Re: [alsa-devel] [BUG] WARNING: at fs/sysfs/dir.c:455 sysfs_add_one+0x83/0x95
On Wed, Mar 09, 2011 at 10:39:49AM -0500, Steven Rostedt wrote:
Not sure who to send this to, but last night on v2.6.38-rc8 ktest triggered this warning:
[ 32.062538] YM3812 and OPL-3 driver Copyright (C) by Hannu Savolainen, Rob Hooft 1993-1996 [ 32.071305] PM: Adding info for No Bus:timer [ 32.075864] PM: Adding info for platform:snd_dummy.0^M [ 32.081232] PM: Adding info for No Bus:card0 [ 32.085853] PM: Adding info for No Bus:pcmC0D0p [ 32.090650] PM: Adding info for No Bus:pcmC0D0c [ 32.095667] PM: Adding info for No Bus:dsp [ 32.099918] ------------[ cut here ]------------ [ 32.104531] WARNING: at /home/rostedt/work/autotest/nobackup/linux-test.git/fs/sysfs/dir.c:455 sysfs_add_one+0x83/0x95() [ 32.115379] Hardware name: [ 32.118859] sysfs: cannot create duplicate filename '/dev/char/14:4'
Two modules wanted to create the same minor number, that's not good.
14:4 is the OSS /dev/audio node.
[ 32.125197] Modules linked in: [ 32.128255] Pid: 1, comm: swapper Not tainted 2.6.38-rc8-test #1 [ 32.134247] Call Trace:^M [ 32.136693] [<c044d095>] ? warn_slowpath_common+0x7c/0x91 [ 32.142168] [<c0552d35>] ? sysfs_add_one+0x83/0x95 [ 32.147036] [<c0552d35>] ? sysfs_add_one+0x83/0x95 [ 32.151904] [<c044d135>] ? warn_slowpath_fmt+0x33/0x35 [ 32.157120] [<c0552d35>] ? sysfs_add_one+0x83/0x95 [ 32.161989] [<c055393d>] ? sysfs_do_create_link+0xdf/0x181^M [ 32.167550] [<c073637f>] ? vsnprintf+0x22c/0x266 [ 32.172246] [<c0553a0f>] ? sysfs_create_link+0x17/0x1c [ 32.177462] [<c0950ce4>] ? device_add+0x183/0x573 [ 32.182245] [<c0958474>] ? device_pm_init+0x5d/0x61 [ 32.187200] [<c09510eb>] ? device_register+0x17/0x1a [ 32.192243] [<c0951165>] ? device_create_vargs+0x77/0x98 [ 32.197631] [<c09511b3>] ? device_create+0x2d/0x2f [ 32.202501] [<c0dc5caa>] ? sound_insert_unit+0x1ba/0x1d5 [ 32.207891] [<c0dc5e8e>] ? register_sound_special_device+0x103/0x11c [ 32.214319] [<c0dd7875>] ? snd_register_oss_device+0x10a/0x169 [ 32.220229] [<c0de6cd6>] ? register_oss_dsp+0x62/0xb5
Yup, sound subsystem...
[ 32.225358] [<c046b95c>] ? sched_clock_cpu+0x16a/0x17a [ 32.230575] [<c04778f2>] ? mark_held_locks+0x43/0x5b [ 32.235616] [<c0fa124c>] ? __mutex_lock_common+0x39f/0x3c7 [ 32.241177] [<c04776d2>] ? mark_lock+0x1e/0x1fb [ 32.245787] [<c04778f2>] ? mark_held_locks+0x43/0x5b [ 32.250838] [<c0fa0cff>] ? __mutex_unlock_slowpath+0x137/0x153 [ 32.256745] [<c0de6d69>] ? snd_pcm_oss_register_minor+0x40/0xe3 [ 32.262737] [<c0fa0d28>] ? mutex_unlock+0xd/0xf [ 32.267345] [<c0dd96b7>] ? snd_timer_dev_register+0x10d/0x117 [ 32.273165] [<c0dd6db9>] ? snd_device_register+0x67/0xd2 [ 32.278551] [<c07363d5>] ? sprintf+0x1c/0x1e^M [ 32.282900] [<c0de2db7>] ? snd_pcm_timer_init+0xba/0xef [ 32.288201] [<c0ddb426>] ? snd_pcm_dev_register+0x1e8/0x224 [ 32.293848] [<c0951000>] ? device_add+0x49f/0x573 [ 32.298630] [<c0dd6d32>] ? snd_device_register_all+0x5e/0x7e [ 32.304363] [<c09511b3>] ? device_create+0x2d/0x2f [ 32.309230] [<c0dd32d3>] ? snd_card_register+0x92/0x1af [ 32.314531] [<c0f87b9b>] ? snd_dummy_probe+0x322/0x3b2 [ 32.319745] [<c0953f57>] ? platform_drv_probe+0x14/0x16 [ 32.325046] [<c0953154>] ? driver_probe_device+0x116/0x1ed [ 32.330606] [<c073b414>] ? do_raw_spin_unlock+0x8f/0x94 [ 32.335906] [<c09532c5>] ? __device_attach+0x2f/0x33^M [ 32.340946] [<c09524ab>] ? bus_for_each_drv+0x44/0x6e [ 32.346073] [<c095334e>] ? device_attach+0x61/0x7e^M [ 32.350941] [<c0953296>] ? __device_attach+0x0/0x33 [ 32.355895] [<c09522e8>] ? bus_probe_device+0x23/0x38 [ 32.361023] [<c0950f61>] ? device_add+0x400/0x573 [ 32.365805] [<c0730ac4>] ? kobject_set_name_vargs+0x46/0x4c [ 32.371453] [<c0954624>] ? platform_device_add+0x139/0x182 [ 32.377015] [<c095487d>] ? platform_device_register_resndata+0x4f/0x65 [ 32.383616] [<c17e2210>] ? alsa_card_dummy_init+0xcc/0x121 [ 32.389177] [<c0403096>] ? do_one_initcall+0x8c/0x161 [ 32.394305] [<c17e2144>] ? alsa_card_dummy_init+0x0/0x121 [ 32.399780] [<c17843cf>] ? kernel_init+0x12c/0x1c2 [ 32.404646] [<c17842a3>] ? kernel_init+0x0/0x1c2 [ 32.409340] [<c040a8c2>] ? kernel_thread_helper+0x6/0x10 [ 32.414782] ---[ end trace 5d9e7ab86ca08fa1 ]--- [ 32.419772] PM: Adding info for No Bus:controlC0 [ 32.424867] PM: Adding info for isa:es18xx.0 [ 32.429167] PM: Removing info for isa:es18xx.0
This was for a 386 build. I could give a .config and full dmesg if needed.
I think the sound developers (now on CC: would probably like to see that.)
thanks,
greg k-h
At Wed, 9 Mar 2011 07:53:09 -0800, Greg KH wrote:
On Wed, Mar 09, 2011 at 10:39:49AM -0500, Steven Rostedt wrote:
Not sure who to send this to, but last night on v2.6.38-rc8 ktest triggered this warning:
[ 32.062538] YM3812 and OPL-3 driver Copyright (C) by Hannu Savolainen, Rob Hooft 1993-1996 [ 32.071305] PM: Adding info for No Bus:timer [ 32.075864] PM: Adding info for platform:snd_dummy.0^M [ 32.081232] PM: Adding info for No Bus:card0 [ 32.085853] PM: Adding info for No Bus:pcmC0D0p [ 32.090650] PM: Adding info for No Bus:pcmC0D0c [ 32.095667] PM: Adding info for No Bus:dsp [ 32.099918] ------------[ cut here ]------------ [ 32.104531] WARNING: at /home/rostedt/work/autotest/nobackup/linux-test.git/fs/sysfs/dir.c:455 sysfs_add_one+0x83/0x95() [ 32.115379] Hardware name: [ 32.118859] sysfs: cannot create duplicate filename '/dev/char/14:4'
Two modules wanted to create the same minor number, that's not good.
14:4 is the OSS /dev/audio node.
Are they modules or built-in?
