d/freebsd-nq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
02dbda9d17
freebsd-nq/sys/dev/sound/pcm/dsp.c

1234 lines
28 KiB
C
Raw Normal View History

Start each of the license/copyright comments with /*-, minor shuffle of lines
2005-01-06 01:43:34 +00:00
/*-
update my email address.
2003-09-07 16:28:03 +00:00
* Copyright (c) 1999 Cameron Grant <cg@freebsd.org>
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR 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 AUTHOR 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/param.h>
#include <sys/queue.h>
repo-copied to make way for newmidi, this commit updates include paths
1999-11-20 16:50:33 +00:00
#include <dev/sound/pcm/sound.h>
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
SND_DECLARE_FILE("$FreeBSD$");
fix the return value of the SOUND_PCM_WRITE_CHANNELS ioctl re-enable old pcm ioctls
2000-05-27 14:40:17 +00:00
#define OLDPCM_IOCTL
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
static d_open_t dsp_open;
static d_close_t dsp_close;
static d_read_t dsp_read;
static d_write_t dsp_write;
static d_ioctl_t dsp_ioctl;
static d_poll_t dsp_poll;
static d_mmap_t dsp_mmap;
Fix a panic when kldloading a sound driver. Do this by replacing the link-list of dev_t's with named variables. Remove used code. Approved by: tanimura (mentor)
2004-01-20 03:58:57 +00:00
struct cdevsw dsp_cdevsw = {
Device megapatch 4/6: Introduce d_version field in struct cdevsw, this must always be initialized to D_VERSION. Flip sense of D_NOGIANT flag to D_NEEDGIANT, this involves removing four D_NOGIANT flags and adding 145 D_NEEDGIANT flags.
2004-02-21 21:10:55 +00:00
.d_version = D_VERSION,
.d_flags = D_NEEDGIANT,
Gigacommit to improve device-driver source compatibility between branches: Initialize struct cdevsw using C99 sparse initializtion and remove all initializations to default values. This patch is automatically generated and has been tested by compiling LINT with all the fields in struct cdevsw in reverse order on alpha, sparc64 and i386. Approved by: re(scottl)
2003-03-03 12:15:54 +00:00
.d_open = dsp_open,
.d_close = dsp_close,
.d_read = dsp_read,
.d_write = dsp_write,
.d_ioctl = dsp_ioctl,
.d_poll = dsp_poll,
.d_mmap = dsp_mmap,
.d_name = "dsp",
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
};
add defines and ifdefs so this code will compile on 4.x add spls so this code will work on 4.x
2001-06-23 17:36:51 +00:00
#ifdef USING_DEVFS
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
static eventhandler_tag dsp_ehtag;
add defines and ifdefs so this code will compile on 4.x add spls so this code will work on 4.x
2001-06-23 17:36:51 +00:00
#endif
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
static struct snddev_info *
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
dsp_get_info(struct cdev *dev)
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
{
struct snddev_info *d;
int unit;
unit = PCMUNIT(dev);
use devclass_get_maxunit() correctly
2001-06-18 00:10:47 +00:00
if (unit >= devclass_get_maxunit(pcm_devclass))
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
return NULL;
d = devclass_get_softc(pcm_devclass, unit);
return d;
}
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
static u_int32_t
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
dsp_get_flags(struct cdev *dev)
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
{
device_t bdev;
int unit;
unit = PCMUNIT(dev);
if (unit >= devclass_get_maxunit(pcm_devclass))
return 0xffffffff;
bdev = devclass_get_device(pcm_devclass, unit);
return pcm_getflags(bdev);
}
static void
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
dsp_set_flags(struct cdev *dev, u_int32_t flags)
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
{
device_t bdev;
int unit;
unit = PCMUNIT(dev);
if (unit >= devclass_get_maxunit(pcm_devclass))
return;
bdev = devclass_get_device(pcm_devclass, unit);
pcm_setflags(bdev, flags);
}
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
/*
Fix typos and clean up some comments. Approved by: ariff
2006-01-25 21:13:46 +00:00
* return the channels associated with an open device instance.
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
* set the priority if the device is simplex and one direction (only) is
* specified.
* lock channels specified.
*/
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
static int
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
getchns(struct cdev *dev, struct pcm_channel **rdch, struct pcm_channel **wrch, u_int32_t prio)
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
{
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
struct snddev_info *d;
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
u_int32_t flags;
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
flags = dsp_get_flags(dev);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
d = dsp_get_info(dev);
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
pcm_inprog(d, 1);
Change KASSERT() in feed_vchan16() into an explicit test and call to panic() so that the buffer overflow just beyond this point is always caught, even when the code is not compiled with INVARIANTS. Change chn_setblocksize() buffer reallocation code to attempt to avoid the feed_vchan16() buffer overflow by attempting to always keep the bufsoft buffer at least as large as the bufhard buffer. Print a diagnositic message Danger! %s bufsoft size increasing from %d to %d after CHANNEL_SETBLOCKSIZE() if our best attempts fail. If feed_vchan16() were to be called by the interrupt handler while locks are dropped in chn_setblocksize() to increase the size bufsoft to match the size of bufhard, the panic() code in feed_vchan16() will be triggered. If the diagnostic message is printed, it is a warning that a panic is possible if the system were to see events in an "unlucky" order. Change the locking code to avoid the need for MTX_RECURSIVE mutexes. Add the MTX_DUPOK option to the channel mutexes and change the locking sequence to always lock the parent channel before its children to avoid the possibility of deadlock. Actually implement locking assertions for the channel mutexes and fix the problems found by the resulting assertion violations. Clean up the locking code in dsp_ioctl(). Allocate the channel buffers using the malloc() M_WAITOK option instead of M_NOWAIT so that buffer allocation won't fail. Drop locks across the malloc() calls. Add/modify KASSERTS() in attempt to detect problems early. Abuse layering by adding a pointer to the snd_dbuf structure that points back to the pcm_channel that owns it. This allows sndbuf_resize() to do proper locking without having to change the its API, which is used by the hardware drivers. Don't dereference a NULL pointer when setting hw.snd.maxautovchans if a hardware driver is not loaded. Noticed by Ryan Sommers <ryans at gamersimpact.com>. Tested by: Stefan Ehmann <shoesoft AT gmx.net> Tested by: matk (Mathew Kanner) Tested by: Gordon Bergling <gbergling AT 0xfce3.net>
2004-01-28 08:02:15 +00:00
pcm_lock(d);
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
KASSERT((flags & SD_F_PRIO_SET) != SD_F_PRIO_SET, \
fix dma underrun issues mutate some panics to kasserts add more spl protection PR: kern/14990 Partially Submitted by: Vladimir N.Silyaev <vns@delta.odessa.ua> Reviewed by: dfr
1999-12-05 19:09:13 +00:00
("getchns: read and write both prioritised"));
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
if ((flags & SD_F_PRIO_SET) == 0 && (prio != (SD_F_PRIO_RD | SD_F_PRIO_WR))) {
flags |= prio & (SD_F_PRIO_RD | SD_F_PRIO_WR);
dsp_set_flags(dev, flags);
}
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
*rdch = dev->si_drv1;
*wrch = dev->si_drv2;
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
if ((flags & SD_F_SIMPLEX) && (flags & SD_F_PRIO_SET)) {
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
if (prio) {
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
if (*rdch && flags & SD_F_PRIO_WR) {
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
dev->si_drv1 = NULL;
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
*rdch = pcm_getfakechan(d);
} else if (*wrch && flags & SD_F_PRIO_RD) {
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
dev->si_drv2 = NULL;
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
*wrch = pcm_getfakechan(d);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
}
}
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
pcm_getfakechan(d)->flags |= CHN_F_BUSY;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
pcm_unlock(d);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
if (*rdch && *rdch != pcm_getfakechan(d) && (prio & SD_F_PRIO_RD))
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
CHN_LOCK(*rdch);
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
if (*wrch && *wrch != pcm_getfakechan(d) && (prio & SD_F_PRIO_WR))
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
CHN_LOCK(*wrch);
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
return 0;
}
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
/* unlock specified channels */
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
static void
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
relchns(struct cdev *dev, struct pcm_channel *rdch, struct pcm_channel *wrch, u_int32_t prio)
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
{
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
struct snddev_info *d;
d = dsp_get_info(dev);
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
if (wrch && wrch != pcm_getfakechan(d) && (prio & SD_F_PRIO_WR))
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
CHN_UNLOCK(wrch);
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
if (rdch && rdch != pcm_getfakechan(d) && (prio & SD_F_PRIO_RD))
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
CHN_UNLOCK(rdch);
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
pcm_inprog(d, -1);
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
static int
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
dsp_open(struct cdev *i_dev, int flags, int mode, struct thread *td)
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
{
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
struct pcm_channel *rdch, *wrch;
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
struct snddev_info *d;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
u_int32_t fmt;
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
int devtype;
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
int rdref;
int error;
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
d = dsp_get_info(i_dev);
devtype = PCMDEV(i_dev);
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
/* decide default format */
switch (devtype) {
case SND_DEV_DSP16:
fmt = AFMT_S16_LE;
break;
case SND_DEV_DSP:
fmt = AFMT_U8;
break;
case SND_DEV_AUDIO:
fmt = AFMT_MU_LAW;
break;
case SND_DEV_NORESET:
fmt = 0;
break;
add a method for recording of specific channels for devices with more than one hardware record channel. new devices, /dev/dsprX.Y where X is unit number and Y is channel index.
2001-09-05 16:28:41 +00:00
case SND_DEV_DSPREC:
fmt = AFMT_U8;
if (mode & FWRITE) {
return EINVAL;
}
break;
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
default:
panic("impossible devtype %d", devtype);
}
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
rdref = 0;
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
/* lock snddev so nobody else can monkey with it */
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
pcm_lock(d);
Marginally simplify dsp_open error handling by adding an early test to determine liklihood of opening device in requested directions. Makes for simpler error handling and change should close kern/35004. PR: kern/35004.
2002-08-28 15:19:30 +00:00
rdch = i_dev->si_drv1;
wrch = i_dev->si_drv2;
if ((dsp_get_flags(i_dev) & SD_F_SIMPLEX) && (rdch || wrch)) {
modify comments
2003-08-15 02:31:13 +00:00
/* we're a simplex device and already open, no go */
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
pcm_unlock(d);
beginnings of virtual playback channel support instead of using two malloced arrays for storing channel lists, use an slist. convert the sndstat device to use sbufs and optionally provide more detail about channel state. vchans are software mixed playback channels. they are not enabled by this commit. they use the feeder infrastructure to emulate normal playback channels in a manner transparent to applications, whilst providing as many channels are desired, especially suitable for devices with only one hardware playback channel. in the future they will provide additional features. those wishing to test this functionality will need to add vchan.c to sys/conf/files and use 'sysctl -w hw.snd.pcm0.vchans' to enable it. blocksize and auto-rate selection are not yet supported.
