| Message ID | s5hwqcy5pit.wl%tiwai@suse.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 06/03/2014 02:08 AM, Takashi Iwai wrote: > At Tue, 03 Jun 2014 00:41:48 -0700, > Alex Fedora wrote: >> On 06/02/2014 05:05 AM, Takashi Iwai wrote: >>> At Mon, 02 Jun 2014 02:49:16 -0700, >>> Alex Fedora wrote: >>>> On 06/02/2014 02:15 AM, Takashi Iwai wrote: >>>>> At Mon, 02 Jun 2014 01:55:16 -0700, >>>>> Alex Fedora wrote: >>>>>> On 06/02/2014 01:03 AM, Takashi Iwai wrote: >>>>>>> At Sat, 31 May 2014 23:54:26 -0700, >>>>>>> Alex Fedora wrote: >>>>>>>> Hi, >>>>>>>> >>>>>>>> Please see problem description below. I've captured alsa info output for >>>>>>>> the latest kernel. Kernel just before the regression. And info from the >>>>>>>> kernel with the offending commit. But I had to calculate diff from first >>>>>>>> one to fit to 100K limit. If you want, I can send all 3 files as an >>>>>>>> attachments. Also the latest version of the kernel was took from >>>>>>>> http://kernel.ubuntu.com/~kernel-ppa/mainline/daily/current/linux-image-3.15.0-999-generic_3.15.0-999.201405310205_amd64.deb >>>>>>>> . >>>>>>>> >>>>>>>> Thank you, >>>>>>>> Alex >>>>>>>> >>>>>>>> [1.] One line summary of the problem: [Lenovo ThinkPad X61s] Speaker >>>>>>>> volume mutes itself in random intervals after upgrade to latest kernel >>>>>>>> [2.] Full description of the problem/report: >>>>>>>> >>>>>>>> Speaker mutes itself in random intervals. Headphones are working fine at >>>>>>>> that moment. I can see what volume of the Speaker drops to 0% and then >>>>>>>> back to 100% using alsamixer. It seems like Internal Mic Boost also >>>>>>>> jumps from 100% to 0% and back. It is regression from >>>>>>>> 5ccc618fee67f0f0b2122dd4b32a02fd2b6a1569 (ALSA: hda - Remove static >>>>>>>> quirks for AD1884/1984 & variants). I've tested previous commit >>>>>>>> aa95d61b43e0fcb0b2ce68e5efa37174fd9e5cd3 (ALSA: hda - Remove static >>>>>>>> quirks for AD1882) and audio works as expected. >>>>>>>> >>>>>>>> New kernel shows two controls "Headphone Playback Volume" and "Speaker >>>>>>>> Playback Volume". But old one shows single control "PCM Playback >>>>>>>> Volume". "Speaker Playback Volume" is the one which gets muted randomly. >>>>>>> Are you using PulseAudio? If yes, does the problem happen without PA? >>>>>>> This kind of random mute/unmute tends to be an issue of bogus >>>>>>> headphone or mic jack detection. >>>>>>> >>>>>>> >>>>>>> Takashi >>>>>>> >>>>>> Hi Takashi, >>>>>> >>>>>> Yes, I'm using PulseAudio. I've removed it and issue gone away. It >>>>>> returned back after reinstalling it. This is still driver bug, right? >>>>> Yes. So PA seems confused by some bogus information by the driver. >>>>> >>>>> Try to run "alsactl monitor 0", maybe better without PA at first. >>>>> Do you see occasional events about the jack detection even if you >>>>> don't plug/unplug? And, try it also with PA. >>>>> >>>>> >>>>> Takashi >>>>> >>>> Yes, I can see a lot of events even if I don't plug/unplug anything. >>>> Even more if I'm using PA. >>>> >>>> No pulseaudio: >>>> >>>> card 0, #22 (0,0,0,Mic Jack,0) VALUE >>>> card 0, #23 (0,0,0,Headphone Jack,0) VALUE >>>> card 0, #23 (0,0,0,Headphone Jack,0) VALUE >>> Thanks. These are definitely bogus events. No wonder that PA gets >>> screwed up. >>> >>> Could you try the patch below? >>> >>> >>> Takashi >>> >>> --- >>> diff --git a/sound/pci/hda/patch_analog.c b/sound/pci/hda/patch_analog.c >>> index 06275f8807a8..fa626648d746 100644 >>> --- a/sound/pci/hda/patch_analog.c >>> +++ b/sound/pci/hda/patch_analog.c >>> @@ -191,7 +191,6 @@ static int ad198x_parse_auto_config(struct hda_codec *codec, bool indep_hp) >>> int err; >>> >>> codec->spdif_status_reset = 1; >>> - codec->no_trigger_sense = 1; >>> codec->no_sticky_stream = 1; >>> >>> spec->gen.indep_hp = indep_hp; >>> @@ -377,6 +376,8 @@ static int patch_ad1986a(struct hda_codec *codec) >>> /* AD1986A has the inverted EAPD implementation */ >>> codec->inv_eapd = 