[ 32.125197] Modules linked in: [ 32.128255] Pid: 1, comm: swapper Not tainted 2.6.38-rc8-test #1 [ 32.134247] Call Trace:^M [ 32.136693] [<c044d095>] ? warn_slowpath_common+0x7c/0x91 [ 32.142168] [<c0552d35>] ? sysfs_add_one+0x83/0x95 [ 32.147036] [<c0552d35>] ? sysfs_add_one+0x83/0x95 [ 32.151904] [<c044d135>] ? warn_slowpath_fmt+0x33/0x35 [ 32.157120] [<c0552d35>] ? sysfs_add_one+0x83/0x95 [ 32.161989] [<c055393d>] ? sysfs_do_create_link+0xdf/0x181^M [ 32.167550] [<c073637f>] ? vsnprintf+0x22c/0x266 [ 32.172246] [<c0553a0f>] ? sysfs_create_link+0x17/0x1c [ 32.177462] [<c0950ce4>] ? device_add+0x183/0x573 [ 32.182245] [<c0958474>] ? device_pm_init+0x5d/0x61 [ 32.187200] [<c09510eb>] ? device_register+0x17/0x1a [ 32.192243] [<c0951165>] ? device_create_vargs+0x77/0x98 [ 32.197631] [<c09511b3>] ? device_create+0x2d/0x2f [ 32.202501] [<c0dc5caa>] ? sound_insert_unit+0x1ba/0x1d5 [ 32.207891] [<c0dc5e8e>] ? register_sound_special_device+0x103/0x11c [ 32.214319] [<c0dd7875>] ? snd_register_oss_device+0x10a/0x169 [ 32.220229] [<c0de6cd6>] ? register_oss_dsp+0x62/0xb5
Yup, sound subsystem...
Yep ;)
This was for a 386 build. I could give a .config and full dmesg if needed.
I think the sound developers (now on CC: would probably like to see that.)
I guess it's a long-standing issue. The dead old OSS driver creates /dev/audio and /dev/dspW by itself without coping with sound_core module.
Takashi
On Wed, 2011-03-09 at 07:53 -0800, Greg KH wrote:
This was for a 386 build. I could give a .config and full dmesg if needed.
I think the sound developers (now on CC: would probably like to see that.)
http://rostedt.homelinux.com/private/mitest-boot-randconfig-fail-20110209042...
http://rostedt.homelinux.com/private/mitest-boot-randconfig-fail-20110209042...
-- Steve
At Wed, 09 Mar 2011 11:20:58 -0500, Steven Rostedt wrote:
On Wed, 2011-03-09 at 07:53 -0800, Greg KH wrote:
This was for a 386 build. I could give a .config and full dmesg if needed.
I think the sound developers (now on CC: would probably like to see that.)
http://rostedt.homelinux.com/private/mitest-boot-randconfig-fail-20110209042...
http://rostedt.homelinux.com/private/mitest-boot-randconfig-fail-20110209042...
Could you try the patch below? Totally untested.
Takashi
--- From: Takashi Iwai tiwai@suse.de Subject: [PATCH] sound: Use sound_register_*() for additional OSS minor devices
Since OSS driver creates the device entries for /dev/audio* and /dev/dspW* by itself without coping with sound_core, it leads to conflicts with others and let sysfs spewing warnings.
This patch rewrites the registration part of OSS driver to use the standard method also for additional minor devices.
Signed-off-by: Takashi Iwai tiwai@suse.de --- sound/oss/soundcard.c | 56 +++++++++++++++++------------------------------- sound/sound_core.c | 3 ++ 2 files changed, 23 insertions(+), 36 deletions(-)
diff --git a/sound/oss/soundcard.c b/sound/oss/soundcard.c index fcb14a09..7c7793a 100644 --- a/sound/oss/soundcard.c +++ b/sound/oss/soundcard.c @@ -526,31 +526,21 @@ bad: }
-/* These device names follow the official Linux device list, - * Documentation/devices.txt. Let us know if there are other - * common names we should support for compatibility. - * Only those devices not created by the generic code in sound_core.c are - * registered here. - */ -static const struct { - unsigned short minor; - char *name; - umode_t mode; - int *num; -} dev_list[] = { /* list of minor devices */ -/* seems to be some confusion here -- this device is not in the device list */ - {SND_DEV_DSP16, "dspW", S_IWUGO | S_IRUSR | S_IRGRP, - &num_audiodevs}, - {SND_DEV_AUDIO, "audio", S_IWUGO | S_IRUSR | S_IRGRP, - &num_audiodevs}, -}; - static int dmabuf; static int dmabug;
module_param(dmabuf, int, 0444); module_param(dmabug, int, 0444);
+/* additional minors for compatibility */ +struct oss_minor_dev { + unsigned short minor; + unsigned int enabled; +} dev_list[] = { + { SND_DEV_DSP16 }, + { SND_DEV_AUDIO }, +}; + static int __init oss_init(void) { int err; @@ -571,18 +561,12 @@ static int __init oss_init(void) sound_dmap_flag = (dmabuf > 0 ? 1 : 0);