2001-05-27 17:22:00 +00:00
return EBUSY;
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
}
fix a panic triggerable by anyone with read/write access to the audio devices. opening /dev/{dsp,dspW,audio}0 and then opening a different device from that list and closing it resulted in a panic when any operation is performed on the first fd. we prevent this happening by denying the second open unless it uses the same minor device as the first. PR: kern/25519
2001-03-13 18:43:22 +00:00
Marginally simplify dsp_open error handling by adding an early test to determine liklihood of opening device in requested directions. Makes for simpler error handling and change should close kern/35004. PR: kern/35004.
2002-08-28 15:19:30 +00:00
if (((flags & FREAD) && rdch) || ((flags & FWRITE) && wrch)) {
modify comments
2003-08-15 02:31:13 +00:00
/*
* device already open in one or both directions that
* the opener wants; we can't handle this.
*/
Marginally simplify dsp_open error handling by adding an early test to determine liklihood of opening device in requested directions. Makes for simpler error handling and change should close kern/35004. PR: kern/35004.
2002-08-28 15:19:30 +00:00
pcm_unlock(d);
return EBUSY;
}
beginnings of virtual playback channel support instead of using two malloced arrays for storing channel lists, use an slist. convert the sndstat device to use sbufs and optionally provide more detail about channel state. vchans are software mixed playback channels. they are not enabled by this commit. they use the feeder infrastructure to emulate normal playback channels in a manner transparent to applications, whilst providing as many channels are desired, especially suitable for devices with only one hardware playback channel. in the future they will provide additional features. those wishing to test this functionality will need to add vchan.c to sys/conf/files and use 'sysctl -w hw.snd.pcm0.vchans' to enable it. blocksize and auto-rate selection are not yet supported.
2001-05-27 17:22:00 +00:00
modify comments
2003-08-15 02:31:13 +00:00
/*
* if we get here, the open request is valid- either:
* * we were previously not open
* * we were open for play xor record and the opener wants
* the non-open direction
*/
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
if (flags & FREAD) {
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
/* open for read */
Release lock for a while during chn_reset() / pcm_chnalloc() operation while malloc()ing, this fixes LOR 129. See - http://lists.freebsd.org/pipermail/freebsd-current/2005-June/051157.html - http://lists.freebsd.org/pipermail/freebsd-current/2005-August/054620.html - http://sources.zabbadoz.net/freebsd/lor.html#129 Submitted by: Ariff Abdullah <skywizard@MyBSD.org.my>
2005-09-10 17:51:38 +00:00
pcm_unlock(d);
Marginally simplify dsp_open error handling by adding an early test to determine liklihood of opening device in requested directions. Makes for simpler error handling and change should close kern/35004. PR: kern/35004.
2002-08-28 15:19:30 +00:00
if (devtype == SND_DEV_DSPREC)
rdch = pcm_chnalloc(d, PCMDIR_REC, td->td_proc->p_pid, PCMCHAN(i_dev));
else
rdch = pcm_chnalloc(d, PCMDIR_REC, td->td_proc->p_pid, -1);
if (!rdch) {
/* no channel available, exit */
beginnings of virtual playback channel support instead of using two malloced arrays for storing channel lists, use an slist. convert the sndstat device to use sbufs and optionally provide more detail about channel state. vchans are software mixed playback channels. they are not enabled by this commit. they use the feeder infrastructure to emulate normal playback channels in a manner transparent to applications, whilst providing as many channels are desired, especially suitable for devices with only one hardware playback channel. in the future they will provide additional features. those wishing to test this functionality will need to add vchan.c to sys/conf/files and use 'sysctl -w hw.snd.pcm0.vchans' to enable it. blocksize and auto-rate selection are not yet supported.
2001-05-27 17:22:00 +00:00
return EBUSY;
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
}
Marginally simplify dsp_open error handling by adding an early test to determine liklihood of opening device in requested directions. Makes for simpler error handling and change should close kern/35004. PR: kern/35004.
2002-08-28 15:19:30 +00:00
/* got a channel, already locked for us */
Restore old compatibility of feeding directly into /dev/dsp at 8000 hz while preserving the New World Order. Discussed with: [1] Michael W. Oliver <michael at gargantuan.com> MFC after: 1 week [1] http://lists.freebsd.org/pipermail/freebsd-multimedia/2006-January/003562.html
2006-01-16 11:07:15 +00:00
if (chn_reset(rdch, fmt) ||
(fmt && chn_setspeed(rdch, DSP_DEFAULT_SPEED))) {
Marginally simplify dsp_open error handling by adding an early test to determine liklihood of opening device in requested directions. Makes for simpler error handling and change should close kern/35004. PR: kern/35004.
2002-08-28 15:19:30 +00:00
pcm_chnrelease(rdch);
Fix LOR #174. Tested with: WITNESS, INVARIANTS and DIAGNOSTICS
2005-12-31 01:58:40 +00:00
pcm_lock(d);
Sound devices were being leaked by dsp_open() not actually releasing the channels it allocates if chn_reset() on them resulted in failure. ARROW'D!
2003-02-11 16:58:54 +00:00
i_dev->si_drv1 = NULL;
Fix LOR #174. Tested with: WITNESS, INVARIANTS and DIAGNOSTICS
2005-12-31 01:58:40 +00:00
pcm_unlock(d);
Marginally simplify dsp_open error handling by adding an early test to determine liklihood of opening device in requested directions. Makes for simpler error handling and change should close kern/35004. PR: kern/35004.
2002-08-28 15:19:30 +00:00
return ENODEV;
}
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
Marginally simplify dsp_open error handling by adding an early test to determine liklihood of opening device in requested directions. Makes for simpler error handling and change should close kern/35004. PR: kern/35004.
2002-08-28 15:19:30 +00:00
if (flags & O_NONBLOCK)
rdch->flags |= CHN_F_NBIO;
- Correctly increment the channel refcount in dsp_open() such that it is no longer possible to unload the driver module while sound is playing (which resulted in a panic). - Fix a similar problem with the sndstat device that I found while looking at the above. - Append a newline character to error messages in pcm_unregister() Reviewed by: cg MFC after: 10 days
2001-09-14 20:26:03 +00:00
pcm_chnref(rdch, 1);
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
CHN_UNLOCK(rdch);
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
rdref = 1;
/*
* Record channel created, ref'ed and unlocked
*/
Fix LOR #174. Tested with: WITNESS, INVARIANTS and DIAGNOSTICS
2005-12-31 01:58:40 +00:00
pcm_lock(d);
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
Marginally simplify dsp_open error handling by adding an early test to determine liklihood of opening device in requested directions. Makes for simpler error handling and change should close kern/35004. PR: kern/35004.
2002-08-28 15:19:30 +00:00
if (flags & FWRITE) {
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
/* open for write */
Release lock for a while during chn_reset() / pcm_chnalloc() operation while malloc()ing, this fixes LOR 129. See - http://lists.freebsd.org/pipermail/freebsd-current/2005-June/051157.html - http://lists.freebsd.org/pipermail/freebsd-current/2005-August/054620.html - http://sources.zabbadoz.net/freebsd/lor.html#129 Submitted by: Ariff Abdullah <skywizard@MyBSD.org.my>
2005-09-10 17:51:38 +00:00
pcm_unlock(d);
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
wrch = pcm_chnalloc(d, PCMDIR_PLAY, td->td_proc->p_pid, -1);
error = 0;
if (!wrch)
error = EBUSY; /* XXX Right return code? */
Restore old compatibility of feeding directly into /dev/dsp at 8000 hz while preserving the New World Order. Discussed with: [1] Michael W. Oliver <michael at gargantuan.com> MFC after: 1 week [1] http://lists.freebsd.org/pipermail/freebsd-multimedia/2006-January/003562.html
2006-01-16 11:07:15 +00:00
else if (chn_reset(wrch, fmt) ||
(fmt && chn_setspeed(wrch, DSP_DEFAULT_SPEED)))
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
error = ENODEV;
if (error != 0) {
if (wrch) {
/*
* Free play channel
*/
pcm_chnrelease(wrch);
Fix LOR #174. Tested with: WITNESS, INVARIANTS and DIAGNOSTICS
2005-12-31 01:58:40 +00:00
pcm_lock(d);
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
i_dev->si_drv2 = NULL;
Fix LOR #174. Tested with: WITNESS, INVARIANTS and DIAGNOSTICS
2005-12-31 01:58:40 +00:00
pcm_unlock(d);
Marginally simplify dsp_open error handling by adding an early test to determine liklihood of opening device in requested directions. Makes for simpler error handling and change should close kern/35004. PR: kern/35004.
2002-08-28 15:19:30 +00:00
}
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
if (rdref) {
/*
* Lock, deref and release previously created record channel
*/
CHN_LOCK(rdch);
pcm_chnref(rdch, -1);
pcm_chnrelease(rdch);
Fix LOR #174. Tested with: WITNESS, INVARIANTS and DIAGNOSTICS
2005-12-31 01:58:40 +00:00
pcm_lock(d);
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
i_dev->si_drv1 = NULL;
Fix LOR #174. Tested with: WITNESS, INVARIANTS and DIAGNOSTICS
2005-12-31 01:58:40 +00:00
pcm_unlock(d);
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
}
return error;
}
if (flags & O_NONBLOCK)
wrch->flags |= CHN_F_NBIO;
pcm_chnref(wrch, 1);
CHN_UNLOCK(wrch);
Fix LOR #174. Tested with: WITNESS, INVARIANTS and DIAGNOSTICS
2005-12-31 01:58:40 +00:00
pcm_lock(d);
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
i_dev->si_drv1 = rdch;
i_dev->si_drv2 = wrch;
Fix lock order reversal when opening device and chn_reset fails. Submitted by: Jan-Espen Pettersen <sigsegv@leakingmemory.org> Tested by: Georg Funk <georgfunk@web.de>
2003-05-01 16:31:21 +00:00
pcm_unlock(d);
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
return 0;
}
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
static int
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
dsp_close(struct cdev *i_dev, int flags, int mode, struct thread *td)
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
{
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
struct pcm_channel *rdch, *wrch;
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
struct snddev_info *d;
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
int refs;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
d = dsp_get_info(i_dev);
Fix LOR #174. Tested with: WITNESS, INVARIANTS and DIAGNOSTICS
2005-12-31 01:58:40 +00:00
pcm_lock(d);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
rdch = i_dev->si_drv1;
wrch = i_dev->si_drv2;
Fix LOR #174. Tested with: WITNESS, INVARIANTS and DIAGNOSTICS
2005-12-31 01:58:40 +00:00
pcm_unlock(d);
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
refs = 0;
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
if (rdch) {
CHN_LOCK(rdch);
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
refs += pcm_chnref(rdch, -1);
CHN_UNLOCK(rdch);
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
}
if (wrch) {
CHN_LOCK(wrch);
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
refs += pcm_chnref(wrch, -1);
CHN_UNLOCK(wrch);
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
}
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
/*
* If there are no more references, release the channels.
