On Fri, 13 Jan 2017 20:42:06 +0100, Enrico Mioso wrote:
Hello guys. I finally can report some interesting news. Now audio stopped working - still, the system didn't panic. So I consider this a good good point. My system is alive and I can keep it alive indefinitely, clearly if hardware doesn't break and/or other events out of my control cause an interruption.
In my dmesg, the following messages started to appear: snd_hda_intel 0000:00:1b.0: azx_get_response timeout, switching to polling mode: last cmd=0x020c0000
I suppose the last_cmd value isn't always same?
snd_hda_codec_realtek hdaudioC0D0: Unable to sync register 0x1f0e00. -5
Hm, this might be an invalid verb.
And they appear apparently when the driver tries to send new commands to the device, but this is only my impression. Let me know what I could do to report on what's happening. My system is acually running with kdump active, so in case I may also invoke the crash sysrq action and send back results.
When a crash no longer happens, kdump isn't needed. Instead, we need to track what triggers the stall.
Try to enable the tracepoint, as mentioned in Documentation/sound/hd-audio/notes.rst.
thanks,
Takashi