the tree: mouse, keyboard, and a few others. For now we only emulate one IOHIDSystem device.
2003-02-16 18:02:05 +03:00
#include <compat/darwin/darwin_iokit.h>
Handle mouseup and mousdown events Add three sysctl to choose wsdisplay unit and screen and wsmux that ioframebuffer and iohidsystem will use.
2003-09-14 13:48:42 +04:00
#include <compat/darwin/darwin_sysctl.h>
Added io_service_close and the io_connect_map_memory method for ioframebuffer's shared meory (used to handle the cursor). XDarwin and WindowServer now fail on framebuffer mapping.
2003-05-14 22:28:04 +04:00
#include <compat/darwin/darwin_iohidsystem.h>
Match IOHIDSystem and IOFramebuffer request in the IOKit and handle them in Darwin specific files. The handler does nothing yet. This needs some improvement. Darwin attaches an IOHIDSystem device to each human-interface device in the tree: mouse, keyboard, and a few others. For now we only emulate one IOHIDSystem device.
2003-02-16 18:02:05 +03:00
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
/* Redefined from sys/dev/wscons/wsmux.c */
extern const struct cdevsw wsmux_cdevsw;
Added the map_memory method for the IOKit, plus an implementation for the IOHIDSystem driver in Darwin. IOHIDSystem provides access to user interface devices. Communication between userspace and kernel is done through a shared memory page. Our plan is to have a kernel thread waiting on wscons events (for now it just does nothing), and have it feeding the shared page.
2003-04-30 02:16:38 +04:00
static struct uvm_object *darwin_iohidsystem_shmem = NULL;
static void darwin_iohidsystem_shmeminit(vaddr_t);
static void darwin_iohidsystem_thread(void *);
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
static int darwin_findwsmux(dev_t *, int);
static void darwin_wscons_to_iohidsystem
(struct wscons_event *, darwin_iohidsystem_event *);
static void mach_notify_iohidsystem(struct lwp *, struct mach_right *);
struct darwin_iohidsystem_thread_args {
vaddr_t dita_shmem;
Sync between parent and child in iohdisystem thread so that the child has time to get its parmeteres before the parent throw them away
2003-09-13 11:56:54 +04:00
struct proc *dita_p;
int dita_done;
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
int *dita_hidsystem_finished;
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
};
Added the map_memory method for the IOKit, plus an implementation for the IOHIDSystem driver in Darwin. IOHIDSystem provides access to user interface devices. Communication between userspace and kernel is done through a shared memory page. Our plan is to have a kernel thread waiting on wscons events (for now it just does nothing), and have it feeding the shared page.
2003-04-30 02:16:38 +04:00
Added IOHIDSystem properties (this is about keyboard mapping) Also added a ssize arg to mach trap args so that trap handler are aware of the sent packet size.
2003-06-03 10:48:47 +04:00
#if 0
static char darwin_iohidsystem_properties[] = "<dict ID=\"0\"><key>IOKit</key><string ID=\"1\">IOService</string><key>AccessMPC106PerformanceRegister</key><string ID=\"2\">AppleGracklePCI is not serializable</string><key>IONVRAM</key><reference IDREF=\"1\"/><key>IOiic0</key><string ID=\"3\">ApplePMU is not serializable</string><key>IORTC</key><reference IDREF=\"3\"/><key>IOBSD</key><reference IDREF=\"1\"/><key>setModemSound</key><string ID=\"4\">AppleScreamerAudio is not serializable</string><key>kdp</key><reference IDREF=\"1\"/></dict>";
#endif
static char darwin_iohidsystem_properties[] = "<dict ID=\"0\"><key>IOClass</key><string ID=\"1\">AppleADBKeyboard</string><key>ADBVirtualKeys</key><string ID=\"2\">0x00,0x01,0x02,0x03,0x04,0x05,0x06,0x07,0x08,0x09,0x0A,0x0B,0x0C,0x0D,0x0E,0x0F,0x10,0x11,0x12,0x13,0x14,0x15,0x16,0x17,0x18,0x19,0x1A,0x1B,0x1C,0x1D,0x1E,0x1F,0x20,0x21,0x22,0x23,0x24,0x25,0x26,0x27,0x28,0x29,0x2A,0x2B,0x2C,0x2D,0x2E,0x2F,0x30,0x31,0x32,0x33,0x34,0x35,0x3B,0x37,0x38,0x39,0x3A,0x7B,0x7C,0x7D,0x7E,0x3F,0x40,0x41,0x42,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0x4A,0x4B,0x4C,0x4D,0x4E,0x4F,0x50,0x51,0x52,0x53,0x54,0x55,0x56,0x57,0x58,0x59,0x5A,0x5B,0x5C,0x5D,0x5E,0x5F,0x60,0x61,0x62,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0x6A,0x6B,0x6C,0x6D,0x6E,0x6F,0x70,0x71,0x72,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7A,0x3C,0x3D,0x3E,0x36,0x7F</string><key>IOProbeScore</key><integer size=\"32\" ID=\"3\">0x3e8</integer><key>IOProviderClass</key><string ID=\"4\">IOADBDevice</string><key>CFBundleIdentifier</key><string ID=\"5\">com.apple.driver.AppleADBKeyboard</string><key>ADB Match</key><string ID=\"6\">2</string><key>PowerBook fn Foward Delete</key><integer size=\"64\" ID=\"7\">0x1</integer><key>IOMatchCategory</key><string ID=\"8\">IODefaultMatchCategory</string><key>HIDKeyMapping</key><data ID=\"10\">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</data><key>HIDKind</key><integer size=\"32\" ID=\"11\">0x1</integer><key>HIDInterfaceID</key><integer size=\"32\" ID=\"12\">0x2</integer><key>HIDSubinterfaceID</key><integer size=\"32\" ID=\"13\">0xc4</integer></dict>";
Introduced IOKit driver methods. Moved back into Mach some IOKit driver support that was in Darwin code. I get an empty darwin_iokit.c, which is getting a good candidate for cvs delete.