*/
if ((rdch || wrch) && refs == 0) {
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
- Fix the locking in dsp.c to prevent a LOR (AFAIK not on the LOR page). - Remove an assertion in sound.c, it's not needed (and causes a panic now). From the conversation via mail between glebius and Ariff: ---snip--- > Well, but which mutex protects now? Do we own anything else > in pcm_chnalloc()? I see some queue(4) macros in pcm_chnalloc(), > they should be protected, shouldn't they? Queue insertion/removal occur during 1) driver loading (which is pretty much single thread / sequential) or unloading (mutex protected, bail out if there is any channel with refcount > 0 or busy). 2) vchan_create()/destroy(), (which is *sigh* quite complicated), but somehow protected by 'master'/parent channel mutex. Other thread cannot add/remove vchan (or even continue traversing that queue) unless it can acquire parent channel mutex. ---snip--- Fix the locking in dsp.c to prevent a LOR (AFAIK not on the LOR page). Submitted by: Ariff Abdullah <skywizard@MyBSD.org.my> Tested with: INVARIANTS[1] and DIAGNOSTICS[2] Tested by: netchild [1,2], David Reid <david@jetnet.co.uk> [1]
2005-09-12 18:33:33 +00:00
pcm_lock(d);
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
if (pcm_getfakechan(d))
pcm_getfakechan(d)->flags = 0;
beginnings of virtual playback channel support instead of using two malloced arrays for storing channel lists, use an slist. convert the sndstat device to use sbufs and optionally provide more detail about channel state. vchans are software mixed playback channels. they are not enabled by this commit. they use the feeder infrastructure to emulate normal playback channels in a manner transparent to applications, whilst providing as many channels are desired, especially suitable for devices with only one hardware playback channel. in the future they will provide additional features. those wishing to test this functionality will need to add vchan.c to sys/conf/files and use 'sysctl -w hw.snd.pcm0.vchans' to enable it. blocksize and auto-rate selection are not yet supported.
2001-05-27 17:22:00 +00:00
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
i_dev->si_drv1 = NULL;
i_dev->si_drv2 = NULL;
Back out last commit. Needs slightly more subtle handling.
2002-08-20 08:02:56 +00:00
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
dsp_set_flags(i_dev, dsp_get_flags(i_dev) & ~SD_F_TRANSIENT);
beginnings of virtual playback channel support instead of using two malloced arrays for storing channel lists, use an slist. convert the sndstat device to use sbufs and optionally provide more detail about channel state. vchans are software mixed playback channels. they are not enabled by this commit. they use the feeder infrastructure to emulate normal playback channels in a manner transparent to applications, whilst providing as many channels are desired, especially suitable for devices with only one hardware playback channel. in the future they will provide additional features. those wishing to test this functionality will need to add vchan.c to sys/conf/files and use 'sysctl -w hw.snd.pcm0.vchans' to enable it. blocksize and auto-rate selection are not yet supported.
2001-05-27 17:22:00 +00:00
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
pcm_unlock(d);
beginnings of virtual playback channel support instead of using two malloced arrays for storing channel lists, use an slist. convert the sndstat device to use sbufs and optionally provide more detail about channel state. vchans are software mixed playback channels. they are not enabled by this commit. they use the feeder infrastructure to emulate normal playback channels in a manner transparent to applications, whilst providing as many channels are desired, especially suitable for devices with only one hardware playback channel. in the future they will provide additional features. those wishing to test this functionality will need to add vchan.c to sys/conf/files and use 'sysctl -w hw.snd.pcm0.vchans' to enable it. blocksize and auto-rate selection are not yet supported.
2001-05-27 17:22:00 +00:00
Fix sound LOR problems: dsp_open: rearrange to only hold one lock at a time dsp_close: ditto mixer_hwvol_init: delete locking, the only consumer seems to be the ess driver and it only call it a creation time, I think the device will be stable across the sleepable malloc. cmi interrupt routine: Release locks while caller chn_intr, either this or do what emu10k1 does which is have no locks at in the interrupt handler. Submitted by: mat@cnd.mcgill.ca
2003-11-11 05:38:28 +00:00
if (rdch) {
CHN_LOCK(rdch);
chn_abort(rdch); /* won't sleep */
rdch->flags &= ~(CHN_F_RUNNING | CHN_F_MAPPED | CHN_F_DEAD);
chn_reset(rdch, 0);
pcm_chnrelease(rdch);
}
if (wrch) {
CHN_LOCK(wrch);
/*
* XXX: Maybe the right behaviour is to abort on non_block.
* It seems that mplayer flushes the audio queue by quickly
* closing and re-opening. In FBSD, there's a long pause
* while the audio queue flushes that I presume isn't there in
* linux.
*/
chn_flush(wrch); /* may sleep */
wrch->flags &= ~(CHN_F_RUNNING | CHN_F_MAPPED | CHN_F_DEAD);
chn_reset(wrch, 0);
pcm_chnrelease(wrch);
}
General spl* cleanup. It doesn't serve any purpose anymore. Submitted by: Ariff Abdullah <skywizard@MyBSD.org.my>
2005-10-02 15:39:07 +00:00
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
return 0;
}
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
static int
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
dsp_read(struct cdev *i_dev, struct uio *buf, int flag)
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
{
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
struct pcm_channel *rdch, *wrch;
int ret;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
getchns(i_dev, &rdch, &wrch, SD_F_PRIO_RD);
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
make dsp_read() assert that it has a read channel, not a write channel Submitted by: green
2000-01-10 07:05:15 +00:00
KASSERT(rdch, ("dsp_read: nonexistant channel"));
KASSERT(rdch->flags & CHN_F_BUSY, ("dsp_read: nonbusy channel"));
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
if (rdch->flags & (CHN_F_MAPPED | CHN_F_DEAD)) {
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
relchns(i_dev, rdch, wrch, SD_F_PRIO_RD);
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
return EINVAL;
}
more mmap fixes
2000-04-17 17:06:47 +00:00
if (!(rdch->flags & CHN_F_RUNNING))
* add a non-reset device- will not reset the channel on open. you will have to mknod yourself for now. * don't eat the first write() * partial rvplayer fix- don't panic on unaligned writes unless our feeder chain requires them for downconversion. a fuller fix is on the way.
1999-09-28 21:43:35 +00:00
rdch->flags |= CHN_F_RUNNING;
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
ret = chn_read(rdch, buf);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
relchns(i_dev, rdch, wrch, SD_F_PRIO_RD);
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
return ret;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
static int
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
dsp_write(struct cdev *i_dev, struct uio *buf, int flag)
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
{
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
struct pcm_channel *rdch, *wrch;
int ret;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
getchns(i_dev, &rdch, &wrch, SD_F_PRIO_WR);
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
fix dma underrun issues mutate some panics to kasserts add more spl protection PR: kern/14990 Partially Submitted by: Vladimir N.Silyaev <vns@delta.odessa.ua> Reviewed by: dfr
1999-12-05 19:09:13 +00:00
KASSERT(wrch, ("dsp_write: nonexistant channel"));
KASSERT(wrch->flags & CHN_F_BUSY, ("dsp_write: nonbusy channel"));
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
if (wrch->flags & (CHN_F_MAPPED | CHN_F_DEAD)) {
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
relchns(i_dev, rdch, wrch, SD_F_PRIO_WR);
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
return EINVAL;
}
more mmap fixes
2000-04-17 17:06:47 +00:00
if (!(wrch->flags & CHN_F_RUNNING))
* add a non-reset device- will not reset the channel on open. you will have to mknod yourself for now. * don't eat the first write() * partial rvplayer fix- don't panic on unaligned writes unless our feeder chain requires them for downconversion. a fuller fix is on the way.
1999-09-28 21:43:35 +00:00
wrch->flags |= CHN_F_RUNNING;
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
ret = chn_write(wrch, buf);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
relchns(i_dev, rdch, wrch, SD_F_PRIO_WR);
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
return ret;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
static int
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
dsp_ioctl(struct cdev *i_dev, u_long cmd, caddr_t arg, int mode, struct thread *td)
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
{
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
struct pcm_channel *chn, *rdch, *wrch;
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
struct snddev_info *d;
int kill;
int ret = 0, *arg_i = (int *)arg, tmp;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
/*
* this is an evil hack to allow broken apps to perform mixer ioctls
* on dsp devices.
*/
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
As previously announced: discontinue use of makedev() call in soundcode. This takes us a lot closer to refcounting dev_t. This patch originally by cg@ with a few minor changes by me. It is largely untested, but has been HEADSUP'ed twice, so presumably people have not found any issues with it. Submitted by: cg@
2004-01-17 10:37:11 +00:00
d = dsp_get_info(i_dev);
if (IOCGROUP(cmd) == 'M')
return mixer_ioctl(d->mixer_dev, cmd, arg, mode, td);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
getchns(i_dev, &rdch, &wrch, 0);
kill = 0;
fix a bug where opening for write would not fail if channel allocation failed when playing, if we stall for 1s with no data advancing, abort and mark the channel dead - fail all future operations
2000-06-20 23:27:12 +00:00
if (wrch && (wrch->flags & CHN_F_DEAD))
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
kill |= 1;
if (rdch && (rdch->flags & CHN_F_DEAD))
kill |= 2;
if (kill == 3) {
relchns(i_dev, rdch, wrch, 0);
fix a bug where opening for write would not fail if channel allocation failed when playing, if we stall for 1s with no data advancing, abort and mark the channel dead - fail all future operations
2000-06-20 23:27:12 +00:00
return EINVAL;
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
}
if (kill & 1)
wrch = NULL;
if (kill & 2)
rdch = NULL;
Change KASSERT() in feed_vchan16() into an explicit test and call to panic() so that the buffer overflow just beyond this point is always caught, even when the code is not compiled with INVARIANTS. Change chn_setblocksize() buffer reallocation code to attempt to avoid the feed_vchan16() buffer overflow by attempting to always keep the bufsoft buffer at least as large as the bufhard buffer. Print a diagnositic message Danger! %s bufsoft size increasing from %d to %d after CHANNEL_SETBLOCKSIZE() if our best attempts fail. If feed_vchan16() were to be called by the interrupt handler while locks are dropped in chn_setblocksize() to increase the size bufsoft to match the size of bufhard, the panic() code in feed_vchan16() will be triggered. If the diagnostic message is printed, it is a warning that a panic is possible if the system were to see events in an "unlucky" order. Change the locking code to avoid the need for MTX_RECURSIVE mutexes. Add the MTX_DUPOK option to the channel mutexes and change the locking sequence to always lock the parent channel before its children to avoid the possibility of deadlock. Actually implement locking assertions for the channel mutexes and fix the problems found by the resulting assertion violations. Clean up the locking code in dsp_ioctl(). Allocate the channel buffers using the malloc() M_WAITOK option instead of M_NOWAIT so that buffer allocation won't fail. Drop locks across the malloc() calls. Add/modify KASSERTS() in attempt to detect problems early. Abuse layering by adding a pointer to the snd_dbuf structure that points back to the pcm_channel that owns it. This allows sndbuf_resize() to do proper locking without having to change the its API, which is used by the hardware drivers. Don't dereference a NULL pointer when setting hw.snd.maxautovchans if a hardware driver is not loaded. Noticed by Ryan Sommers <ryans at gamersimpact.com>. Tested by: Stefan Ehmann <shoesoft AT gmx.net> Tested by: matk (Mathew Kanner) Tested by: Gordon Bergling <gbergling AT 0xfce3.net>
2004-01-28 08:02:15 +00:00
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
switch(cmd) {
make mmap sort-of work. there seem to be interactions with certain hw drivers, so still work in progress. do various mmap-related ioctls right. improve blocksize control. bits of cleanup.
