[alsa-devel] [PATCH v2] ALSA: ASoC: soc-compress.c: fix NULL dereference
In soc_new_compress() when rtd->dai_link->daynmic is set, we create the pcm substreams with this call:
ret = snd_pcm_new_internal(rtd->card->snd_card, new_name, num, 1, 0, &be_pcm);
which passes 0 as capture_count leading to
be_pcm->streams[SNDRV_PCM_STREAM_CAPTURE].substream
being NULL, hence when trying to set rtd a few lines below we get an oops.
Fix by using rtd->dai_link->dpcm_playback and rtd->dai_link->dpcm_capture as playback_count and capture_count to snd_pcm_new_internal().
Signed-off-by: Qais Yousef qais.yousef@imgtec.com Cc: Vinod Koul vinod.koul@intel.com Cc: Liam Girdwood lgirdwood@gmail.com Cc: Mark Brown broonie@kernel.org Cc: Jaroslav Kysela perex@perex.cz Cc: Takashi Iwai tiwai@suse.de Cc: linux-kernel@vger.kernel.org --- v2->v1: - use better way to fix it than just removing the line that caused the oops
sound/soc/soc-compress.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/sound/soc/soc-compress.c b/sound/soc/soc-compress.c index 590a82f01d0b..27a668463ad7 100644 --- a/sound/soc/soc-compress.c +++ b/sound/soc/soc-compress.c @@ -659,7 +659,8 @@ int soc_new_compress(struct snd_soc_pcm_runtime *rtd, int num) rtd->dai_link->stream_name);
ret = snd_pcm_new_internal(rtd->card->snd_card, new_name, num, - 1, 0, &be_pcm); + rtd->dai_link->dpcm_playback, + rtd->dai_link->dpcm_capture, &be_pcm); if (ret < 0) { dev_err(rtd->card->dev, "ASoC: can't create compressed for %s\n", rtd->dai_link->name); @@ -668,8 +669,10 @@ int soc_new_compress(struct snd_soc_pcm_runtime *rtd, int num)
rtd->pcm = be_pcm; rtd->fe_compr = 1; - be_pcm->streams[SNDRV_PCM_STREAM_PLAYBACK].substream->private_data = rtd; - be_pcm->streams[SNDRV_PCM_STREAM_CAPTURE].substream->private_data = rtd; + if (rtd->dai_link->dpcm_playback) + be_pcm->streams[SNDRV_PCM_STREAM_PLAYBACK].substream->private_data = rtd; + if (rtd->dai_link->dpcm_capture) + be_pcm->streams[SNDRV_PCM_STREAM_CAPTURE].substream->private_data = rtd; memcpy(compr->ops, &soc_compr_dyn_ops, sizeof(soc_compr_dyn_ops)); } else memcpy(compr->ops, &soc_compr_ops, sizeof(soc_compr_ops));
On Tue, Jan 13, 2015 at 11:18:53AM +0000, Qais Yousef wrote:
In soc_new_compress() when rtd->dai_link->daynmic is set, we create the pcm
^^^^^^^^ typo
substreams with this call:
ret = snd_pcm_new_internal(rtd->card->snd_card, new_name, num, 1, 0, &be_pcm);
which passes 0 as capture_count leading to
be_pcm->streams[SNDRV_PCM_STREAM_CAPTURE].substream
being NULL, hence when trying to set rtd a few lines below we get an oops.
It is a good practice to add the oops here
Fix by using rtd->dai_link->dpcm_playback and rtd->dai_link->dpcm_capture as playback_count and capture_count to snd_pcm_new_internal().