2003-02-21 01:39:42 +03:00
struct mach_iokit_devclass darwin_iohidsystem_devclass = {
"<dict ID=\"0\"><key>IOProviderClass</key>"
"<string ID=\"1\">IOHIDSystem</string></dict>",
Added IOHIDSystem properties (this is about keyboard mapping) Also added a ssize arg to mach trap args so that trap handler are aware of the sent packet size.
2003-06-03 10:48:47 +04:00
darwin_iohidsystem_properties,
Quick and dirty version of io_registry_entry_get_properties.
2003-03-09 21:33:28 +03:00
NULL,
Introduced IOKit driver methods. Moved back into Mach some IOKit driver support that was in Darwin code. I get an empty darwin_iokit.c, which is getting a good candidate for cvs delete.
2003-02-21 01:39:42 +03:00
darwin_iohidsystem_connect_method_scalari_scalaro,
More efforts on the XDarwin front. Introduce io_connect_method_scalari_structo and io_connect_method_structi_structo. We don't know at all how this stuff works, but at least we can imitate what Darwin does.
2003-05-14 18:41:04 +04:00
NULL,
Remove references to wsmux_cdevsw and wsdisplay_cdevsw (how to remove them completely?) Add SETCURSORPOSITION operation
2003-10-18 17:27:17 +04:00
darwin_iohidsystem_connect_method_structi_structo,
+ In ioframebuffer driver: We try to map the console framebuffer through wsdisplay. It seems to work, but we still need to _see_ something. XXX Framebuffer size discovery is borken. + In ioframebuffer and iohidsystem: More methods and more debug for io_connect_method_* + A few new IOKit mach traps.
2003-05-23 02:07:36 +04:00
NULL,
Added the map_memory method for the IOKit, plus an implementation for the IOHIDSystem driver in Darwin. IOHIDSystem provides access to user interface devices. Communication between userspace and kernel is done through a shared memory page. Our plan is to have a kernel thread waiting on wscons events (for now it just does nothing), and have it feeding the shared page.
2003-04-30 02:16:38 +04:00
darwin_iohidsystem_connect_map_memory,
Introduced IOKit driver methods. Moved back into Mach some IOKit driver support that was in Darwin code. I get an empty darwin_iokit.c, which is getting a good candidate for cvs delete.
2003-02-21 01:39:42 +03:00
"IOHIDSystem",
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
NULL,
Introduced IOKit driver methods. Moved back into Mach some IOKit driver support that was in Darwin code. I get an empty darwin_iokit.c, which is getting a good candidate for cvs delete.
2003-02-21 01:39:42 +03:00
};
Match IOHIDSystem and IOFramebuffer request in the IOKit and handle them in Darwin specific files. The handler does nothing yet. This needs some improvement. Darwin attaches an IOHIDSystem device to each human-interface device in the tree: mouse, keyboard, and a few others. For now we only emulate one IOHIDSystem device.
2003-02-16 18:02:05 +03:00
int
Introduced IOKit driver methods. Moved back into Mach some IOKit driver support that was in Darwin code. I get an empty darwin_iokit.c, which is getting a good candidate for cvs delete.
2003-02-21 01:39:42 +03:00
darwin_iohidsystem_connect_method_scalari_scalaro(args)
Match IOHIDSystem and IOFramebuffer request in the IOKit and handle them in Darwin specific files. The handler does nothing yet. This needs some improvement. Darwin attaches an IOHIDSystem device to each human-interface device in the tree: mouse, keyboard, and a few others. For now we only emulate one IOHIDSystem device.
2003-02-16 18:02:05 +03:00
struct mach_trap_args *args;
{
mach_io_connect_method_scalari_scalaro_request_t *req = args->smsg;
mach_io_connect_method_scalari_scalaro_reply_t *rep = args->rmsg;
size_t *msglen = args->rsize;
+ In ioframebuffer driver: We try to map the console framebuffer through wsdisplay. It seems to work, but we still need to _see_ something. XXX Framebuffer size discovery is borken. + In ioframebuffer and iohidsystem: More methods and more debug for io_connect_method_* + A few new IOKit mach traps.
2003-05-23 02:07:36 +04:00
int maxoutcount;
Match IOHIDSystem and IOFramebuffer request in the IOKit and handle them in Darwin specific files. The handler does nothing yet. This needs some improvement. Darwin attaches an IOHIDSystem device to each human-interface device in the tree: mouse, keyboard, and a few others. For now we only emulate one IOHIDSystem device.
2003-02-16 18:02:05 +03:00
#ifdef DEBUG_DARWIN
Introduced IOKit driver methods. Moved back into Mach some IOKit driver support that was in Darwin code. I get an empty darwin_iokit.c, which is getting a good candidate for cvs delete.