2000-04-15 05:04:12 +00:00
#ifdef OLDPCM_IOCTL
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
/*
* we start with the new ioctl interface.
*/
case AIONWRITE: /* how many bytes can write ? */
Fix some potential NULL pointer dereferences. This is supposed to fix some Coverity Prevent errors (Ariff didn't looked at the CID's (ENOTIME), I just told him that there are some problems in function dsp_ioctl()). CID: 215-218 Found with: Coverity Prevent(tm) Submitted by: ariff MFC after: 5 days
2006-01-29 16:48:41 +00:00
if (wrch) {
CHN_LOCK(wrch);
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
/*
if (wrch && wrch->bufhard.dl)
while (chn_wrfeed(wrch) == 0);
*/
Fix some potential NULL pointer dereferences. This is supposed to fix some Coverity Prevent errors (Ariff didn't looked at the CID's (ENOTIME), I just told him that there are some problems in function dsp_ioctl()). CID: 215-218 Found with: Coverity Prevent(tm) Submitted by: ariff MFC after: 5 days
2006-01-29 16:48:41 +00:00
*arg_i = sndbuf_getfree(wrch->bufsoft);
CHN_UNLOCK(wrch);
} else {
*arg_i = 0;
ret = EINVAL;
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
break;
case AIOSSIZE: /* set the current blocksize */
{
struct snd_size *p = (struct snd_size *)arg;
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
p->play_size = 0;
p->rec_size = 0;
if (wrch) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(wrch);
fix the return value of the SOUND_PCM_WRITE_CHANNELS ioctl re-enable old pcm ioctls
2000-05-27 14:40:17 +00:00
chn_setblocksize(wrch, 2, p->play_size);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
p->play_size = sndbuf_getblksz(wrch->bufsoft);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(wrch);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
}
if (rdch) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(rdch);
fix the return value of the SOUND_PCM_WRITE_CHANNELS ioctl re-enable old pcm ioctls
2000-05-27 14:40:17 +00:00
chn_setblocksize(rdch, 2, p->rec_size);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
p->rec_size = sndbuf_getblksz(rdch->bufsoft);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(rdch);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
break;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
case AIOGSIZE: /* get the current blocksize */
{
struct snd_size *p = (struct snd_size *)arg;
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
if (wrch) {
CHN_LOCK(wrch);
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
p->play_size = sndbuf_getblksz(wrch->bufsoft);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(wrch);
}
if (rdch) {
CHN_LOCK(rdch);
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
p->rec_size = sndbuf_getblksz(rdch->bufsoft);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(rdch);
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
break;
case AIOSFMT:
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
case AIOGFMT:
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
{
snd_chan_param *p = (snd_chan_param *)arg;
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
Add some safeguards to AIOSFMT: - Return EINVAL if play_format or rec_format is set but the corresponding sample rate is 0. - Don't try to set the playback or recording format to 0. Previously, issuing an AIOSFMT ioctl with an all-zeroes snd_chan_param would trigger a KASSERT in chn_fmtchain(); I'm unsure about the effects on a kernel without INVARIANTS. After this commit, issuing AIOSFMT with an all-zeroes snd_chan_param is equivalent to issuing AIOGFMT. MFC after: 2 weeks
2005-10-30 10:03:11 +00:00
if (cmd == AIOSFMT &&
((p->play_format != 0 && p->play_rate == 0) ||
(p->rec_format != 0 && p->rec_rate == 0))) {
ret = EINVAL;
break;
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
if (wrch) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(wrch);
Add some safeguards to AIOSFMT: - Return EINVAL if play_format or rec_format is set but the corresponding sample rate is 0. - Don't try to set the playback or recording format to 0. Previously, issuing an AIOSFMT ioctl with an all-zeroes snd_chan_param would trigger a KASSERT in chn_fmtchain(); I'm unsure about the effects on a kernel without INVARIANTS. After this commit, issuing AIOSFMT with an all-zeroes snd_chan_param is equivalent to issuing AIOGFMT. MFC after: 2 weeks
2005-10-30 10:03:11 +00:00
if (cmd == AIOSFMT && p->play_format != 0) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
chn_setformat(wrch, p->play_format);
chn_setspeed(wrch, p->play_rate);
}
p->play_rate = wrch->speed;
p->play_format = wrch->format;
CHN_UNLOCK(wrch);
} else {
p->play_rate = 0;
p->play_format = 0;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
if (rdch) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(rdch);
Add some safeguards to AIOSFMT: - Return EINVAL if play_format or rec_format is set but the corresponding sample rate is 0. - Don't try to set the playback or recording format to 0. Previously, issuing an AIOSFMT ioctl with an all-zeroes snd_chan_param would trigger a KASSERT in chn_fmtchain(); I'm unsure about the effects on a kernel without INVARIANTS. After this commit, issuing AIOSFMT with an all-zeroes snd_chan_param is equivalent to issuing AIOGFMT. MFC after: 2 weeks
2005-10-30 10:03:11 +00:00
if (cmd == AIOSFMT && p->rec_format != 0) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
chn_setformat(rdch, p->rec_format);
chn_setspeed(rdch, p->rec_rate);
}
p->rec_rate = rdch->speed;
p->rec_format = rdch->format;
CHN_UNLOCK(rdch);
} else {
p->rec_rate = 0;
p->rec_format = 0;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
}
break;
case AIOGCAP: /* get capabilities */
{
snd_capabilities *p = (snd_capabilities *)arg;
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
struct pcmchan_caps *pcaps = NULL, *rcaps = NULL;
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
struct cdev *pdev;
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
if (rdch) {
CHN_LOCK(rdch);
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
rcaps = chn_getcaps(rdch);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
}
if (wrch) {
CHN_LOCK(wrch);
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
pcaps = chn_getcaps(wrch);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
p->rate_min = max(rcaps? rcaps->minspeed : 0,
pcaps? pcaps->minspeed : 0);
p->rate_max = min(rcaps? rcaps->maxspeed : 1000000,
pcaps? pcaps->maxspeed : 1000000);
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
p->bufsize = min(rdch? sndbuf_getsize(rdch->bufsoft) : 1000000,
wrch? sndbuf_getsize(wrch->bufsoft) : 1000000);
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
/* XXX bad on sb16 */
rework feeder sytem to allow feeders in klds modify driver capability reporting format to list every audio format seperately- required for above and because we could not previously indicate that mono was unsupported. there should be no functional impact.
2000-08-20 22:18:56 +00:00
p->formats = (rdch? chn_getformats(rdch) : 0xffffffff) &
(wrch? chn_getformats(wrch) : 0xffffffff);
handle duplex properly in the AIOGCAP ioctl - this may (partially?) fix rat
2000-06-07 11:51:30 +00:00
if (rdch && wrch)
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
p->formats |= (dsp_get_flags(i_dev) & SD_F_SIMPLEX)? 0 : AFMT_FULLDUPLEX;
As previously announced: discontinue use of makedev() call in soundcode. This takes us a lot closer to refcounting dev_t. This patch originally by cg@ with a few minor changes by me. It is largely untested, but has been HEADSUP'ed twice, so presumably people have not found any issues with it. Submitted by: cg@
2004-01-17 10:37:11 +00:00
pdev = d->mixer_dev;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
p->mixers = 1; /* default: one mixer */
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
p->inputs = pdev->si_drv1? mix_getdevs(pdev->si_drv1) : 0;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
p->left = p->right = 100;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
if (rdch)
CHN_UNLOCK(rdch);
if (wrch)
CHN_UNLOCK(wrch);
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
break;
case AIOSTOP:
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
if (*arg_i == AIOSYNC_PLAY && wrch) {
CHN_LOCK(wrch);
update code dealing with snd_dbuf objects to do so using a functional interface modify chn_setblocksize() to pick a default soft-blocksize appropriate to the sample rate and format in use. it will aim for a power of two size small enough to generate block sizes of at most 20ms. it will also set the hard-blocksize taking into account rate/format conversions in use. update drivers to implement setblocksize correctly: updated, tested: sb16, emu10k1, maestro, solo updated, untested: ad1816, ess, mss, sb8, csa not updated: ds1, es137x, fm801, neomagic, t4dwave, via82c686 i lack hardware to test: ad1816, csa, fm801, neomagic others will be updated/tested in the next few days.
2000-12-23 03:16:13 +00:00
*arg_i = chn_abort(wrch);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(wrch);
} else if (*arg_i == AIOSYNC_CAPTURE && rdch) {
CHN_LOCK(rdch);
update code dealing with snd_dbuf objects to do so using a functional interface modify chn_setblocksize() to pick a default soft-blocksize appropriate to the sample rate and format in use. it will aim for a power of two size small enough to generate block sizes of at most 20ms. it will also set the hard-blocksize taking into account rate/format conversions in use. update drivers to implement setblocksize correctly: updated, tested: sb16, emu10k1, maestro, solo updated, untested: ad1816, ess, mss, sb8, csa not updated: ds1, es137x, fm801, neomagic, t4dwave, via82c686 i lack hardware to test: ad1816, csa, fm801, neomagic others will be updated/tested in the next few days.
2000-12-23 03:16:13 +00:00
*arg_i = chn_abort(rdch);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(rdch);
} else {
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
printf("AIOSTOP: bad channel 0x%x\n", *arg_i);
*arg_i = 0;
}
break;
case AIOSYNC:
printf("AIOSYNC chan 0x%03lx pos %lu unimplemented\n",
((snd_sync_parm *)arg)->chan, ((snd_sync_parm *)arg)->pos);
break;
implement settrigger according to spec fixes for non-blocking mode
2000-04-23 18:09:18 +00:00
#endif
Fix the "dsp sync" debugging printf that was unconditional. Fix a few formatting nits/glitches/whatever.
1999-09-19 09:53:57 +00:00
/*
* here follow the standard ioctls (filio.h etc.)