Signed-off-by: Qais Yousef qais.yousef@imgtec.com Cc: Vinod Koul vinod.koul@intel.com Cc: Liam Girdwood lgirdwood@gmail.com Cc: Mark Brown broonie@kernel.org Cc: Jaroslav Kysela perex@perex.cz Cc: Takashi Iwai tiwai@suse.de Cc: linux-kernel@vger.kernel.org
v2->v1:
- use better way to fix it than just removing the line that caused the oops
sound/soc/soc-compress.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/sound/soc/soc-compress.c b/sound/soc/soc-compress.c index 590a82f01d0b..27a668463ad7 100644 --- a/sound/soc/soc-compress.c +++ b/sound/soc/soc-compress.c @@ -659,7 +659,8 @@ int soc_new_compress(struct snd_soc_pcm_runtime *rtd, int num) rtd->dai_link->stream_name);
ret = snd_pcm_new_internal(rtd->card->snd_card, new_name, num,
1, 0, &be_pcm);
rtd->dai_link->dpcm_playback,rtd->dai_link->dpcm_capture, &be_pcm);@@ -668,8 +669,10 @@ int soc_new_compress(struct snd_soc_pcm_runtime *rtd, int num)
rtd->pcm = be_pcm; rtd->fe_compr = 1;
be_pcm->streams[SNDRV_PCM_STREAM_PLAYBACK].substream->private_data = rtd;be_pcm->streams[SNDRV_PCM_STREAM_CAPTURE].substream->private_data = rtd;
if (rtd->dai_link->dpcm_playback)be_pcm->streams[SNDRV_PCM_STREAM_PLAYBACK].substream->private_data = rtd;if (rtd->dai_link->dpcm_capture)
this should be else if, as for compressed device we can have playback or capture not both
be_pcm->streams[SNDRV_PCM_STREAM_CAPTURE].substream->private_data = rtd;-- 2.1.0
On 01/13/2015 02:59 PM, Vinod Koul wrote:
On Tue, Jan 13, 2015 at 11:18:53AM +0000, Qais Yousef wrote:
In soc_new_compress() when rtd->dai_link->daynmic is set, we create the pcm
^^^^^^^^typo
substreams with this call:
ret = snd_pcm_new_internal(rtd->card->snd_card, new_name, num, 1, 0, &be_pcm);which passes 0 as capture_count leading to
be_pcm->streams[SNDRV_PCM_STREAM_CAPTURE].substreambeing NULL, hence when trying to set rtd a few lines below we get an oops.
It is a good practice to add the oops here
Will this really be helpful? I think it'll be more clutter (the backtrace on metag arch is not great):
Oops: err 8007 (Unknown fault) addr 00000008 [#1] Modules linked in: CPU: 0 PID: 6 Comm: kworker/u2:0 Not tainted 3.18.0-rc4+ #1904 Workqueue: deferwq _deferred_probe_work_func task: 4f030780 ti: 4f044000 task.ti: 4f044000 pt_regs @ 4f044388 SaveMask = 0x4041 Flags = 0x0008 (Znoc) TXRPT = 0x00000000 PC = 0x402e6c58 A0StP = 0x4f044388 A1GbP = 0x60001000 A0FrP = 0x4f044110 A1LbP = 0x40000048 A0.2 = 0x00000000 A1.2 = 0x00000000 A0.3 = 0x40090000 A1.3 = 0x00000001 D0Re0 = 0x00000000 D1Re0 = 0x00000001 D0Ar6 = 0x00000000 D1Ar5 = 0x4b5c1a00 D0Ar4 = 0x4f044330 D1Ar3 = 0x405833a8 D0Ar2 = 0x4f1d7170 D1Ar1 = 0x4b5c25a0 D0FrT = 0x00000001 D1RtP = 0x402e6c20 D0.5 = 0x00000000 D1.5 = 0x4f1f65c4 D0.6 = 0x4f1f65c4 D1.6 = 0x4f1d0500 D0.7 = 0x00000001 D1.7 = 0x4f1e3e40
Call trace: [<40410004>] _ieee80211_change_bss+0x1b4/0x220 [<400f8034>] _kernfs_add_one+0x10c/0x17c [<400fa2b0>] ___kernfs_create_file+0x94/0xdc [<402d8bf0>] _snd_soc_register_card+0x12b8/0x1380 [<400170e8>] ___request_region+0x58/0x150 [<402068b4>] _devres_add+0x14/0x2c [<402e97f0>] _zero1xx_probe+0x2b8/0x37c [<40205004>] _platform_drv_probe+0x4c/0xc0 [<40204fb4>] _platform_drv_remove+0x3c/0x40 [<402032e0>] _driver_probe_device+0xc8/0x294 [<40204fb4>] _platform_drv_remove+0x3c/0x40 [<40203624>] _wait_for_device_probe+0x7c/0x80 [<40201cac>] _bus_for_each_drv+0x5c/0xb0 [<40203708>] _device_attach+0x84/0x9c [<40202258>] _bus_probe_device+0x90/0xd0 [<4020354c>] _deferred_probe_work_func+0x70/0xac [<40025d84>] _process_one_work+0x110/0x364 [<402034d8>] _device_bind_driver+0x2c/0x30 [<40046150>] _mod_timer+0xc4/0x178 [<400286a4>] _worker_thread+0x14c/0x4d4 [<4002b90c>] _kthread_parkme+0x14/0x18 [<40028554>] _pool_mayday_timeout+0xe8/0xec [<4002ba08>] _kthread+0xf8/0x100 [<4000aeb4>] _ret_from_fork+0x44/0x110 [<4002b90c>] _kthread_parkme+0x14/0x18 [<40000044>] _text+0x44/0x48 [<40000044>] _text+0x44/0x48 [<4002b90c>] _kthread_parkme+0x14/0x18
Process: kworker/u2:0 (pid: 6, stack limit = 4f046000) ---[ end trace fabdbb359f5c60d8 ]---
Fix by using rtd->dai_link->dpcm_playback and rtd->dai_link->dpcm_capture as playback_count and capture_count to snd_pcm_new_internal().