for (i = 0; i < ARRAY_SIZE(dev_list); i++) { - device_create(sound_class, NULL, - MKDEV(SOUND_MAJOR, dev_list[i].minor), NULL, - "%s", dev_list[i].name); - - if (!dev_list[i].num) - continue; - - for (j = 1; j < *dev_list[i].num; j++) - device_create(sound_class, NULL, - MKDEV(SOUND_MAJOR, - dev_list[i].minor + (j*0x10)), - NULL, "%s%d", dev_list[i].name, j); + j = 0; + do { + unsigned short minor = dev_list[i].minor + j * 0x10; + if (!register_sound_special(&oss_sound_fops, minor)) + dev_list[i].enabled = (1 << j); + } while (++j < num_audiodevs); }
if (sound_nblocks >= MAX_MEM_BLOCKS - 1) @@ -596,11 +580,11 @@ static void __exit oss_cleanup(void) int i, j;
for (i = 0; i < ARRAY_SIZE(dev_list); i++) { - device_destroy(sound_class, MKDEV(SOUND_MAJOR, dev_list[i].minor)); - if (!dev_list[i].num) - continue; - for (j = 1; j < *dev_list[i].num; j++) - device_destroy(sound_class, MKDEV(SOUND_MAJOR, dev_list[i].minor + (j*0x10))); + j = 0; + do { + if (dev_list[i].enabled & (1 << j)) + unregister_sound_special(dev_list[i].minor); + } while (++j < num_audiodevs); } unregister_sound_special(1); diff --git a/sound/sound_core.c b/sound/sound_core.c index 5580ace..6ce2778 100644 --- a/sound/sound_core.c +++ b/sound/sound_core.c @@ -384,6 +384,9 @@ int register_sound_special_device(const struct file_operations *fops, int unit, case 4: name = "audio"; break; + case 5: + name = "dspW"; + break; case 8: name = "sequencer2"; if (unit >= SOUND_STEP)
On Wed, 2011-03-09 at 17:38 +0100, Takashi Iwai wrote:
At Wed, 09 Mar 2011 11:20:58 -0500, Steven Rostedt wrote:
On Wed, 2011-03-09 at 07:53 -0800, Greg KH wrote:
This was for a 386 build. I could give a .config and full dmesg if needed.
I think the sound developers (now on CC: would probably like to see that.)
http://rostedt.homelinux.com/private/mitest-boot-randconfig-fail-20110209042...
http://rostedt.homelinux.com/private/mitest-boot-randconfig-fail-20110209042...
Could you try the patch below? Totally untested.
Will do, but first I'll make sure that this is reproducible on every boot.
Thanks,
-- Steve
On Wed, 2011-03-09 at 17:38 +0100, Takashi Iwai wrote:
Could you try the patch below? Totally untested.
It seems to work. I added the patch and the warning is gone, and removed it, and it reappeared. Now, I don't have and speakers attached to this test box, so I never tested if sound works ;)
Takashi
From: Takashi Iwai tiwai@suse.de Subject: [PATCH] sound: Use sound_register_*() for additional OSS minor devices
Since OSS driver creates the device entries for /dev/audio* and /dev/dspW* by itself without coping with sound_core, it leads to conflicts with others and let sysfs spewing warnings.
This patch rewrites the registration part of OSS driver to use the standard method also for additional minor devices.
Reported-by: Steven Rostedt rostedt@goodmis.org (with ktest.pl) Tested-by: Steven Rostedt rostedt@goodmis.org (with ktest.pl)
-- Steve
Signed-off-by: Takashi Iwai tiwai@suse.de
At Wed, 09 Mar 2011 12:21:53 -0500, Steven Rostedt wrote:
On Wed, 2011-03-09 at 17:38 +0100, Takashi Iwai wrote:
Could you try the patch below? Totally untested.
It seems to work. I added the patch and the warning is gone, and removed it, and it reappeared. Now, I don't have and speakers attached to this test box, so I never tested if sound works ;)
Takashi
From: Takashi Iwai tiwai@suse.de Subject: [PATCH] sound: Use sound_register_*() for additional OSS minor devices
Since OSS driver creates the device entries for /dev/audio* and /dev/dspW* by itself without coping with sound_core, it leads to conflicts with others and let sysfs spewing warnings.
This patch rewrites the registration part of OSS driver to use the standard method also for additional minor devices.
Reported-by: Steven Rostedt rostedt@goodmis.org (with ktest.pl) Tested-by: Steven Rostedt rostedt@goodmis.org (with ktest.pl)
Thanks for a quick check! I merged the patch now.
Takashi
participants (3)
-
Greg KH -
Steven Rostedt -
Takashi Iwai