*/
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
case FIONREAD: /* get # bytes to read */
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
if (rdch) {
CHN_LOCK(rdch);
/* if (rdch && rdch->bufhard.dl)
while (chn_rdfeed(rdch) == 0);
*/
*arg_i = sndbuf_getready(rdch->bufsoft);
CHN_UNLOCK(rdch);
Fix some potential NULL pointer dereferences. This is supposed to fix some Coverity Prevent errors (Ariff didn't looked at the CID's (ENOTIME), I just told him that there are some problems in function dsp_ioctl()). CID: 215-218 Found with: Coverity Prevent(tm) Submitted by: ariff MFC after: 5 days
2006-01-29 16:48:41 +00:00
} else {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
*arg_i = 0;
Fix some potential NULL pointer dereferences. This is supposed to fix some Coverity Prevent errors (Ariff didn't looked at the CID's (ENOTIME), I just told him that there are some problems in function dsp_ioctl()). CID: 215-218 Found with: Coverity Prevent(tm) Submitted by: ariff MFC after: 5 days
2006-01-29 16:48:41 +00:00
ret = EINVAL;
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
break;
case FIOASYNC: /*set/clear async i/o */
DEB( printf("FIOASYNC\n") ; )
break;
implement settrigger according to spec fixes for non-blocking mode
2000-04-23 18:09:18 +00:00
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
case SNDCTL_DSP_NONBLOCK:
case FIONBIO: /* set/clear non-blocking i/o */
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
if (rdch) {
CHN_LOCK(rdch);
if (*arg_i)
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
rdch->flags |= CHN_F_NBIO;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
else
rdch->flags &= ~CHN_F_NBIO;
CHN_UNLOCK(rdch);
}
if (wrch) {
CHN_LOCK(wrch);
if (*arg_i)
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
wrch->flags |= CHN_F_NBIO;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
else
wrch->flags &= ~CHN_F_NBIO;
CHN_UNLOCK(wrch);
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
break;
/*
Fix the "dsp sync" debugging printf that was unconditional. Fix a few formatting nits/glitches/whatever.
1999-09-19 09:53:57 +00:00
* Finally, here is the linux-compatible ioctl interface
*/
#define THE_REAL_SNDCTL_DSP_GETBLKSIZE _IOWR('P', 4, int)
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
case THE_REAL_SNDCTL_DSP_GETBLKSIZE:
case SNDCTL_DSP_GETBLKSIZE:
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
chn = wrch ? wrch : rdch;
Fix some potential NULL pointer dereferences. This is supposed to fix some Coverity Prevent errors (Ariff didn't looked at the CID's (ENOTIME), I just told him that there are some problems in function dsp_ioctl()). CID: 215-218 Found with: Coverity Prevent(tm) Submitted by: ariff MFC after: 5 days
2006-01-29 16:48:41 +00:00
if (chn) {
CHN_LOCK(chn);
*arg_i = sndbuf_getblksz(chn->bufsoft);
CHN_UNLOCK(chn);
} else {
*arg_i = 0;
ret = EINVAL;
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
break ;
case SNDCTL_DSP_SETBLKSIZE:
make mmap sort-of work. there seem to be interactions with certain hw drivers, so still work in progress. do various mmap-related ioctls right. improve blocksize control. bits of cleanup.
2000-04-15 05:04:12 +00:00
RANGE(*arg_i, 16, 65536);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
if (wrch) {
CHN_LOCK(wrch);
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
chn_setblocksize(wrch, 2, *arg_i);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(wrch);
}
if (rdch) {
CHN_LOCK(rdch);
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
chn_setblocksize(rdch, 2, *arg_i);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(rdch);
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
break;
case SNDCTL_DSP_RESET:
DEB(printf("dsp reset\n"));
Reduce latency when using the SNDCTL_DSP_RESET ioctl by calling chn_resetbuf(). Submited by: Pyun YongHyeon <yongari@kt-is.co.kr> Approved by: tanimura (mentor)
2004-01-20 05:30:09 +00:00
if (wrch) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(wrch);
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
chn_abort(wrch);
Reduce latency when using the SNDCTL_DSP_RESET ioctl by calling chn_resetbuf(). Submited by: Pyun YongHyeon <yongari@kt-is.co.kr> Approved by: tanimura (mentor)
2004-01-20 05:30:09 +00:00
chn_resetbuf(wrch);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(wrch);
Reduce latency when using the SNDCTL_DSP_RESET ioctl by calling chn_resetbuf(). Submited by: Pyun YongHyeon <yongari@kt-is.co.kr> Approved by: tanimura (mentor)
2004-01-20 05:30:09 +00:00
}
if (rdch) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(rdch);
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
chn_abort(rdch);
Reduce latency when using the SNDCTL_DSP_RESET ioctl by calling chn_resetbuf(). Submited by: Pyun YongHyeon <yongari@kt-is.co.kr> Approved by: tanimura (mentor)
2004-01-20 05:30:09 +00:00
chn_resetbuf(rdch);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(rdch);
Reduce latency when using the SNDCTL_DSP_RESET ioctl by calling chn_resetbuf(). Submited by: Pyun YongHyeon <yongari@kt-is.co.kr> Approved by: tanimura (mentor)
2004-01-20 05:30:09 +00:00
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
break;
case SNDCTL_DSP_SYNC:
Fix the "dsp sync" debugging printf that was unconditional. Fix a few formatting nits/glitches/whatever.
1999-09-19 09:53:57 +00:00
DEB(printf("dsp sync\n"));
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
/* chn_sync may sleep */
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
if (wrch) {
CHN_LOCK(wrch);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
chn_sync(wrch, sndbuf_getsize(wrch->bufsoft) - 4);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(wrch);
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
break;
case SNDCTL_DSP_SPEED:
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
/* chn_setspeed may sleep */
tmp = 0;
if (wrch) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(wrch);
move various fields from pcm_channel to snd_dbuf improve chn_read code- don't stop/restart on overrun, just dump data more error checking on ioctls
2000-05-26 21:55:13 +00:00
ret = chn_setspeed(wrch, *arg_i);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
tmp = wrch->speed;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(wrch);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
}
if (rdch && ret == 0) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(rdch);
move various fields from pcm_channel to snd_dbuf improve chn_read code- don't stop/restart on overrun, just dump data more error checking on ioctls
2000-05-26 21:55:13 +00:00
ret = chn_setspeed(rdch, *arg_i);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
if (tmp == 0)
tmp = rdch->speed;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(rdch);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
}
*arg_i = tmp;
break;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
case SOUND_PCM_READ_RATE:
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
chn = wrch ? wrch : rdch;
Fix some potential NULL pointer dereferences. This is supposed to fix some Coverity Prevent errors (Ariff didn't looked at the CID's (ENOTIME), I just told him that there are some problems in function dsp_ioctl()). CID: 215-218 Found with: Coverity Prevent(tm) Submitted by: ariff MFC after: 5 days
2006-01-29 16:48:41 +00:00
if (chn) {
CHN_LOCK(chn);
*arg_i = chn->speed;
CHN_UNLOCK(chn);
} else {
*arg_i = 0;
ret = EINVAL;
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
break;
case SNDCTL_DSP_STEREO:
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
tmp = -1;
*arg_i = (*arg_i)? AFMT_STEREO : 0;
if (wrch) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(wrch);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
ret = chn_setformat(wrch, (wrch->format & ~AFMT_STEREO) | *arg_i);
tmp = (wrch->format & AFMT_STEREO)? 1 : 0;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(wrch);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
}
if (rdch && ret == 0) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(rdch);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
ret = chn_setformat(rdch, (rdch->format & ~AFMT_STEREO) | *arg_i);
if (tmp == -1)
tmp = (rdch->format & AFMT_STEREO)? 1 : 0;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(rdch);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
}
*arg_i = tmp;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
break;
case SOUND_PCM_WRITE_CHANNELS:
move various fields from pcm_channel to snd_dbuf improve chn_read code- don't stop/restart on overrun, just dump data more error checking on ioctls
2000-05-26 21:55:13 +00:00
/* case SNDCTL_DSP_CHANNELS: ( == SOUND_PCM_WRITE_CHANNELS) */
handle bogus values passed to SOUND_PCM_WRITE_CHANNELS/SNDCTL_DSP_CHANNELS ioctls better- if 0 is passed, return the current setting. if 1 or 2 are passed, behave accordingly. treat values >2 as 2. PR: kern/29964
2001-08-23 12:21:12 +00:00
if (*arg_i != 0) {
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
tmp = 0;
handle bogus values passed to SOUND_PCM_WRITE_CHANNELS/SNDCTL_DSP_CHANNELS ioctls better- if 0 is passed, return the current setting. if 1 or 2 are passed, behave accordingly. treat values >2 as 2. PR: kern/29964
2001-08-23 12:21:12 +00:00
*arg_i = (*arg_i != 1)? AFMT_STEREO : 0;
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
if (wrch) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(wrch);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
ret = chn_setformat(wrch, (wrch->format & ~AFMT_STEREO) | *arg_i);
tmp = (wrch->format & AFMT_STEREO)? 2 : 1;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(wrch);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
}
if (rdch && ret == 0) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(rdch);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
ret = chn_setformat(rdch, (rdch->format & ~AFMT_STEREO) | *arg_i);
if (tmp == 0)
tmp = (rdch->format & AFMT_STEREO)? 2 : 1;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(rdch);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
}
*arg_i = tmp;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
} else {
chn = wrch ? wrch : rdch;
CHN_LOCK(chn);
*arg_i = (chn->format & AFMT_STEREO) ? 2 : 1;
CHN_UNLOCK(chn);
}
move various fields from pcm_channel to snd_dbuf improve chn_read code- don't stop/restart on overrun, just dump data more error checking on ioctls
2000-05-26 21:55:13 +00:00
break;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
case SOUND_PCM_READ_CHANNELS:
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
chn = wrch ? wrch : rdch;
Fix some potential NULL pointer dereferences. This is supposed to fix some Coverity Prevent errors (Ariff didn't looked at the CID's (ENOTIME), I just told him that there are some problems in function dsp_ioctl()). CID: 215-218 Found with: Coverity Prevent(tm) Submitted by: ariff MFC after: 5 days
2006-01-29 16:48:41 +00:00
if (chn) {
CHN_LOCK(chn);
*arg_i = (chn->format & AFMT_STEREO) ? 2 : 1;
CHN_UNLOCK(chn);
} else {
*arg_i = 0;
ret = EINVAL;
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
break;
case SNDCTL_DSP_GETFMTS: /* returns a mask of supported fmts */
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
chn = wrch ? wrch : rdch;
Fix some potential NULL pointer dereferences. This is supposed to fix some Coverity Prevent errors (Ariff didn't looked at the CID's (ENOTIME), I just told him that there are some problems in function dsp_ioctl()). CID: 215-218 Found with: Coverity Prevent(tm) Submitted by: ariff MFC after: 5 days
2006-01-29 16:48:41 +00:00
if (chn) {
CHN_LOCK(chn);
*arg_i = chn_getformats(chn);
CHN_UNLOCK(chn);
} else {
*arg_i = 0;
ret = EINVAL;
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
break ;
case SNDCTL_DSP_SETFMT: /* sets _one_ format */
handle AFMT_QUERY right
2000-10-26 01:34:54 +00:00
if ((*arg_i != AFMT_QUERY)) {
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
tmp = 0;
if (wrch) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(wrch);
handle AFMT_QUERY right
2000-10-26 01:34:54 +00:00
ret = chn_setformat(wrch, (*arg_i) | (wrch->format & AFMT_STEREO));
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
tmp = wrch->format & ~AFMT_STEREO;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(wrch);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
}
if (rdch && ret == 0) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(rdch);
handle AFMT_QUERY right
2000-10-26 01:34:54 +00:00
ret = chn_setformat(rdch, (*arg_i) | (rdch->format & AFMT_STEREO));
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
if (tmp == 0)
tmp = rdch->format & ~AFMT_STEREO;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(rdch);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
}
*arg_i = tmp;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
} else {
chn = wrch ? wrch : rdch;
CHN_LOCK(chn);
*arg_i = chn->format & ~AFMT_STEREO;
CHN_UNLOCK(chn);
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
break;
case SNDCTL_DSP_SETFRAGMENT:
DEB(printf("SNDCTL_DSP_SETFRAGMENT 0x%08x\n", *(int *)arg));
{
make mmap sort-of work. there seem to be interactions with certain hw drivers, so still work in progress. do various mmap-related ioctls right. improve blocksize control. bits of cleanup.
