Bug 103431
| Summary: | NULL dereference at hdmi_chmap_ctl_info | ||
|---|---|---|---|
| Product: | Drivers | Reporter: | Andreas Reis (andreas.reis) |
| Component: | Sound(ALSA) | Assignee: | Jaroslav Kysela (perex) |
| Status: | RESOLVED CODE_FIX | ||
| Severity: | high | CC: | tiwai |
| Priority: | P1 | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Kernel Version: | drm-intel-nightly / sound-unstable | Subsystem: | |
| Regression: | No | Bisected commit-id: | |
| Attachments: |
Test patch
desktop kernel config dmesg from sound/for-next output of alsa-info.sh Fix patch |
||
|
Description
Andreas Reis
2015-08-24 13:54:42 UTC
Reverting the commit fa4f18b4f402d3654 helps? Yes. (Two more unrelated things I might add: 1. The "Bug Tracker" link on alsa-project.org still 404's and should probably better redirect to the Sound(ALSA) list here. 2. On my CA0132 the Master volume slider every now and then just ceases to have an effect. Haven't filed a bug as I have no idea how to reproduce it, and it might as well be another firmware issue.) (In reply to Takashi Iwai from comment #1) > Reverting the commit fa4f18b4f402d3654 helps? It works on my machine. How is the way to reproduce? Please give your kernel config. I ignore the rest of your unrelated comments. Also, does the patch below have any influence on the behavior? Created attachment 185701 [details]
Test patch
No change with the test patch. grep 'SND.*=' on my desktop's config with the bug: CONFIG_SND=y CONFIG_SND_TIMER=y CONFIG_SND_PCM=y CONFIG_SND_HWDEP=y CONFIG_SND_JACK=y CONFIG_SND_SEQUENCER=y CONFIG_SND_OSSEMUL=y CONFIG_SND_MIXER_OSS=m CONFIG_SND_PCM_OSS=m CONFIG_SND_PCM_OSS_PLUGINS=y CONFIG_SND_SEQUENCER_OSS=y CONFIG_SND_HRTIMER=y CONFIG_SND_SEQ_HRTIMER_DEFAULT=y CONFIG_SND_DYNAMIC_MINORS=y CONFIG_SND_MAX_CARDS=32 CONFIG_SND_PROC_FS=y CONFIG_SND_VERBOSE_PROCFS=y CONFIG_SND_VMASTER=y CONFIG_SND_DMA_SGBUF=y CONFIG_SND_DRIVERS=y CONFIG_SND_PCI=y CONFIG_SND_HDA=y CONFIG_SND_HDA_INTEL=y CONFIG_SND_HDA_HWDEP=y CONFIG_SND_HDA_RECONFIG=y CONFIG_SND_HDA_INPUT_BEEP=y CONFIG_SND_HDA_INPUT_BEEP_MODE=1 CONFIG_SND_HDA_PATCH_LOADER=y CONFIG_SND_HDA_CODEC_HDMI=y CONFIG_SND_HDA_CODEC_CA0132=y CONFIG_SND_HDA_CODEC_CA0132_DSP=y CONFIG_SND_HDA_GENERIC=y CONFIG_SND_HDA_POWER_SAVE_DEFAULT=0 CONFIG_SND_HDA_CORE=y CONFIG_SND_HDA_DSP_LOADER=y CONFIG_SND_HDA_I915=y CONFIG_SND_HDA_PREALLOC_SIZE=4096 The deference does not occur when I switch alsa to modules, in which case: zcat /proc/config.gz | grep 'SND.*=m' CONFIG_SND=m CONFIG_SND_TIMER=m CONFIG_SND_PCM=m CONFIG_SND_HWDEP=m CONFIG_SND_SEQUENCER=m CONFIG_SND_MIXER_OSS=m CONFIG_SND_PCM_OSS=m CONFIG_SND_HRTIMER=m CONFIG_SND_HDA=m CONFIG_SND_HDA_INTEL=m CONFIG_SND_HDA_CODEC_HDMI=m CONFIG_SND_HDA_CODEC_CA0132=m CONFIG_SND_HDA_GENERIC=m CONFIG_SND_HDA_CORE=m Could you explain *how* to reproduce this? Also, please give the full kernel messages containing Oops (preferably from the kernel with sound modules). As mentioned, I couldn't reproduce this on my local machines (tried drm-intel-nightly, too). So, we need to align the environment. Let's try the following: - Build the kernel without drm-intel-nightly. Use vanilla merged sound/for-next as the test basis. Build sound modules for ease of tests. - Give the full kernel config. A difference of debug config might influence on such behavior. - Remove ca0132 board once and test. If it's not possible, try to disable ca0132 part via enable option of snd-hda-intel module. Created attachment 185851 [details]
desktop kernel config
I get it by simply booting when the modules are compiled-in. ie, immediately and automatically.
- Just tried sound/for-next, same.
- See attachment. Laptop's display is one eDP, desktop's one HDMI connected to the mainboard.
- Can't, it's the mainboard's sound chip. Not relevant either, a) the laptop (with a Realtek ALC282) exhibits the same bug and b) the bug occurs even with CA0132 completely disabled.
Created attachment 185861 [details]
dmesg from sound/for-next
desktop, SND modules compiled-in, CA0132 disabled
Why you don't enable CONFIG_SND_HDA_REALTEK=y? And which device is what? Please give alsa-info.sh taken from the working kernel. Created attachment 185871 [details]
output of alsa-info.sh
Because this is my desktop kernel config, not my laptop one.
Desktop: CA0132
Laptop ALC282
Both onboard chips, both on Haswell boards.
Ah, please explain clearly at the next time... It was confusing! Meanwhile, I found the culprit. It's seen in dmesg before the Oops, [ 0.958225] WARNING: CPU: 0 PID: 4 at fs/sysfs/dir.c:31 sysfs_warn_dup+0x66/0x80() [ 0.958226] sysfs: cannot create duplicate filename '/devices/pci0000:00/0000:00:03.0/hdaudioC0D0/widgets' So Oops was actually a red herring. The fix patch is attached below. At the next time, please give the full dmesg and alsa-info.sh output from the beginning. This would reduce the time for bug hunt significantly. Created attachment 185891 [details]
Fix patch
Forgot to confirm, yes that indeed fixes it. You're right ofc about what I should have included from the start. Sorry. |