2.6.36: Sound stop working
in [Kernel]
|
Prev: [tip:perf/urgent] perf: Add back list_head data types
Next: [tip:x86/uv] x86, UV: Make kdump avoid stack dumps - fix !CONFIG_KEXEC breakage
From: Thomas Meyer on 12 Aug 2010 16:10 My sound card stop to work since a few commits. I tried to bisect it and ended up with this: There are only 'skip'ped commits left to test. The first bad commit could be any of: 3bcf3860a4ff9bbc522820b4b765e65e4deceb3e c1e5c954020e123d30b4abf4038ce501861bcf9f We cannot bisect more! any ideas what to do now? -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Pekka Enberg on 12 Aug 2010 16:20 On Thu, Aug 12, 2010 at 11:00 PM, Thomas Meyer <thomas(a)m3y3r.de> wrote: > My sound card stop to work since a few commits. I tried to bisect it and ended up with this: > > There are only 'skip'ped commits left to test. > The first bad commit could be any of: > 3bcf3860a4ff9bbc522820b4b765e65e4deceb3e > c1e5c954020e123d30b4abf4038ce501861bcf9f > We cannot bisect more! > > any ideas what to do now? Takashi, maybe this is related to Linus' problem? -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Takashi Iwai on 12 Aug 2010 16:30 At Thu, 12 Aug 2010 23:10:15 +0300, Pekka Enberg wrote: > > On Thu, Aug 12, 2010 at 11:00 PM, Thomas Meyer <thomas(a)m3y3r.de> wrote: > > My sound card stop to work since a few commits. I tried to bisect it and ended up with this: > > > > There are only 'skip'ped commits left to test. > > The first bad commit could be any of: > > 3bcf3860a4ff9bbc522820b4b765e65e4deceb3e > > c1e5c954020e123d30b4abf4038ce501861bcf9f > > We cannot bisect more! > > > > any ideas what to do now? > > Takashi, maybe this is related to Linus' problem? Yeah, I guess so, since I couldn't see any obvious problem in sound/*. PulseAudio uses inotify in udev-detection module, at least. It's possible that inotify change may hit the sound in the end. thanks, Takashi -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Linus Torvalds on 12 Aug 2010 16:30 On Thu, Aug 12, 2010 at 1:10 PM, Pekka Enberg <penberg(a)kernel.org> wrote: > On Thu, Aug 12, 2010 at 11:00 PM, Thomas Meyer <thomas(a)m3y3r.de> wrote: >> My sound card stop to work since a few commits. I tried to bisect it and ended up with this: >> >> There are only 'skip'ped commits left to test. >> The first bad commit could be any of: >> 3bcf3860a4ff9bbc522820b4b765e65e4deceb3e >> c1e5c954020e123d30b4abf4038ce501861bcf9f >> We cannot bisect more! >> >> any ideas what to do now? > > Takashi, maybe this is related to Linus' problem? Yes, I'm 99% sure it is. I haven't bisected my problem all the way, but I've bisected away all the sound changes (and as mentioned, doing a 'cat' to /dev/audio works). And yes, the fanotify changes are right in the middle of my bisect. So I suspect it's not sound that is broken at all, but pulseaudio that got broken by the fanotify changes. Eric - over to you. Linus -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Linus Torvalds on 12 Aug 2010 17:10
On Thu, Aug 12, 2010 at 1:20 PM, Linus Torvalds <torvalds(a)linux-foundation.org> wrote: > > So I suspect it's not sound that is broken at all, but pulseaudio that > got broken by the fanotify changes. Confirmed. That broken commit 3bcf3860a4ff9bb doesn't even boot for me (looks like a stack smash recursion), and the commit to "fix" it (c1e5c954020e12) is really too ugly to live. I think we need to totally undo the whole "struct file" thing, and just admit that it was a mistake. The code really wants a "struct path", and using a struct file screws up all the refcounting and is just not right. The fact that dentry_open() may have some problem needs to be fixed -there- rather than make the callers do crazy things that they don't want to do and can't do sanely. Linus -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/ |