2000-04-15 05:04:12 +00:00
u_int32_t fragln = (*arg_i) & 0x0000ffff;
u_int32_t maxfrags = ((*arg_i) & 0xffff0000) >> 16;
u_int32_t fragsz;
From luigi: In SNDCTL_DSP_SETFRAGMENT, if you specify both read and write channels, the existing code first acts on the read channel, but as a side effect it updates the arguments (maxfrags, fragsz) passed by the caller according to acceptable values for the read channel, and then uses the modified values to act on the write channel. The problem with this approach is that, given a (maxfrags, fragsz) user-specified value, the actual values computed by the read and write channels may differ: e.g. the read channel might want to allocate more fragments than what the user specified because it has no side-effects on the delay and it helps in case of slow readers, whereas the write channel needs to use as few fragments as possible to keep the audio latency low (very important with telephony apps). This patch stores the values computed by the read channel into temproary variables so the write channel will use the actual arguments of the ioctl. This patch is very helpful with telephony apps such as asterisk. Submitted by: luigi Approved by: netchild (mentor)
2005-11-14 18:20:47 +00:00
u_int32_t r_maxfrags, r_fragsz;
make mmap sort-of work. there seem to be interactions with certain hw drivers, so still work in progress. do various mmap-related ioctls right. improve blocksize control. bits of cleanup.
2000-04-15 05:04:12 +00:00
RANGE(fragln, 4, 16);
fragsz = 1 << fragln;
if (maxfrags == 0)
maxfrags = CHN_2NDBUFMAXSIZE / fragsz;
Set a lower bound on fragment size rather than returning a failure when the user specifies a maximum fragment size < 2. This is the behavior that Linux provides and fixes the problem I've observed in Tribes2 where sounds effects are delayed by 1/2 a second.
2003-06-26 13:13:18 +00:00
if (maxfrags < 2)
maxfrags = 2;
make mmap sort-of work. there seem to be interactions with certain hw drivers, so still work in progress. do various mmap-related ioctls right. improve blocksize control. bits of cleanup.
2000-04-15 05:04:12 +00:00
if (maxfrags * fragsz > CHN_2NDBUFMAXSIZE)
maxfrags = CHN_2NDBUFMAXSIZE / fragsz;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
implement settrigger according to spec fixes for non-blocking mode
2000-04-23 18:09:18 +00:00
DEB(printf("SNDCTL_DSP_SETFRAGMENT %d frags, %d sz\n", maxfrags, fragsz));
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
if (rdch) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(rdch);
make mmap sort-of work. there seem to be interactions with certain hw drivers, so still work in progress. do various mmap-related ioctls right. improve blocksize control. bits of cleanup.
2000-04-15 05:04:12 +00:00
ret = chn_setblocksize(rdch, maxfrags, fragsz);
From luigi: In SNDCTL_DSP_SETFRAGMENT, if you specify both read and write channels, the existing code first acts on the read channel, but as a side effect it updates the arguments (maxfrags, fragsz) passed by the caller according to acceptable values for the read channel, and then uses the modified values to act on the write channel. The problem with this approach is that, given a (maxfrags, fragsz) user-specified value, the actual values computed by the read and write channels may differ: e.g. the read channel might want to allocate more fragments than what the user specified because it has no side-effects on the delay and it helps in case of slow readers, whereas the write channel needs to use as few fragments as possible to keep the audio latency low (very important with telephony apps). This patch stores the values computed by the read channel into temproary variables so the write channel will use the actual arguments of the ioctl. This patch is very helpful with telephony apps such as asterisk. Submitted by: luigi Approved by: netchild (mentor)
2005-11-14 18:20:47 +00:00
r_maxfrags = sndbuf_getblkcnt(rdch->bufsoft);
r_fragsz = sndbuf_getblksz(rdch->bufsoft);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(rdch);
From luigi: In SNDCTL_DSP_SETFRAGMENT, if you specify both read and write channels, the existing code first acts on the read channel, but as a side effect it updates the arguments (maxfrags, fragsz) passed by the caller according to acceptable values for the read channel, and then uses the modified values to act on the write channel. The problem with this approach is that, given a (maxfrags, fragsz) user-specified value, the actual values computed by the read and write channels may differ: e.g. the read channel might want to allocate more fragments than what the user specified because it has no side-effects on the delay and it helps in case of slow readers, whereas the write channel needs to use as few fragments as possible to keep the audio latency low (very important with telephony apps). This patch stores the values computed by the read channel into temproary variables so the write channel will use the actual arguments of the ioctl. This patch is very helpful with telephony apps such as asterisk. Submitted by: luigi Approved by: netchild (mentor)
2005-11-14 18:20:47 +00:00
} else {
r_maxfrags = maxfrags;
r_fragsz = fragsz;
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
}
if (wrch && ret == 0) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(wrch);
make mmap sort-of work. there seem to be interactions with certain hw drivers, so still work in progress. do various mmap-related ioctls right. improve blocksize control. bits of cleanup.
2000-04-15 05:04:12 +00:00
ret = chn_setblocksize(wrch, maxfrags, fragsz);
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
maxfrags = sndbuf_getblkcnt(wrch->bufsoft);
fragsz = sndbuf_getblksz(wrch->bufsoft);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(wrch);
From luigi: In SNDCTL_DSP_SETFRAGMENT, if you specify both read and write channels, the existing code first acts on the read channel, but as a side effect it updates the arguments (maxfrags, fragsz) passed by the caller according to acceptable values for the read channel, and then uses the modified values to act on the write channel. The problem with this approach is that, given a (maxfrags, fragsz) user-specified value, the actual values computed by the read and write channels may differ: e.g. the read channel might want to allocate more fragments than what the user specified because it has no side-effects on the delay and it helps in case of slow readers, whereas the write channel needs to use as few fragments as possible to keep the audio latency low (very important with telephony apps). This patch stores the values computed by the read channel into temproary variables so the write channel will use the actual arguments of the ioctl. This patch is very helpful with telephony apps such as asterisk. Submitted by: luigi Approved by: netchild (mentor)
2005-11-14 18:20:47 +00:00
} else { /* use whatever came from the read channel */
maxfrags = r_maxfrags;
fragsz = r_fragsz;
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
}
make mmap sort-of work. there seem to be interactions with certain hw drivers, so still work in progress. do various mmap-related ioctls right. improve blocksize control. bits of cleanup.
2000-04-15 05:04:12 +00:00
move various fields from pcm_channel to snd_dbuf improve chn_read code- don't stop/restart on overrun, just dump data more error checking on ioctls
2000-05-26 21:55:13 +00:00
fragln = 0;
while (fragsz > 1) {
fragln++;
fragsz >>= 1;
}
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
*arg_i = (maxfrags << 16) | fragln;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
break;
fix SNDCTL_DSP_GETISPACE behaviour
2001-10-24 12:40:54 +00:00
case SNDCTL_DSP_GETISPACE:
- latest 2ndbuffer patch - make chn_setdir work for rec on isa cards - note: es1371 does not irq in smp Submitted by: tanimura
1999-12-29 03:46:54 +00:00
/* return the size of data available in the input queue */
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
{
audio_buf_info *a = (audio_buf_info *)arg;
if (rdch) {
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
struct snd_dbuf *bs = rdch->bufsoft;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(rdch);
fix SNDCTL_DSP_GETISPACE behaviour
2001-10-24 12:40:54 +00:00
a->bytes = sndbuf_getready(bs);
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
a->fragments = a->bytes / sndbuf_getblksz(bs);
a->fragstotal = sndbuf_getblkcnt(bs);
a->fragsize = sndbuf_getblksz(bs);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(rdch);
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
}
break;
case SNDCTL_DSP_GETOSPACE:
/* return space available in the output queue */
{
audio_buf_info *a = (audio_buf_info *)arg;
if (wrch) {
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
struct snd_dbuf *bs = wrch->bufsoft;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(wrch);
Disable frantic DMA update within few SNDCTL_DSP_* ioctl. This should reduce huge playback / recording latency for applications that try to act smarter and manage their own buffering (XMMS, Skype, etc.). Note to Skype + via8xxx users: Remove previous hackish "hint.pcm.<unit>.via_dxs_disabled" from kernel hint and see whether this changes cure all those annoying sound issues.
2005-12-30 07:33:28 +00:00
/* XXX abusive DMA update: chn_wrupdate(wrch); */
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
a->bytes = sndbuf_getfree(bs);
a->fragments = a->bytes / sndbuf_getblksz(bs);
a->fragstotal = sndbuf_getblkcnt(bs);
a->fragsize = sndbuf_getblksz(bs);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(wrch);
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
}
break;
case SNDCTL_DSP_GETIPTR:
{
count_info *a = (count_info *)arg;
if (rdch) {
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
struct snd_dbuf *bs = rdch->bufsoft;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(rdch);
Disable frantic DMA update within few SNDCTL_DSP_* ioctl. This should reduce huge playback / recording latency for applications that try to act smarter and manage their own buffering (XMMS, Skype, etc.). Note to Skype + via8xxx users: Remove previous hackish "hint.pcm.<unit>.via_dxs_disabled" from kernel hint and see whether this changes cure all those annoying sound issues.
