* ac97 oops in ali modem driver
@ 2005-10-03 1:01 Dave Jones
2005-10-03 14:34 ` Sasha Khapyorsky
0 siblings, 1 reply; 6+ messages in thread
From: Dave Jones @ 2005-10-03 1:01 UTC (permalink / raw)
To: alsa-devel
I just had a user file a Fedora bug with the following oops.
Unable to handle kernel NULL pointer dereference at virtual address 000001a4
printing eip:
df969005
*pde = 12620067
Oops: 0000 [#1]
Modules linked in: nfs radeon drm nfsd exportfs lockd nfs_acl autofs4 sunrpc pcmcia dm_mod vid
eo button battery ac yenta_socket rsrc_nonstatic pcmcia_core ohci_hcd shpchp i2c_ali1535 i2c_c
ore snd_ali5451 snd_ac97_codec snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq snd_seq_de
vice snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd soundcore snd_page_alloc prism54 8139cp m
ii floppy joydev ext3 jbd
CPU: 0
EIP: 0060:[<df969005>] Not tainted VLI
EFLAGS: 00010202 (2.6.13-1.1526_FC4)
EIP is at snd_ac97_valid_reg+0x5/0x135 [snd_ac97_codec]
eax: 00000000 ebx: 00000040 ecx: 00000000 edx: 00000040
esi: 00000000 edi: 00000040 ebp: 00001f40 esp: ca515edc
ds: 007b es: 007b ss: 0068
Process gstreamer-prope (pid: 10503, threadinfo=ca515000 task=da809aa0)
Stack: 00000040 df969149 de2ac800 00000001 d4fb1000 de3bf560 df9207ca 00000000
fffffff4 de3bf560 db6b9000 df92d6b2 00000000 00000001 00000246 00000000
00000000 00000000 0000025c 0000025c d4fb1000 d4fb1000 fffffff4 de3bf560
Call Trace:
[<df969149>] snd_ac97_write+0x14/0x5a [snd_ac97_codec]
[<df9207ca>] snd_ali_modem_hw_params+0x2c/0x4c [snd_ali5451]
[<df92d6b2>] snd_pcm_hw_params+0x290/0x609 [snd_pcm]
[<df92da7b>] snd_pcm_hw_params_user+0x50/0x6a [snd_pcm]
[<df9349b5>] snd_pcm_capture_ioctl1+0x2e/0x1ce [snd_pcm]
[<df934b70>] snd_pcm_capture_ioctl+0x0/0x1b [snd_pcm]
[<c01bf0e9>] do_ioctl+0x19/0x55
[<c01bf217>] vfs_ioctl+0x50/0x1aa
[<c01bf3ce>] sys_ioctl+0x5d/0x6b
[<c0104409>] sysenter_past_esp+0x56/0x79
Code: Bad EIP value.
It looks like we're somehow ending up with card->ac97 == NULL.
This kernel is based on 2.6.13.2, with no ALSA patches other than
an unrelated emu10k1 diff. From eyeballing a quick diff of
the ali driver between 2.6.13.2 and current upstream 2.6.14rc3,
it doesn't look like this has been addressed already.
Dave
-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: ac97 oops in ali modem driver 2005-10-03 1:01 ac97 oops in ali modem driver Dave Jones @ 2005-10-03 14:34 ` Sasha Khapyorsky 2005-10-04 11:05 ` Takashi Iwai 0 siblings, 1 reply; 6+ messages in thread From: Sasha Khapyorsky @ 2005-10-03 14:34 UTC (permalink / raw) To: Dave Jones; +Cc: alsa-devel On 21:01 Sun 02 Oct , Dave Jones wrote: > I just had a user file a Fedora bug with the following oops. Is there other messages from ali5451 loading? The sound works I guess? What do you see when 'ls /proc/asound/ali5451/codec97#0'? Sasha. > > Unable to handle kernel NULL pointer dereference at virtual address 000001a4 > printing eip: > df969005 > *pde = 12620067 > Oops: 0000 [#1] > Modules linked in: nfs radeon drm nfsd exportfs lockd nfs_acl autofs4 sunrpc pcmcia dm_mod vid > eo button battery ac yenta_socket rsrc_nonstatic pcmcia_core ohci_hcd shpchp i2c_ali1535 i2c_c > ore snd_ali5451 snd_ac97_codec snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq snd_seq_de > vice snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd soundcore snd_page_alloc prism54 8139cp m > ii floppy joydev ext3 jbd > CPU: 0 > EIP: 0060:[<df969005>] Not tainted VLI > EFLAGS: 00010202 (2.6.13-1.1526_FC4) > EIP is at snd_ac97_valid_reg+0x5/0x135 [snd_ac97_codec] > eax: 00000000 ebx: 00000040 ecx: 00000000 edx: 00000040 > esi: 00000000 edi: 00000040 ebp: 00001f40 esp: ca515edc > ds: 007b es: 007b ss: 0068 > Process gstreamer-prope (pid: 10503, threadinfo=ca515000 task=da809aa0) > Stack: 00000040 df969149 de2ac800 00000001 d4fb1000 de3bf560 df9207ca 00000000 > fffffff4 de3bf560 db6b9000 df92d6b2 00000000 00000001 00000246 00000000 > 00000000 00000000 0000025c 0000025c d4fb1000 d4fb1000 fffffff4 de3bf560 > Call Trace: > [<df969149>] snd_ac97_write+0x14/0x5a [snd_ac97_codec] > [<df9207ca>] snd_ali_modem_hw_params+0x2c/0x4c [snd_ali5451] > [<df92d6b2>] snd_pcm_hw_params+0x290/0x609 [snd_pcm] > [<df92da7b>] snd_pcm_hw_params_user+0x50/0x6a [snd_pcm] > [<df9349b5>] snd_pcm_capture_ioctl1+0x2e/0x1ce [snd_pcm] > [<df934b70>] snd_pcm_capture_ioctl+0x0/0x1b [snd_pcm] > [<c01bf0e9>] do_ioctl+0x19/0x55 > [<c01bf217>] vfs_ioctl+0x50/0x1aa > [<c01bf3ce>] sys_ioctl+0x5d/0x6b > [<c0104409>] sysenter_past_esp+0x56/0x79 > Code: Bad EIP value. > > > It looks like we're somehow ending up with card->ac97 == NULL. > This kernel is based on 2.6.13.2, with no ALSA patches other than > an unrelated emu10k1 diff. From eyeballing a quick diff of > the ali driver between 2.6.13.2 and current upstream 2.6.14rc3, > it doesn't look like this has been addressed already. > > Dave > > > > ------------------------------------------------------- > This SF.Net email is sponsored by: > Power Architecture Resource Center: Free content, downloads, discussions, > and more. http://solutions.newsforge.com/ibmarch.tmpl > _______________________________________________ > Alsa-devel mailing list > Alsa-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/alsa-devel ------------------------------------------------------- This SF.Net email is sponsored by: Power Architecture Resource Center: Free content, downloads, discussions, and more. http://solutions.newsforge.com/ibmarch.tmpl ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: ac97 oops in ali modem driver 2005-10-03 14:34 ` Sasha Khapyorsky @ 2005-10-04 11:05 ` Takashi Iwai 2005-10-05 0:01 ` Sasha Khapyorsky 0 siblings, 1 reply; 6+ messages in thread From: Takashi Iwai @ 2005-10-04 11:05 UTC (permalink / raw) To: Sasha Khapyorsky; +Cc: Dave Jones, alsa-devel At Mon, 3 Oct 2005 17:34:48 +0300, Sasha Khapyorsky wrote: > > On 21:01 Sun 02 Oct , Dave Jones wrote: > > I just had a user file a Fedora bug with the following oops. > > Is there other messages from ali5451 loading? The sound works I guess? > What do you see when 'ls /proc/asound/ali5451/codec97#0'? Maybe the probe of secondary codec failed? Then the patch below might fix. Takashi --- linux/sound/pci/ali5451/ali5451.c 29 Sep 2005 10:47:07 -0000 1.71 +++ linux/sound/pci/ali5451/ali5451.c 4 Oct 2005 11:03:38 -0000 @@ -1994,8 +1994,10 @@ if ((err = snd_ac97_mixer(codec->ac97_bus, &ac97, &codec->ac97[i])) < 0) { snd_printk("ali mixer %d creating error.\n", i); if(i == 0) - return err; - } + return err; + codec->num_of_codecs = 1; + break; + } } if (codec->spdif_support) { ------------------------------------------------------- This SF.Net email is sponsored by: Power Architecture Resource Center: Free content, downloads, discussions, and more. http://solutions.newsforge.com/ibmarch.tmpl ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: ac97 oops in ali modem driver 2005-10-04 11:05 ` Takashi Iwai @ 2005-10-05 0:01 ` Sasha Khapyorsky 2005-10-05 9:57 ` Takashi Iwai 0 siblings, 1 reply; 6+ messages in thread From: Sasha Khapyorsky @ 2005-10-05 0:01 UTC (permalink / raw) To: Takashi Iwai; +Cc: Dave Jones, alsa-devel On 13:05 Tue 04 Oct , Takashi Iwai wrote: > At Mon, 3 Oct 2005 17:34:48 +0300, > Sasha Khapyorsky wrote: > > > > On 21:01 Sun 02 Oct , Dave Jones wrote: > > > I just had a user file a Fedora bug with the following oops. > > > > Is there other messages from ali5451 loading? The sound works I guess? > > What do you see when 'ls /proc/asound/ali5451/codec97#0'? > > Maybe the probe of secondary codec failed? > Then the patch below might fix. Quite possible and we expect to see something about it in dmesg. The patch is obviously correct and will help in such cases. Frankly I would prefer like this: --- alsa-kernel/pci/ali5451/ali5451.c 4 Oct 2005 12:50:44 -0000 1.73 +++ alsa-kernel/pci/ali5451/ali5451.c 4 Oct 2005 23:06:43 -0000 @@ -1844,7 +1844,7 @@ static int __devinit snd_ali_build_pcms( { int i, err; for(i = 0 ; i < codec->num_of_codecs && i < ARRAY_SIZE(ali_pcms) ; i++) - if((err = snd_ali_pcm(codec, i, &ali_pcms[i])) < 0) + if(codec->ac97[i] && (err = snd_ali_pcm(codec, i, &ali_pcms[i])) < 0) return err; return 0; } So it will be "more" on controller's side. But not critical. Thanks, Sasha. > > > --- linux/sound/pci/ali5451/ali5451.c 29 Sep 2005 10:47:07 -0000 1.71 > +++ linux/sound/pci/ali5451/ali5451.c 4 Oct 2005 11:03:38 -0000 > @@ -1994,8 +1994,10 @@ > if ((err = snd_ac97_mixer(codec->ac97_bus, &ac97, &codec->ac97[i])) < 0) { > snd_printk("ali mixer %d creating error.