Signed-off-by: Qais Yousef qais.yousef@imgtec.com Cc: Vinod Koul vinod.koul@intel.com Cc: Liam Girdwood lgirdwood@gmail.com Cc: Mark Brown broonie@kernel.org Cc: Jaroslav Kysela perex@perex.cz Cc: Takashi Iwai tiwai@suse.de Cc: linux-kernel@vger.kernel.org
v2->v1: - use better way to fix it than just removing the line that caused the oops
sound/soc/soc-compress.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/sound/soc/soc-compress.c b/sound/soc/soc-compress.c index 590a82f01d0b..27a668463ad7 100644 --- a/sound/soc/soc-compress.c +++ b/sound/soc/soc-compress.c @@ -659,7 +659,8 @@ int soc_new_compress(struct snd_soc_pcm_runtime *rtd, int num) rtd->dai_link->stream_name);
ret = snd_pcm_new_internal(rtd->card->snd_card, new_name, num,
1, 0, &be_pcm);
rtd->dai_link->dpcm_playback,rtd->dai_link->dpcm_capture, &be_pcm);@@ -668,8 +669,10 @@ int soc_new_compress(struct snd_soc_pcm_runtime *rtd, int num)
rtd->pcm = be_pcm; rtd->fe_compr = 1;
be_pcm->streams[SNDRV_PCM_STREAM_PLAYBACK].substream->private_data = rtd;be_pcm->streams[SNDRV_PCM_STREAM_CAPTURE].substream->private_data = rtd;
if (rtd->dai_link->dpcm_playback)be_pcm->streams[SNDRV_PCM_STREAM_PLAYBACK].substream->private_data = rtd;if (rtd->dai_link->dpcm_capture)this should be else if, as for compressed device we can have playback or capture not both
be_pcm->streams[SNDRV_PCM_STREAM_CAPTURE].substream->private_data = rtd;-- 2.1.0
On Tue, Jan 13, 2015 at 03:16:10PM +0000, Qais Yousef wrote:
On 01/13/2015 02:59 PM, Vinod Koul wrote:
being NULL, hence when trying to set rtd a few lines below we get an oops.
It is a good practice to add the oops here
Will this really be helpful? I think it'll be more clutter (the backtrace on metag arch is not great):
It's better in general to leave it out unless it's adding something (for example sometimes the particular call path is important) and even there edit it down to relevant details - the splat from the full oops normally overwhelms the commit message.
On 13 January 2015 15:16:10 GMT+00:00, Qais Yousef qais.yousef@imgtec.com wrote:
On 01/13/2015 02:59 PM, Vinod Koul wrote:
On Tue, Jan 13, 2015 at 11:18:53AM +0000, Qais Yousef wrote:
In soc_new_compress() when rtd->dai_link->daynmic is set, we create
the pcm
^^^^^^^^typo
substreams with this call:
ret = snd_pcm_new_internal(rtd->card->snd_card, new_name, num, 1, 0, &be_pcm);which passes 0 as capture_count leading to
be_pcm->streams[SNDRV_PCM_STREAM_CAPTURE].substreambeing NULL, hence when trying to set rtd a few lines below we get an
oops.
It is a good practice to add the oops here
Will this really be helpful? I think it'll be more clutter (the backtrace on metag arch is not great):
I suspect you don't have frame pointers enabled in your kernel config. That should improve the meaningfulness of the backtrace.