2005-12-30 07:33:28 +00:00
/* XXX abusive DMA update: chn_rdupdate(rdch); */
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
a->bytes = sndbuf_gettotal(bs);
a->blocks = sndbuf_getblocks(bs) - rdch->blocks;
a->ptr = sndbuf_getreadyptr(bs);
rdch->blocks = sndbuf_getblocks(bs);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(rdch);
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
} else
ret = EINVAL;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
break;
case SNDCTL_DSP_GETOPTR:
{
count_info *a = (count_info *)arg;
if (wrch) {
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
struct snd_dbuf *bs = wrch->bufsoft;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(wrch);
Disable frantic DMA update within few SNDCTL_DSP_* ioctl. This should reduce huge playback / recording latency for applications that try to act smarter and manage their own buffering (XMMS, Skype, etc.). Note to Skype + via8xxx users: Remove previous hackish "hint.pcm.<unit>.via_dxs_disabled" from kernel hint and see whether this changes cure all those annoying sound issues.
2005-12-30 07:33:28 +00:00
/* XXX abusive DMA update: chn_wrupdate(wrch); */
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
a->bytes = sndbuf_gettotal(bs);
a->blocks = sndbuf_getblocks(bs) - wrch->blocks;
a->ptr = sndbuf_getreadyptr(bs);
wrch->blocks = sndbuf_getblocks(bs);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(wrch);
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
} else
ret = EINVAL;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
break;
case SNDCTL_DSP_GETCAPS:
*arg_i = DSP_CAP_REALTIME | DSP_CAP_MMAP | DSP_CAP_TRIGGER;
many changes: * add new channels to the end of the list so channels used in order of addition * de-globalise definition of struct snddev_info and provide accessor functions where necessary. * move the $FreeBSD$ tag in each .c file into a macro and allow the /dev/sndstat handler to display these when set to maximum verbosity to aid debugging. * allow each device to register its own sndstat handler to reduce the amount of groping sndstat must do in foreign structs.
2001-08-23 11:30:52 +00:00
if (rdch && wrch && !(dsp_get_flags(i_dev) & SD_F_SIMPLEX))
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
*arg_i |= DSP_CAP_DUPLEX;
break;
case SOUND_PCM_READ_BITS:
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
chn = wrch ? wrch : rdch;
Fix some potential NULL pointer dereferences. This is supposed to fix some Coverity Prevent errors (Ariff didn't looked at the CID's (ENOTIME), I just told him that there are some problems in function dsp_ioctl()). CID: 215-218 Found with: Coverity Prevent(tm) Submitted by: ariff MFC after: 5 days
2006-01-29 16:48:41 +00:00
if (chn) {
CHN_LOCK(chn);
if (chn->format & AFMT_8BIT)
*arg_i = 8;
else if (chn->format & AFMT_16BIT)
*arg_i = 16;
else if (chn->format & AFMT_24BIT)
*arg_i = 24;
else if (chn->format & AFMT_32BIT)
*arg_i = 32;
else
ret = EINVAL;
CHN_UNLOCK(chn);
} else {
*arg_i = 0;
Release lock for a while during chn_reset() / pcm_chnalloc() operation while malloc()ing, this fixes LOR 129. See - http://lists.freebsd.org/pipermail/freebsd-current/2005-June/051157.html - http://lists.freebsd.org/pipermail/freebsd-current/2005-August/054620.html - http://sources.zabbadoz.net/freebsd/lor.html#129 Submitted by: Ariff Abdullah <skywizard@MyBSD.org.my>
2005-09-10 17:51:38 +00:00
ret = EINVAL;
Fix some potential NULL pointer dereferences. This is supposed to fix some Coverity Prevent errors (Ariff didn't looked at the CID's (ENOTIME), I just told him that there are some problems in function dsp_ioctl()). CID: 215-218 Found with: Coverity Prevent(tm) Submitted by: ariff MFC after: 5 days
2006-01-29 16:48:41 +00:00
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
break;
case SNDCTL_DSP_SETTRIGGER:
if (rdch) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(rdch);
implement settrigger according to spec fixes for non-blocking mode
2000-04-23 18:09:18 +00:00
rdch->flags &= ~(CHN_F_TRIGGERED | CHN_F_NOTRIGGER);
MFS: make SNDCTL_DSP_SETTRIGGER work
2001-08-23 11:58:38 +00:00
if (*arg_i & PCM_ENABLE_INPUT)
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
chn_start(rdch, 1);
MFS: make SNDCTL_DSP_SETTRIGGER work
2001-08-23 11:58:38 +00:00
else
implement settrigger according to spec fixes for non-blocking mode
2000-04-23 18:09:18 +00:00
rdch->flags |= CHN_F_NOTRIGGER;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(rdch);
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
if (wrch) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(wrch);
implement settrigger according to spec fixes for non-blocking mode
2000-04-23 18:09:18 +00:00
wrch->flags &= ~(CHN_F_TRIGGERED | CHN_F_NOTRIGGER);
MFS: make SNDCTL_DSP_SETTRIGGER work
2001-08-23 11:58:38 +00:00
if (*arg_i & PCM_ENABLE_OUTPUT)
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
chn_start(wrch, 1);
MFS: make SNDCTL_DSP_SETTRIGGER work
2001-08-23 11:58:38 +00:00
else
implement settrigger according to spec fixes for non-blocking mode
2000-04-23 18:09:18 +00:00
wrch->flags |= CHN_F_NOTRIGGER;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(wrch);
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
break;
case SNDCTL_DSP_GETTRIGGER:
*arg_i = 0;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
if (wrch) {
CHN_LOCK(wrch);
if (wrch->flags & CHN_F_TRIGGERED)
*arg_i |= PCM_ENABLE_OUTPUT;
CHN_UNLOCK(wrch);
}
if (rdch) {
CHN_LOCK(rdch);
if (rdch->flags & CHN_F_TRIGGERED)
*arg_i |= PCM_ENABLE_INPUT;
CHN_UNLOCK(rdch);
}
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
break;
- latest 2ndbuffer patch - make chn_setdir work for rec on isa cards - note: es1371 does not irq in smp Submitted by: tanimura
1999-12-29 03:46:54 +00:00
case SNDCTL_DSP_GETODELAY:
if (wrch) {
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
struct snd_dbuf *b = wrch->bufhard;
struct snd_dbuf *bs = wrch->bufsoft;
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(wrch);
Disable frantic DMA update within few SNDCTL_DSP_* ioctl. This should reduce huge playback / recording latency for applications that try to act smarter and manage their own buffering (XMMS, Skype, etc.). Note to Skype + via8xxx users: Remove previous hackish "hint.pcm.<unit>.via_dxs_disabled" from kernel hint and see whether this changes cure all those annoying sound issues.
2005-12-30 07:33:28 +00:00
/* XXX abusive DMA update: chn_wrupdate(wrch); */
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
*arg_i = sndbuf_getready(b) + sndbuf_getready(bs);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(wrch);
- latest 2ndbuffer patch - make chn_setdir work for rec on isa cards - note: es1371 does not irq in smp Submitted by: tanimura
1999-12-29 03:46:54 +00:00
} else
ret = EINVAL;
break;
implement SNDCTL_DSP_POST
2000-10-27 22:20:45 +00:00
handle AFMT_QUERY right
2000-10-26 01:34:54 +00:00
case SNDCTL_DSP_POST:
if (wrch) {
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_LOCK(wrch);
handle AFMT_QUERY right
2000-10-26 01:34:54 +00:00
wrch->flags &= ~CHN_F_NOTRIGGER;
chn_start(wrch, 1);
Lock channels only as necessary in dsp_ioctl(), and only lock one channel at a time unless it is actually necessary to lock both. This avoids problems with lock order reversal and malloc() calls with a mutex held when lower level code unlocks a channel, calls malloc(), and relocks the channel. This also avoids the cost of some unnecessary locking and unlocking. Tested by: matk
2004-02-28 19:42:48 +00:00
CHN_UNLOCK(wrch);
handle AFMT_QUERY right
2000-10-26 01:34:54 +00:00
}
break;
implement SNDCTL_DSP_POST
2000-10-27 22:20:45 +00:00
Implement SNDCTL_DSP_SETDUPLEX. This may fix sound apps which want to use full duplex mode. Approved by: matk
2004-07-02 15:31:44 +00:00
case SNDCTL_DSP_SETDUPLEX:
/*
* switch to full-duplex mode if card is in half-duplex
* mode and is able to work in full-duplex mode
*/
if (rdch && wrch && (dsp_get_flags(i_dev) & SD_F_SIMPLEX))
dsp_set_flags(i_dev, dsp_get_flags(i_dev)^SD_F_SIMPLEX);
break;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
case SNDCTL_DSP_MAPINBUF:
case SNDCTL_DSP_MAPOUTBUF:
case SNDCTL_DSP_SETSYNCRO:
/* undocumented */
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
case SNDCTL_DSP_SUBDIVIDE:
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
case SOUND_PCM_WRITE_FILTER:
case SOUND_PCM_READ_FILTER:
/* dunno what these do, don't sound important */
Implement SNDCTL_DSP_SETDUPLEX. This may fix sound apps which want to use full duplex mode. Approved by: matk
2004-07-02 15:31:44 +00:00
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
default:
Fix code that had rotted behind debugging macros. Approved by: cg (in principle) MFC after: 2 weeks
2002-01-25 04:14:12 +00:00
DEB(printf("default ioctl fn 0x%08lx fail\n", cmd));
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
ret = EINVAL;
break;
}
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
relchns(i_dev, rdch, wrch, 0);
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
return ret;
}
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
static int
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
dsp_poll(struct cdev *i_dev, int events, struct thread *td)
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
{
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
struct pcm_channel *wrch = NULL, *rdch = NULL;
add defines and ifdefs so this code will compile on 4.x add spls so this code will work on 4.x
2001-06-23 17:36:51 +00:00
int ret, e;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
ret = 0;
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
getchns(i_dev, &rdch, &wrch, SD_F_PRIO_RD | SD_F_PRIO_WR);
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
if (wrch) {
e = (events & (POLLOUT | POLLWRNORM));
if (e)
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
ret |= chn_poll(wrch, e, td);
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
}
if (rdch) {
e = (events & (POLLIN | POLLRDNORM));
if (e)
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
ret |= chn_poll(rdch, e, td);
various locking fixes, rework open logic and channel registration PR: kern/28084
2001-06-14 13:31:30 +00:00
}
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
relchns(i_dev, rdch, wrch, SD_F_PRIO_RD | SD_F_PRIO_WR);
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
return ret;
}
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
static int
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
dsp_mmap(struct cdev *i_dev, vm_offset_t offset, vm_paddr_t *paddr, int nprot)
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
{
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
struct pcm_channel *wrch = NULL, *rdch = NULL, *c;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
if (nprot & PROT_EXEC)
return -1;
getchns(i_dev, &rdch, &wrch, SD_F_PRIO_RD | SD_F_PRIO_WR);
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
#if 0
/*
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
* XXX the linux api uses the nprot to select read/write buffer
* our vm system doesn't allow this, so force write buffer
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
*/
add defines and ifdefs so this code will compile on 4.x add spls so this code will work on 4.x
2001-06-23 17:36:51 +00:00
if (wrch && (nprot & PROT_WRITE)) {
move various fields from pcm_channel to snd_dbuf improve chn_read code- don't stop/restart on overrun, just dump data more error checking on ioctls
2000-05-26 21:55:13 +00:00
c = wrch;
add defines and ifdefs so this code will compile on 4.x add spls so this code will work on 4.x
2001-06-23 17:36:51 +00:00
} else if (rdch && (nprot & PROT_READ)) {
move various fields from pcm_channel to snd_dbuf improve chn_read code- don't stop/restart on overrun, just dump data more error checking on ioctls
2000-05-26 21:55:13 +00:00
c = rdch;
add defines and ifdefs so this code will compile on 4.x add spls so this code will work on 4.x
2001-06-23 17:36:51 +00:00
} else {
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
return -1;
add defines and ifdefs so this code will compile on 4.x add spls so this code will work on 4.x
2001-06-23 17:36:51 +00:00
}
mega-commit. this introduces a new buffering mechanism which results in dramatic simplification of the channel manager. as several structures have changed, we take the opportunity to move their definitions into the source files where they are used, make them private and de-typedef them. the sound drivers are updated to use snd_setup_intr instead of bus_setup_intr, and to comply with the de-typedefed structures. the ac97, mixer and channel layers have been updated with finegrained locking, as have some drivers- not all though. the rest will follow soon.