1; >>> >>> + codec->no_trigger_sense = 1; >>> + >>> spec->gen.mixer_nid = 0x07; >>> spec->gen.beep_nid = 0x19; >>> set_beep_amp(spec, 0x18, 0, HDA_OUTPUT); >>> @@ -500,6 +501,8 @@ static int patch_ad1983(struct hda_codec *codec) >>> return err; >>> spec = codec->spec; >>> >>> + codec->no_trigger_sense = 1; >>> + >>> spec->gen.mixer_nid = 0x0e; >>> spec->gen.beep_nid = 0x10; >>> set_beep_amp(spec, 0x10, 0, HDA_OUTPUT); >>> @@ -588,6 +591,8 @@ static int patch_ad1981(struct hda_codec *codec) >>> return -ENOMEM; >>> spec = codec->spec; >>> >>> + codec->no_trigger_sense = 1; >>> + >>> spec->gen.mixer_nid = 0x0e; >>> spec->gen.beep_nid = 0x10; >>> set_beep_amp(spec, 0x0d, 0, HDA_OUTPUT); >>> @@ -913,6 +918,8 @@ static int patch_ad1988(struct hda_codec *codec) >>> return err; >>> spec = codec->spec; >>> >>> + codec->no_trigger_sense = 1; >>> + >>> spec->gen.mixer_nid = 0x20; >>> spec->gen.mixer_merge_nid = 0x21; >>> spec->gen.beep_nid = 0x10; >>> @@ -1140,6 +1147,8 @@ static int patch_ad1882(struct hda_codec *codec) >>> return err; >>> spec = codec->spec; >>> >>> + codec->no_trigger_sense = 1; >>> + >>> spec->gen.mixer_nid = 0x20; >>> spec->gen.mixer_merge_nid = 0x21; >>> spec->gen.beep_nid = 0x10; >>> . >>> >> I've applied the patch and rebuild the kernel. But it didn't solve the >> problem. I have applied the patch to kernel 3.13.11.2. Does it matter? >> Should I try to download latest kernel and try to apply patch to it? > It should be OK with 3.13.x as long as the patch is cleanly applied, > but better to test with the very latest kernel, I'd say. Yes, it have applied to that version without conflicts. I've used that repo for bisection of that problem. > > >> Here is the output of the monitor with PA enabled: > At this moment, let's concentrate on non-PA case. Then there should > be any events reported. Especially "XXX Jack" ones shouldn't be > reported unless plugged/unplugged. I've applied the latest patch you send. It didn't solve the problem. Still can see random events (non-PA): card 0, #22 (0,0,0,Mic Jack,0) VALUE card 0, #23 (0,0,0,Headphone Jack,0) VALUE card 0, #23 (0,0,0,Headphone Jack,0) VALUE card 0, #22 (0,0,0,Mic Jack,0) VALUE card 0, #22 (0,0,0,Mic Jack,0) VALUE card 0, #23 (0,0,0,Headphone Jack,0) VALUE card 0, #23 (0,0,0,Headphone Jack,0) VALUE card 0, #22 (0,0,0,Mic Jack,0) VALUE card 0, #22 (0,0,0,Mic Jack,0) VALUE card 0, #23 (0,0,0,Headphone Jack,0) VALUE card 0, #23 (0,0,0,Headphone Jack,0) VALUE card 0, #22 (0,0,0,Mic Jack,0) VALUE card 0, #22 (0,0,0,Mic Jack,0) VALUE > > Is the power-saving mode disabled? Check > /sys/module/snd_hda_intel/parameters/power_save. Echo 0 to that file > for disabling it. It was set to 0. I tried to set to 1 and back. Didn't help. Still can see Headphone Jack events. > > Maybe the surest way (and equivalent with the previous for your > device) is a oneliner like below. If it doesn't work either (you > still see the spurious jack events), the problem isn't about the > trigger sense setup. Unfortunately it didn't help. > Takashi > > --- > diff --git a/sound/pci/hda/patch_analog.c b/sound/pci/hda/patch_analog.c > index 06275f8807a8..fcbda351e340 100644 > --- a/sound/pci/hda/patch_analog.c > +++ b/sound/pci/hda/patch_analog.c > @@ -191,7 +191,7 @@ static int ad198x_parse_auto_config(struct hda_codec *codec, bool indep_hp) > int err; > > codec->spdif_status_reset = 1; > - codec->no_trigger_sense = 1; > + //codec->no_trigger_sense = 1; > codec->no_sticky_stream = 1; > > spec->gen.indep_hp = indep_hp; > . > Thanks, Alex
diff --git a/sound/pci/hda/patch_analog.c b/sound/pci/hda/patch_analog.c index 06275f8807a8..fcbda351e340 100644 --- a/sound/pci/hda/patch_analog.c +++ b/sound/pci/hda/patch_analog.c @@ -191,7 +191,7 @@ static int ad198x_parse_auto_config(struct hda_codec *codec, bool indep_hp) int err; codec->spdif_status_reset = 1; - codec->no_trigger_sense = 1; + //codec->no_trigger_sense = 1; codec->no_sticky_stream = 1; spec->gen.indep_hp = indep_hp;