2003-02-21 01:39:42 +03:00
printf("darwin_iohidsystem_connect_method_scalari_scalaro()\n");
Match IOHIDSystem and IOFramebuffer request in the IOKit and handle them in Darwin specific files. The handler does nothing yet. This needs some improvement. Darwin attaches an IOHIDSystem device to each human-interface device in the tree: mouse, keyboard, and a few others. For now we only emulate one IOHIDSystem device.
2003-02-16 18:02:05 +03:00
#endif
rep->rep_msgh.msgh_bits =
MACH_MSGH_REPLY_LOCAL_BITS(MACH_MSG_TYPE_MOVE_SEND_ONCE);
rep->rep_msgh.msgh_size = sizeof(*rep) - sizeof(rep->rep_trailer);
rep->rep_msgh.msgh_local_port = req->req_msgh.msgh_local_port;
rep->rep_msgh.msgh_id = req->req_msgh.msgh_id + 100;
rep->rep_outcount = 0;
+ In ioframebuffer driver: We try to map the console framebuffer through wsdisplay. It seems to work, but we still need to _see_ something. XXX Framebuffer size discovery is borken. + In ioframebuffer and iohidsystem: More methods and more debug for io_connect_method_* + A few new IOKit mach traps.
2003-05-23 02:07:36 +04:00
Added a IOFBSETCOLORCONVERTTABLE method in ioframebuffer driver. Added sanity check for user provided offsets in messages.
2003-07-01 23:15:47 +04:00
/* Sanity check req->req_incount */
if (MACH_REQMSG_OVERFLOW(args, req->req_in[req->req_incount]))
return mach_msg_error(args, EINVAL);
maxoutcount = req->req_in[req->req_incount];
+ In ioframebuffer driver: We try to map the console framebuffer through wsdisplay. It seems to work, but we still need to _see_ something. XXX Framebuffer size discovery is borken. + In ioframebuffer and iohidsystem: More methods and more debug for io_connect_method_* + A few new IOKit mach traps.
2003-05-23 02:07:36 +04:00
switch (req->req_selector) {
case DARWIN_IOHIDCREATESHMEM: {
/* Create the shared memory for HID events */
int version;
int error;
size_t memsize;
vaddr_t kvaddr;
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
struct proc *p;
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
struct darwin_iohidsystem_thread_args *dita;
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
struct darwin_emuldata *ded;
+ In ioframebuffer driver: We try to map the console framebuffer through wsdisplay. It seems to work, but we still need to _see_ something. XXX Framebuffer size discovery is borken. + In ioframebuffer and iohidsystem: More methods and more debug for io_connect_method_* + A few new IOKit mach traps.
2003-05-23 02:07:36 +04:00
version = req->req_in[0]; /* 1 */
#ifdef DEBUG_DARWIN
printf("DARWIN_IOHIDCREATESHMEM: version = %d\n", version);
#endif
memsize = round_page(sizeof(struct darwin_iohidsystem_shmem));
/* If it has not been used yet, initialize it */
if (darwin_iohidsystem_shmem == NULL) {
darwin_iohidsystem_shmem = uao_create(memsize, 0);
error = uvm_map(kernel_map, &kvaddr, memsize,
darwin_iohidsystem_shmem, 0, PAGE_SIZE,
UVM_MAPFLAG(UVM_PROT_RW, UVM_PROT_RW,
UVM_INH_SHARE, UVM_ADV_RANDOM, 0));
if (error != 0) {
uao_detach(darwin_iohidsystem_shmem);
darwin_iohidsystem_shmem = NULL;
return mach_msg_error(args, error);
}
error = uvm_map_pageable(kernel_map, kvaddr,
kvaddr + memsize, FALSE, 0);
if (error != 0) {
uao_detach(darwin_iohidsystem_shmem);
darwin_iohidsystem_shmem = NULL;
return mach_msg_error(args, error);
}
darwin_iohidsystem_shmeminit(kvaddr);
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
p = args->l->l_proc;
ded = (struct darwin_emuldata *)p->p_emuldata;
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
dita = malloc(sizeof(*dita), M_TEMP, M_WAITOK);
dita->dita_shmem = kvaddr;
Sync between parent and child in iohdisystem thread so that the child has time to get its parmeteres before the parent throw them away
2003-09-13 11:56:54 +04:00
dita->dita_done = 0;
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
+ In ioframebuffer driver: We try to map the console framebuffer through wsdisplay. It seems to work, but we still need to _see_ something. XXX Framebuffer size discovery is borken. + In ioframebuffer and iohidsystem: More methods and more debug for io_connect_method_* + A few new IOKit mach traps.
2003-05-23 02:07:36 +04:00
kthread_create1(darwin_iohidsystem_thread,
Sync between parent and child in iohdisystem thread so that the child has time to get its parmeteres before the parent throw them away
2003-09-13 11:56:54 +04:00
(void *)dita, &dita->dita_p, "iohidsystem");
/*
* Make sure the thread got the informations
* before exitting and destroying dita.