2001-03-24 23:10:29 +00:00
#else
c = wrch;
#endif
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
if (c == NULL) {
relchns(i_dev, rdch, wrch, SD_F_PRIO_RD | SD_F_PRIO_WR);
move various fields from pcm_channel to snd_dbuf improve chn_read code- don't stop/restart on overrun, just dump data more error checking on ioctls
2000-05-26 21:55:13 +00:00
return -1;
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
}
don't allow mmap beyond the end of the buffer Submitted by: Philippe Anel <philippe.anel@noos.fr> (partially)
2002-01-23 04:32:18 +00:00
if (offset >= sndbuf_getsize(c->bufsoft)) {
relchns(i_dev, rdch, wrch, SD_F_PRIO_RD | SD_F_PRIO_WR);
return -1;
}
don't set a flag that is already set
2001-07-01 18:57:16 +00:00
if (!(c->flags & CHN_F_MAPPED))
c->flags |= CHN_F_MAPPED;
Cleanup of the d_mmap_t interface. - Get rid of the useless atop() / pmap_phys_address() detour. The device mmap handlers must now give back the physical address without atop()'ing it. - Don't borrow the physical address of the mapping in the returned int. Now we properly pass a vm_offset_t * and expect it to be filled by the mmap handler when the mapping was successful. The mmap handler must now return 0 when successful, any other value is considered as an error. Previously, returning -1 was the only way to fail. This change thus accidentally fixes some devices which were bogusly returning errno constants which would have been considered as addresses by the device pager. - Garbage collect the poorly named pmap_phys_address() now that it's no longer used. - Convert all the d_mmap_t consumers to the new API. I'm still not sure wheter we need a __FreeBSD_version bump for this, since and we didn't guarantee API/ABI stability until 5.1-RELEASE. Discussed with: alc, phk, jake Reviewed by: peter Compile-tested on: LINT (i386), GENERIC (alpha and sparc64) Runtime-tested on: i386
2003-02-25 03:21:22 +00:00
*paddr = vtophys(sndbuf_getbufofs(c->bufsoft, offset));
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
relchns(i_dev, rdch, wrch, SD_F_PRIO_RD | SD_F_PRIO_WR);
Cleanup of the d_mmap_t interface. - Get rid of the useless atop() / pmap_phys_address() detour. The device mmap handlers must now give back the physical address without atop()'ing it. - Don't borrow the physical address of the mapping in the returned int. Now we properly pass a vm_offset_t * and expect it to be filled by the mmap handler when the mapping was successful. The mmap handler must now return 0 when successful, any other value is considered as an error. Previously, returning -1 was the only way to fail. This change thus accidentally fixes some devices which were bogusly returning errno constants which would have been considered as addresses by the device pager. - Garbage collect the poorly named pmap_phys_address() now that it's no longer used. - Convert all the d_mmap_t consumers to the new API. I'm still not sure wheter we need a __FreeBSD_version bump for this, since and we didn't guarantee API/ABI stability until 5.1-RELEASE. Discussed with: alc, phk, jake Reviewed by: peter Compile-tested on: LINT (i386), GENERIC (alpha and sparc64) Runtime-tested on: i386
2003-02-25 03:21:22 +00:00
return 0;
say hello to newpcm. it is not yet enabled, requiring new pnp code from dfr to compile successfully. further details will be provided in the commit enabling newpcm.
1999-09-01 04:08:39 +00:00
}
add defines and ifdefs so this code will compile on 4.x add spls so this code will work on 4.x
2001-06-23 17:36:51 +00:00
#ifdef USING_DEVFS
Fix a panic when kldloading a sound driver. Do this by replacing the link-list of dev_t's with named variables. Remove used code. Approved by: tanimura (mentor)
2004-01-20 03:58:57 +00:00
/*
* Clone logic is this:
* x E X = {dsp, dspW, audio}
* x -> x${sysctl("hw.snd.unit")}
* xN->
* for i N = 1 to channels of device N
* if xN.i isn't busy, return its dev_t
*/
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
static void
Merge the dev_clone and dev_clone_cred event handlers into a single event handler, dev_clone, which accepts a credential argument. Implementors of the event can ignore it if they're not interested, and most do. This avoids having multiple event handler types and fall-back/precedence logic in devfs. This changes the kernel API for /dev cloning, and may affect third party packages containg cloning kernel modules. Requested by: phk MFC after: 3 days
2005-08-08 19:55:32 +00:00
dsp_clone(void *arg, struct ucred *cred, char *name, int namelen,
struct cdev **dev)
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
{
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
struct cdev *pdev;
Fix a panic when kldloading a sound driver. Do this by replacing the link-list of dev_t's with named variables. Remove used code. Approved by: tanimura (mentor)
2004-01-20 03:58:57 +00:00
struct snddev_info *pcm_dev;
struct snddev_channel *pcm_chan;
int i, unit, devtype;
tidy up/simplify dsp_clone()
2001-06-27 19:51:02 +00:00
int devtypes[3] = {SND_DEV_DSP, SND_DEV_DSP16, SND_DEV_AUDIO};
char *devnames[3] = {"dsp", "dspW", "audio"};
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
Second half of the dev_t cleanup. The big lines are: NODEV -> NULL NOUDEV -> NODEV udev_t -> dev_t udev2dev() -> findcdev() Various minor adjustments including handling of userland access to kernel space struct cdev etc.
2004-06-17 17:16:53 +00:00
if (*dev != NULL)
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
return;
fix a potential panic in dsp_clone() if no pcm devices were detected
2001-06-17 20:15:29 +00:00
if (pcm_devclass == NULL)
return;
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
tidy up/simplify dsp_clone()
2001-06-27 19:51:02 +00:00
devtype = 0;
unit = -1;
for (i = 0; (i < 3) && (unit == -1); i++) {
devtype = devtypes[i];
if (strcmp(name, devnames[i]) == 0) {
unit = snd_unit;
} else {
if (dev_stdclone(name, NULL, devnames[i], &unit) != 1)
unit = -1;
}
}
use devclass_get_maxunit() correctly
2001-06-18 00:10:47 +00:00
if (unit == -1 || unit >= devclass_get_maxunit(pcm_devclass))
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
return;
Fix a panic when kldloading a sound driver. Do this by replacing the link-list of dev_t's with named variables. Remove used code. Approved by: tanimura (mentor)
2004-01-20 03:58:57 +00:00
pcm_dev = devclass_get_softc(pcm_devclass, unit);
Fix a panic in dsp_clone when trying to access a sound device that doesn't exists. I'm using my discretion and committing without mentor approval since Seigo is away. Noticed by: Maxime Henrion <mux@freebsd.org>
2004-01-25 22:46:22 +00:00
if (pcm_dev == NULL)
return;
Fix a panic when kldloading a sound driver. Do this by replacing the link-list of dev_t's with named variables. Remove used code. Approved by: tanimura (mentor)
2004-01-20 03:58:57 +00:00
SLIST_FOREACH(pcm_chan, &pcm_dev->channels, link) {
switch(devtype) {
case SND_DEV_DSP:
pdev = pcm_chan->dsp_devt;
break;
case SND_DEV_DSP16:
pdev = pcm_chan->dspW_devt;
break;
case SND_DEV_AUDIO:
pdev = pcm_chan->audio_devt;
break;
default:
panic("Unknown devtype %d", devtype);
}
Appropriate NULL pointer checking to avoid mysterious panic during device cloning. Approved by: netchild (mentor)
2005-11-03 08:19:04 +00:00
if ((pdev != NULL) && (pdev->si_drv1 == NULL) && (pdev->si_drv2 == NULL)) {
Fix a panic when kldloading a sound driver. Do this by replacing the link-list of dev_t's with named variables. Remove used code. Approved by: tanimura (mentor)
2004-01-20 03:58:57 +00:00
*dev = pdev;
Explicitly hold a reference to the cdev we have just cloned. This closes the race where the cdev was reclaimed before it ever made it back to devfs lookup.
2005-03-31 12:19:44 +00:00
dev_ref(*dev);
Fix a panic when kldloading a sound driver. Do this by replacing the link-list of dev_t's with named variables. Remove used code. Approved by: tanimura (mentor)
2004-01-20 03:58:57 +00:00
return;
revise dsp_clone() to return the first nonbusy channel instead of simply cycling channel numbers. remove unused fields from struct snddev_info.
2001-06-17 23:23:06 +00:00
}
}
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
}
static void
dsp_sysinit(void *p)
{
dsp_ehtag = EVENTHANDLER_REGISTER(dev_clone, dsp_clone, 0, 1000);
}
static void
dsp_sysuninit(void *p)
{
if (dsp_ehtag != NULL)
EVENTHANDLER_DEREGISTER(dev_clone, dsp_ehtag);
}
SYSINIT(dsp_sysinit, SI_SUB_DRIVERS, SI_ORDER_MIDDLE, dsp_sysinit, NULL);
SYSUNINIT(dsp_sysuninit, SI_SUB_DRIVERS, SI_ORDER_MIDDLE, dsp_sysuninit, NULL);
add defines and ifdefs so this code will compile on 4.x add spls so this code will work on 4.x
2001-06-23 17:36:51 +00:00
#endif
use a global devclass for all drivers - i'm not entirely sure why this worked before. mixer, dsp and sndstat are seperate devices - give them their own cdevsws instead of demuxing requests sent to a single cdevsw. use the si_drv1/si_drv2 fields in dev_t structures for holding information specific to an open instance of mixer/dsp. nuke /dev/{dsp,dspW,audio}[0-9]* links - this functionality is now provided using cloning. various locking fixes.
2001-06-16 21:25:10 +00:00
Reference in New Issue Copy Permalink