Oops: err 8007 (Unknown fault) addr 00000008 [#1] Modules linked in: CPU: 0 PID: 6 Comm: kworker/u2:0 Not tainted 3.18.0-rc4+ #1904 Workqueue: deferwq _deferred_probe_work_func task: 4f030780 ti: 4f044000 task.ti: 4f044000 pt_regs @ 4f044388 SaveMask = 0x4041 Flags = 0x0008 (Znoc) TXRPT = 0x00000000 PC = 0x402e6c58 A0StP = 0x4f044388 A1GbP = 0x60001000 A0FrP = 0x4f044110 A1LbP = 0x40000048 A0.2 = 0x00000000 A1.2 = 0x00000000 A0.3 = 0x40090000 A1.3 = 0x00000001 D0Re0 = 0x00000000 D1Re0 = 0x00000001 D0Ar6 = 0x00000000 D1Ar5 = 0x4b5c1a00 D0Ar4 = 0x4f044330 D1Ar3 = 0x405833a8 D0Ar2 = 0x4f1d7170 D1Ar1 = 0x4b5c25a0 D0FrT = 0x00000001 D1RtP = 0x402e6c20 D0.5 = 0x00000000 D1.5 = 0x4f1f65c4 D0.6 = 0x4f1f65c4 D1.6 = 0x4f1d0500 D0.7 = 0x00000001 D1.7 = 0x4f1e3e40
Call trace: [<40410004>] _ieee80211_change_bss+0x1b4/0x220 [<400f8034>] _kernfs_add_one+0x10c/0x17c [<400fa2b0>] ___kernfs_create_file+0x94/0xdc [<402d8bf0>] _snd_soc_register_card+0x12b8/0x1380 [<400170e8>] ___request_region+0x58/0x150 [<402068b4>] _devres_add+0x14/0x2c [<402e97f0>] _zero1xx_probe+0x2b8/0x37c [<40205004>] _platform_drv_probe+0x4c/0xc0 [<40204fb4>] _platform_drv_remove+0x3c/0x40 [<402032e0>] _driver_probe_device+0xc8/0x294 [<40204fb4>] _platform_drv_remove+0x3c/0x40 [<40203624>] _wait_for_device_probe+0x7c/0x80 [<40201cac>] _bus_for_each_drv+0x5c/0xb0 [<40203708>] _device_attach+0x84/0x9c [<40202258>] _bus_probe_device+0x90/0xd0 [<4020354c>] _deferred_probe_work_func+0x70/0xac [<40025d84>] _process_one_work+0x110/0x364 [<402034d8>] _device_bind_driver+0x2c/0x30 [<40046150>] _mod_timer+0xc4/0x178 [<400286a4>] _worker_thread+0x14c/0x4d4 [<4002b90c>] _kthread_parkme+0x14/0x18 [<40028554>] _pool_mayday_timeout+0xe8/0xec [<4002ba08>] _kthread+0xf8/0x100 [<4000aeb4>] _ret_from_fork+0x44/0x110 [<4002b90c>] _kthread_parkme+0x14/0x18 [<40000044>] _text+0x44/0x48 [<40000044>] _text+0x44/0x48 [<4002b90c>] _kthread_parkme+0x14/0x18
Process: kworker/u2:0 (pid: 6, stack limit = 4f046000) ---[ end trace fabdbb359f5c60d8 ]---
-- James Hogan
On 01/13/2015 04:20 PM, Mark Brown wrote:
On Tue, Jan 13, 2015 at 03:16:10PM +0000, Qais Yousef wrote:
On 01/13/2015 02:59 PM, Vinod Koul wrote:
being NULL, hence when trying to set rtd a few lines below we get an oops.
It is a good practice to add the oops here
Will this really be helpful? I think it'll be more clutter (the backtrace on metag arch is not great):
It's better in general to leave it out unless it's adding something (for example sometimes the particular call path is important) and even there edit it down to relevant details - the splat from the full oops normally overwhelms the commit message.
I think the commit message explains what's going. So unless Vinod insists I'll send v3 with the other 2 requested fixes.
Thanks for the review!
On 01/13/2015 05:21 PM, James Hogan wrote:
On 13 January 2015 15:16:10 GMT+00:00, Qais Yousef qais.yousef@imgtec.com wrote:
Will this really be helpful? I think it'll be more clutter (the backtrace on metag arch is not great):
I suspect you don't have frame pointers enabled in your kernel config. That should improve the meaningfulness of the backtrace.
I forgot about this option. Thanks.
participants (4)
-
James Hogan -
Mark Brown -
Qais Yousef -
Vinod Koul