*/
while (!dita->dita_done)
(void)tsleep(&dita->dita_done,
PZERO, "iohid_done", 0);
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
ded->ded_hidsystem_finished =
dita->dita_hidsystem_finished;
Sync between parent and child in iohdisystem thread so that the child has time to get its parmeteres before the parent throw them away
2003-09-13 11:56:54 +04:00
free(dita, M_TEMP);
+ In ioframebuffer driver: We try to map the console framebuffer through wsdisplay. It seems to work, but we still need to _see_ something. XXX Framebuffer size discovery is borken. + In ioframebuffer and iohidsystem: More methods and more debug for io_connect_method_* + A few new IOKit mach traps.
2003-05-23 02:07:36 +04:00
}
rep->rep_outcount = 0;
break;
}
case DARWIN_IOHIDSETEVENTSENABLE: {
/* Enable or disable events */
int enable;
enable = req->req_in[0];
#ifdef DEBUG_DARWIN
printf("DARWIN_IOHIDSETEVENTSENABLE: enable = %d\n", enable);
#endif
/* For now, this is a no-op */
rep->rep_outcount = 0;
break;
}
case DARWIN_IOHIDSETCURSORENABLE: {
/* Enable or disable the cursor */
int enable;
enable = req->req_in[0];
#ifdef DEBUG_DARWIN
printf("DARWIN_IOHIDSETCURSORENABLE: enable = %d\n", enable);
#endif
/* We don't support it */
rep->rep_outcount = 0;
break;
}
default:
#ifdef DEBUG_DARWIN
printf("Unknown selector %d\n", req->req_selector);
#endif
return mach_msg_error(args, EINVAL);
break;
}
Match IOHIDSystem and IOFramebuffer request in the IOKit and handle them in Darwin specific files. The handler does nothing yet. This needs some improvement. Darwin attaches an IOHIDSystem device to each human-interface device in the tree: mouse, keyboard, and a few others. For now we only emulate one IOHIDSystem device.
2003-02-16 18:02:05 +03:00
rep->rep_out[rep->rep_outcount + 1] = 8; /* XXX Trailer */
+ In ioframebuffer driver: We try to map the console framebuffer through wsdisplay. It seems to work, but we still need to _see_ something. XXX Framebuffer size discovery is borken. + In ioframebuffer and iohidsystem: More methods and more debug for io_connect_method_* + A few new IOKit mach traps.
2003-05-23 02:07:36 +04:00
*msglen = sizeof(*rep) - ((16 + rep->rep_outcount) * sizeof(int));
rep->rep_msgh.msgh_size = *msglen - sizeof(rep->rep_trailer);
Match IOHIDSystem and IOFramebuffer request in the IOKit and handle them in Darwin specific files. The handler does nothing yet. This needs some improvement. Darwin attaches an IOHIDSystem device to each human-interface device in the tree: mouse, keyboard, and a few others. For now we only emulate one IOHIDSystem device.
2003-02-16 18:02:05 +03:00
return 0;
}
Added the map_memory method for the IOKit, plus an implementation for the IOHIDSystem driver in Darwin. IOHIDSystem provides access to user interface devices. Communication between userspace and kernel is done through a shared memory page. Our plan is to have a kernel thread waiting on wscons events (for now it just does nothing), and have it feeding the shared page.
2003-04-30 02:16:38 +04:00
Remove references to wsmux_cdevsw and wsdisplay_cdevsw (how to remove them completely?) Add SETCURSORPOSITION operation
2003-10-18 17:27:17 +04:00
int
darwin_iohidsystem_connect_method_structi_structo(args)
struct mach_trap_args *args;
{
mach_io_connect_method_structi_structo_request_t *req = args->smsg;
mach_io_connect_method_structi_structo_reply_t *rep = args->rmsg;
size_t *msglen = args->rsize;
struct proc *p = args->l->l_proc;
int maxoutcount;
int error;
#ifdef DEBUG_DARWIN
printf("darwin_iohidsystem_connect_method_structi_structo()\n");
#endif
rep->rep_msgh.msgh_bits =
MACH_MSGH_REPLY_LOCAL_BITS(MACH_MSG_TYPE_MOVE_SEND_ONCE);
rep->rep_msgh.msgh_size = sizeof(*rep) - sizeof(rep->rep_trailer);
rep->rep_msgh.msgh_local_port = req->req_msgh.msgh_local_port;
rep->rep_msgh.msgh_id = req->req_msgh.msgh_id + 100;
rep->rep_outcount = 0;
/* Sanity check req->req_incount */
if (MACH_REQMSG_OVERFLOW(args, req->req_in[req->req_incount]))
return mach_msg_error(args, EINVAL);
maxoutcount = req->req_in[req->req_incount];
switch (req->req_selector) {
case DARWIN_IOHIDSETMOUSELOCATION: {
struct wscons_event wsevt;
dev_t dev;
const struct cdevsw *wsmux;
darwin_iogpoint *pt = (darwin_iogpoint *)&req->req_in[0];
#ifdef DEBUG_DARWIN
printf("DARWIN_IOHIDSETMOUSELOCATION: %d,%d\n", pt->x, pt->y);
#endif
/*
* Use the wsmux given by sysctl emul.darwin.iohidsystem_mux
*/
error = darwin_findwsmux(&dev, darwin_iohidsystem_mux);
if (error != 0)
return mach_msg_error(args, error);
if ((wsmux = cdevsw_lookup(dev)) == NULL)
return mach_msg_error(args, ENXIO);
wsevt.type = WSCONS_EVENT_MOUSE_ABSOLUTE_X;
wsevt.value = pt->x;
if ((error = (wsmux->d_ioctl)(dev,
WSMUXIO_INJECTEVENT, (caddr_t)&wsevt, 0, p)) != 0)
return mach_msg_error(args, error);
wsevt.type = WSCONS_EVENT_MOUSE_ABSOLUTE_Y;
wsevt.value = pt->y;
if ((error = (wsmux->d_ioctl)(dev,
WSMUXIO_INJECTEVENT, (caddr_t)&wsevt, 0, p)) != 0)
return mach_msg_error(args, error);
rep->rep_outcount = 0;
break;
}
default:
#ifdef DEBUG_DARWIN
printf("Unknown selector %d\n", req->req_selector);
#endif
return mach_msg_error(args, EINVAL);
break;
}
rep->rep_out[rep->rep_outcount + 1] = 8; /* XXX Trailer */
*msglen = sizeof(*rep) - (4096 - rep->rep_outcount);
rep->rep_msgh.msgh_size = *msglen - sizeof(rep->rep_trailer);
return 0;
}
Added the map_memory method for the IOKit, plus an implementation for the IOHIDSystem driver in Darwin. IOHIDSystem provides access to user interface devices. Communication between userspace and kernel is done through a shared memory page. Our plan is to have a kernel thread waiting on wscons events (for now it just does nothing), and have it feeding the shared page.