\n", i); > if(i == 0) > - return err; > - } > + return err; > + codec->num_of_codecs = 1; > + break; > + } > } > > if (codec->spdif_support) { > ------------------------------------------------------- This SF.Net email is sponsored by: Power Architecture Resource Center: Free content, downloads, discussions, and more. http://solutions.newsforge.com/ibmarch.tmpl ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: ac97 oops in ali modem driver 2005-10-05 0:01 ` Sasha Khapyorsky @ 2005-10-05 9:57 ` Takashi Iwai 2005-10-06 10:00 ` Sasha Khapyorsky 0 siblings, 1 reply; 6+ messages in thread From: Takashi Iwai @ 2005-10-05 9:57 UTC (permalink / raw) To: Sasha Khapyorsky; +Cc: Dave Jones, alsa-devel At Wed, 5 Oct 2005 03:01:32 +0300, Sasha Khapyorsky wrote: > > On 13:05 Tue 04 Oct , Takashi Iwai wrote: > > At Mon, 3 Oct 2005 17:34:48 +0300, > > Sasha Khapyorsky wrote: > > > > > > On 21:01 Sun 02 Oct , Dave Jones wrote: > > > > I just had a user file a Fedora bug with the following oops. > > > > > > Is there other messages from ali5451 loading? The sound works I guess? > > > What do you see when 'ls /proc/asound/ali5451/codec97#0'? > > > > Maybe the probe of secondary codec failed? > > Then the patch below might fix. > > Quite possible and we expect to see something about it in dmesg. > The patch is obviously correct and will help in such cases. Frankly I > would prefer like this: > > > --- alsa-kernel/pci/ali5451/ali5451.c 4 Oct 2005 12:50:44 -0000 1.73 > +++ alsa-kernel/pci/ali5451/ali5451.c 4 Oct 2005 23:06:43 -0000 > @@ -1844,7 +1844,7 @@ static int __devinit snd_ali_build_pcms( > { > int i, err; > for(i = 0 ; i < codec->num_of_codecs && i < ARRAY_SIZE(ali_pcms) ; i++) > - if((err = snd_ali_pcm(codec, i, &ali_pcms[i])) < 0) > + if(codec->ac97[i] && (err = snd_ali_pcm(codec, i, &ali_pcms[i])) < 0) > return err; > return 0; > } > > > So it will be "more" on controller's side. But not critical. The NULL ac97[] and pcm[] are referred in suspend/resume code, too. It should be fixed in your case. Or correct codec->num_of_codecs as in my patch. Anyway, I already committed my patch to ALSA CVS tree. A clean up patch is welcome, of course :) Takashi ------------------------------------------------------- This SF.Net email is sponsored by: Power Architecture Resource Center: Free content, downloads, discussions, and more. http://solutions.newsforge.com/ibmarch.tmpl ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: ac97 oops in ali modem driver 2005-10-05 9:57 ` Takashi Iwai @ 2005-10-06 10:00 ` Sasha Khapyorsky 0 siblings, 0 replies; 6+ messages in thread From: Sasha Khapyorsky @ 2005-10-06 10:00 UTC (permalink / raw) To: Takashi Iwai; +Cc: Dave Jones, alsa-devel On 11:57 Wed 05 Oct , Takashi Iwai wrote: > > The NULL ac97[] and pcm[] are referred in suspend/resume code, too. Sure, and there are already pcm[i] and ac97[i] != NULL checkings. > It should be fixed in your case. Or correct codec->num_of_codecs as > in my patch. > > Anyway, I already committed my patch to ALSA CVS tree. Yeah, I see :). Let's run your patch - my note was really cosmetic. Sasha. > A clean up patch is welcome, of course :) > > > Takashi > ------------------------------------------------------- This SF.Net email is sponsored by: Power Architecture Resource Center: Free content, downloads, discussions, and more. http://solutions.newsforge.com/ibmarch.tmpl ^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2005-10-06 9:44 UTC | newest] Thread overview: 6+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2005-10-03 1:01 ac97 oops in ali modem driver Dave Jones 2005-10-03 14:34 ` Sasha Khapyorsky 2005-10-04 11:05 ` Takashi Iwai 2005-10-05 0:01 ` Sasha Khapyorsky 2005-10-05 9:57 ` Takashi Iwai 2005-10-06 10:00 ` Sasha Khapyorsky
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.