2003-04-30 02:16:38 +04:00
int
darwin_iohidsystem_connect_map_memory(args)
struct mach_trap_args *args;
{
mach_io_connect_map_memory_request_t *req = args->smsg;
mach_io_connect_map_memory_reply_t *rep = args->rmsg;
size_t *msglen = args->rsize;
struct proc *p = args->l->l_proc;
int error;
size_t memsize;
vaddr_t pvaddr;
#ifdef DEBUG_DARWIN
printf("darwin_iohidsystem_connect_map_memory()\n");
#endif
memsize = round_page(sizeof(struct darwin_iohidsystem_shmem));
+ In ioframebuffer driver: We try to map the console framebuffer through wsdisplay. It seems to work, but we still need to _see_ something. XXX Framebuffer size discovery is borken. + In ioframebuffer and iohidsystem: More methods and more debug for io_connect_method_* + A few new IOKit mach traps.
2003-05-23 02:07:36 +04:00
if (darwin_iohidsystem_shmem == NULL)
return mach_msg_error(args, ENOMEM);
Added the map_memory method for the IOKit, plus an implementation for the IOHIDSystem driver in Darwin. IOHIDSystem provides access to user interface devices. Communication between userspace and kernel is done through a shared memory page. Our plan is to have a kernel thread waiting on wscons events (for now it just does nothing), and have it feeding the shared page.
2003-04-30 02:16:38 +04:00
uao_reference(darwin_iohidsystem_shmem);
pvaddr = VM_DEFAULT_ADDRESS(p->p_vmspace->vm_daddr, memsize);
if ((error = uvm_map(&p->p_vmspace->vm_map, &pvaddr,
memsize, darwin_iohidsystem_shmem, 0, PAGE_SIZE,
UVM_MAPFLAG(UVM_PROT_RW, UVM_PROT_RW,
UVM_INH_SHARE, UVM_ADV_RANDOM, 0))) != 0)
return mach_msg_error(args, error);
#ifdef DEBUG_DARWIN
printf("pvaddr = 0x%08lx\n", (long)pvaddr);
#endif
rep->rep_msgh.msgh_bits =
MACH_MSGH_REPLY_LOCAL_BITS(MACH_MSG_TYPE_MOVE_SEND_ONCE);
rep->rep_msgh.msgh_size = sizeof(*rep) - sizeof(rep->rep_trailer);
rep->rep_msgh.msgh_local_port = req->req_msgh.msgh_local_port;
rep->rep_msgh.msgh_id = req->req_msgh.msgh_id + 100;
rep->rep_retval = 0;
rep->rep_addr = pvaddr;
rep->rep_len = sizeof(struct darwin_iohidsystem_shmem);
rep->rep_trailer.msgh_trailer_size = 8;
*msglen = sizeof(*rep);
return 0;
}
static void
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
darwin_iohidsystem_thread(args)
void *args;
Added the map_memory method for the IOKit, plus an implementation for the IOHIDSystem driver in Darwin. IOHIDSystem provides access to user interface devices. Communication between userspace and kernel is done through a shared memory page. Our plan is to have a kernel thread waiting on wscons events (for now it just does nothing), and have it feeding the shared page.
2003-04-30 02:16:38 +04:00
{
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
struct darwin_iohidsystem_thread_args *dita;
struct darwin_iohidsystem_shmem *shmem;
struct darwin_iohidsystem_evglobals *evg;
darwin_iohidsystem_event_item *diei;
darwin_iohidsystem_event *die;
dev_t dev;
Remove references to wsmux_cdevsw and wsdisplay_cdevsw (how to remove them completely?) Add SETCURSORPOSITION operation
2003-10-18 17:27:17 +04:00
const struct cdevsw *wsmux;
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
struct uio auio;
struct iovec aiov;
struct wscons_event wsevt;
struct proc *p;
int error = 0;
struct mach_right *mr;
struct lwp *l;
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
int finished = 0;
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
Added the map_memory method for the IOKit, plus an implementation for the IOHIDSystem driver in Darwin. IOHIDSystem provides access to user interface devices. Communication between userspace and kernel is done through a shared memory page. Our plan is to have a kernel thread waiting on wscons events (for now it just does nothing), and have it feeding the shared page.
2003-04-30 02:16:38 +04:00
#ifdef DEBUG_DARWIN
printf("darwin_iohidsystem_thread: start\n");
#endif
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
dita = (struct darwin_iohidsystem_thread_args *)args;
Sync between parent and child in iohdisystem thread so that the child has time to get its parmeteres before the parent throw them away
2003-09-13 11:56:54 +04:00
shmem = (struct darwin_iohidsystem_shmem *)dita->dita_shmem;
p = dita->dita_p;
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
l = proc_representative_lwp(p);
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
dita->dita_hidsystem_finished = &finished;
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
Sync between parent and child in iohdisystem thread so that the child has time to get its parmeteres before the parent throw them away
2003-09-13 11:56:54 +04:00
/*
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
* Allow the parent to read dita_hidsystem_finished
* and to get rid of dita. Once the parent is awaken,
* it holds a reference to our on-stack finished flag,
* hence we cannot exi before the parent sets this flag.
* This is done in darwin_proc_exit()
Sync between parent and child in iohdisystem thread so that the child has time to get its parmeteres before the parent throw them away
2003-09-13 11:56:54 +04:00
*/
dita->dita_done = 1;
wakeup(&dita->dita_done);
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
evg = (struct darwin_iohidsystem_evglobals *)&shmem->dis_evglobals;
Handle mouseup and mousdown events Add three sysctl to choose wsdisplay unit and screen and wsmux that ioframebuffer and iohidsystem will use.
2003-09-14 13:48:42 +04:00
/*
* Use the wsmux given by sysctl emul.darwin.iohidsystem_mux
*/
if ((error = darwin_findwsmux(&dev, darwin_iohidsystem_mux)) != 0)
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
goto out2;
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
Remove references to wsmux_cdevsw and wsdisplay_cdevsw (how to remove them completely?) Add SETCURSORPOSITION operation
2003-10-18 17:27:17 +04:00
if ((wsmux = cdevsw_lookup(dev)) == NULL) {
error = ENXIO;
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
goto out2;
Remove references to wsmux_cdevsw and wsdisplay_cdevsw (how to remove them completely?) Add SETCURSORPOSITION operation
2003-10-18 17:27:17 +04:00
}
if ((error = (wsmux->d_open)(dev, FREAD|FWRITE, 0, p)) != 0)
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
goto out2;
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
while(!finished) {
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
auio.uio_iov = &aiov;
auio.uio_iovcnt = 1;
aiov.iov_base = &wsevt;
aiov.iov_len = sizeof(wsevt);
auio.uio_resid = sizeof(wsevt);
auio.uio_offset = 0;
auio.uio_segflg = UIO_SYSSPACE;
auio.uio_rw = UIO_READ;
auio.uio_procp = p;
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
if ((error = (wsmux->d_read)(dev, &auio, 0)) != 0) {
#ifdef DEBUG_DARWIN
printf("iohidsystem: read error %d\n", error);
#endif
goto out1;
}
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
diei = &evg->evg_evqueue[evg->evg_event_last];
while (diei->diei_sem != 0)
tsleep((void *)&diei->diei_sem, PZERO, "iohid_lock", 1);
/*
* No need to take the lock since we will not
* yield control to the user process.
*/
diei->diei_next = evg->evg_event_tail;
diei = &evg->evg_evqueue[evg->evg_event_tail];
die = (darwin_iohidsystem_event *)&diei->diei_event;
darwin_wscons_to_iohidsystem(&wsevt, die);
evg->evg_event_last = evg->evg_event_tail;
evg->evg_event_tail++;
if (evg->evg_event_tail == DARWIN_IOHIDSYSTEM_EVENTQUEUE_LEN)
evg->evg_event_tail = 0;
Report keyup and keydown events Qhen the event queue is full, sleep so that the process will be able to consume entries
2003-09-30 23:56:54 +04:00
/*
* Send a I/O notification if the process
* has consumed all available entries
*/
if (evg->evg_event_last == evg->evg_event_head) {
mr = darwin_iohidsystem_devclass.mid_notify;
if (mr != NULL)
mach_notify_iohidsystem(l, mr);
}
/*
* If the queue is full, ie: the next event slot is not
* yet consumed, sleep until the process consumes it.
*/
diei = &evg->evg_evqueue[evg->evg_event_tail];
die = (darwin_iohidsystem_event *)&diei->diei_event;
while (die->die_type != 0)
tsleep((void *)&die->die_type, PZERO, "iohid_full", 1);
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
}
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
out1:
(wsmux->d_close)(dev, FREAD|FWRITE, 0, p);
out2:
while (!finished)
tsleep((void *)&finished, PZERO, "iohid_exit", 0);
uao_detach(darwin_iohidsystem_shmem);
darwin_iohidsystem_shmem = NULL;
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
kthread_exit(error);
/* NOTREACHED */
Added the map_memory method for the IOKit, plus an implementation for the IOHIDSystem driver in Darwin. IOHIDSystem provides access to user interface devices. Communication between userspace and kernel is done through a shared memory page. Our plan is to have a kernel thread waiting on wscons events (for now it just does nothing), and have it feeding the shared page.
2003-04-30 02:16:38 +04:00
};
static void
darwin_iohidsystem_shmeminit(kvaddr)
vaddr_t kvaddr;
{
struct darwin_iohidsystem_shmem *shmem;
struct darwin_iohidsystem_evglobals *evglobals;
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
int i;
Added the map_memory method for the IOKit, plus an implementation for the IOHIDSystem driver in Darwin. IOHIDSystem provides access to user interface devices. Communication between userspace and kernel is done through a shared memory page. Our plan is to have a kernel thread waiting on wscons events (for now it just does nothing), and have it feeding the shared page.
2003-04-30 02:16:38 +04:00
shmem = (struct darwin_iohidsystem_shmem *)kvaddr;
shmem->dis_global_offset =
(size_t)&shmem->dis_evglobals - (size_t)&shmem->dis_global_offset;
shmem->dis_private_offset =
shmem->dis_global_offset + sizeof(*evglobals);
evglobals = &shmem->dis_evglobals;
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
evglobals->evg_struct_size = sizeof(*evglobals);
for (i = 0; i < DARWIN_IOHIDSYSTEM_EVENTQUEUE_LEN; i++)
evglobals->evg_evqueue[i].diei_next = i + 1;
evglobals->
evg_evqueue[DARWIN_IOHIDSYSTEM_EVENTQUEUE_LEN - 1].diei_next = 0;
return;
}
static int
darwin_findwsmux(dev, mux)
dev_t *dev;
int mux;
{
struct wsmux_softc *wsm_sc;
int minor, major;
if ((wsm_sc = wsmux_getmux(mux)) == NULL)
return ENODEV;
major = cdevsw_lookup_major(&wsmux_cdevsw);
minor = wsm_sc->sc_base.me_dv.dv_unit;
*dev = makedev(major, minor);
return 0;
}
static void
darwin_wscons_to_iohidsystem(wsevt, hidevt)
struct wscons_event *wsevt;
darwin_iohidsystem_event *hidevt;
{
struct timeval tv;
static int px = 0; /* Previous mouse location */
static int py = 0;
static int pf = 0; /* previous kbd flags */
microtime(&tv);
(void)memset((void *)hidevt, 0, sizeof(*hidevt));
hidevt->die_time_hi = tv.tv_sec;
hidevt->die_time_lo = tv.tv_usec;
hidevt->die_location_x = px;
hidevt->die_location_y = py;
hidevt->die_flags = pf;
switch (wsevt->type) {
case WSCONS_EVENT_MOUSE_DELTA_X:
hidevt->die_type = DARWIN_NX_MOUSEMOVED;
px += wsevt->value;
hidevt->die_data.mouse_move.dx = wsevt->value;
hidevt->die_location_x = px;
break;
Added the map_memory method for the IOKit, plus an implementation for the IOHIDSystem driver in Darwin. IOHIDSystem provides access to user interface devices. Communication between userspace and kernel is done through a shared memory page. Our plan is to have a kernel thread waiting on wscons events (for now it just does nothing), and have it feeding the shared page.
2003-04-30 02:16:38 +04:00
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
case WSCONS_EVENT_MOUSE_DELTA_Y:
hidevt->die_type = DARWIN_NX_MOUSEMOVED;
Darwin's I/O reverse the mouse's Y axis.
2003-09-13 20:35:47 +04:00
py -= wsevt->value;
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
hidevt->die_data.mouse_move.dy = wsevt->value;
hidevt->die_location_y = py;
break;
case WSCONS_EVENT_MOUSE_ABSOLUTE_X:
hidevt->die_type = DARWIN_NX_MOUSEMOVED;
hidevt->die_location_x = wsevt->value;
px = wsevt->value;
break;
case WSCONS_EVENT_MOUSE_ABSOLUTE_Y:
hidevt->die_type = DARWIN_NX_MOUSEMOVED;
hidevt->die_location_y = wsevt->value;
py = wsevt->value;
break;
Handle mouseup and mousdown events Add three sysctl to choose wsdisplay unit and screen and wsmux that ioframebuffer and iohidsystem will use.
2003-09-14 13:48:42 +04:00
case WSCONS_EVENT_MOUSE_DOWN:
if (wsevt->value == 0)
hidevt->die_type = DARWIN_NX_LMOUSEDOWN;
else if (wsevt->value == 1)
hidevt->die_type = DARWIN_NX_RMOUSEDOWN;
else {
#ifdef DEBUG_DARWIN
printf("Unknown mouse button %d\n", wsevt->value);
#endif
break;
}
hidevt->die_data.mouse.subx = px;
hidevt->die_data.mouse.suby = py;
hidevt->die_data.mouse.buttonid = wsevt->value;
break;
case WSCONS_EVENT_MOUSE_UP:
if (wsevt->value == 0)
hidevt->die_type = DARWIN_NX_LMOUSEUP;
else if (wsevt->value == 1)
hidevt->die_type = DARWIN_NX_RMOUSEUP;
else {
#ifdef DEBUG_DARWIN
printf("Unknown mouse button %d\n", wsevt->value);
#endif
break;
}
hidevt->die_data.mouse.subx = px;
hidevt->die_data.mouse.suby = py;
hidevt->die_data.mouse.buttonid = wsevt->value;
break;
Report keyup and keydown events Qhen the event queue is full, sleep so that the process will be able to consume entries
2003-09-30 23:56:54 +04:00
case WSCONS_EVENT_KEY_DOWN:
hidevt->die_type = DARWIN_NX_KEYDOWN;
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
hidevt->die_data.key.charcode = wsevt->value;
hidevt->die_data.key.orig_charcode = wsevt->value;
hidevt->die_data.key.keycode = wsevt->value; /* Translate */
hidevt->die_data.key.keyboardtype = 0xcd; /* XXX */
Report keyup and keydown events Qhen the event queue is full, sleep so that the process will be able to consume entries
2003-09-30 23:56:54 +04:00
break;
case WSCONS_EVENT_KEY_UP:
hidevt->die_type = DARWIN_NX_KEYUP;
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
hidevt->die_data.key.charcode = wsevt->value;
hidevt->die_data.key.orig_charcode = wsevt->value;
hidevt->die_data.key.keycode = wsevt->value; /* Translate */
hidevt->die_data.key.keyboardtype = 0xcd; /* XXX */
Report keyup and keydown events Qhen the event queue is full, sleep so that the process will be able to consume entries
2003-09-30 23:56:54 +04:00
break;
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
default:
Handle mouseup and mousdown events Add three sysctl to choose wsdisplay unit and screen and wsmux that ioframebuffer and iohidsystem will use.
2003-09-14 13:48:42 +04:00
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
#ifdef DEBUG_DARWIN
printf("Untranslated wsevt->type = %d, wsevt->value = %d\n",
wsevt->type, wsevt->value);
#endif
break;
}
return;
}
static void
mach_notify_iohidsystem(l, mr)
struct lwp *l;
struct mach_right *mr;
{
struct mach_port *mp;
mach_notify_iohidsystem_request_t *req;
mp = mr->mr_port;
#ifdef DEBUG_MACH
if (mp == NULL) {
printf("Warning: notification right without a port\n");
return;
}
#endif
req = malloc(sizeof(*req), M_EMULDATA, M_WAITOK | M_ZERO);
req->req_msgh.msgh_bits =
MACH_MSGH_REPLY_LOCAL_BITS(MACH_MSG_TYPE_MOVE_SEND_ONCE);
req->req_msgh.msgh_size = sizeof(*req) - sizeof(req->req_trailer);
req->req_msgh.msgh_local_port = mr->mr_name;
req->req_msgh.msgh_id = 0;
req->req_trailer.msgh_trailer_size = 8;
#ifdef KTRACE
ktruser(l->l_proc, "notify_iohidsystem", NULL, 0, 0);
#endif
mr->mr_refcount++;
Bugfix: notifications sent by iohidsystem kernel threads are to be hnalded as mmessages coming from the kernel, since the threas has no p->p_emuldata
2003-09-13 20:07:44 +04:00
/*
* Call mach_message_get with a NULL lwp: the message is sent
* by the kernel, not a userprocess. If we do not do that,
* we crash because iohidsystem thread has p->p_emuldata == NULL.
*/
mach_message_get((mach_msg_header_t *)req, sizeof(*req), mp, NULL);
Some initial support for IOHIDSystem. This gives us some pointer moves in X11 (not really related the the actual mouse movement, but this will come). The darwin_iohidsystem_thread reads events from wscons, translates them into IOHIDSystem events and wakes up the userland client with a notification. To do this, I had to improve the void implementation of io_connect_set_notification_port() to actually register something (I assumed a single notification port, which makes some sense since only one process seems to be able to open the driver) Missing bits: - we do not take event masks given by the process into account. - the notification message has not been checked against Darwin - events are badly translated
2003-09-12 03:16:18 +04:00
#ifdef DEBUG_MACH_MSG
printf("pid %d: message queued on port %p (%d) [%p]\n",
l->l_proc->p_pid, mp, req->req_msgh.msgh_id,
mp->mp_recv->mr_sethead);
#endif
wakeup(mp->mp_recv->mr_sethead);
Added the map_memory method for the IOKit, plus an implementation for the IOHIDSystem driver in Darwin. IOHIDSystem provides access to user interface devices. Communication between userspace and kernel is done through a shared memory page. Our plan is to have a kernel thread waiting on wscons events (for now it just does nothing), and have it feeding the shared page.
2003-04-30 02:16:38 +04:00
return;
}
Correctly terminate the iohidsystem thread (with appropriate cleanups of the shared memory), when the display server quits. We are now able to restart the X server without the need to reboot the machine. That's better.
2003-10-25 14:43:45 +04:00
void
darwin_iohidsystem_postfake(p)
struct proc *p;
{
const struct cdevsw *wsmux;
dev_t dev;
int error;
struct wscons_event wsevt;
/*
* Use the wsmux given by sysctl emul.darwin.iohidsystem_mux
*/
if ((error = darwin_findwsmux(&dev, darwin_iohidsystem_mux)) != 0)
return;
if ((wsmux = cdevsw_lookup(dev)) == NULL)
return;
wsevt.type = 0;
wsevt.value = 0;
(wsmux->d_ioctl)(dev, WSMUXIO_INJECTEVENT, (caddr_t)&wsevt, 0, p);
return;
}
Reference in New Issue Copy Permalink