At Tue, 19 Nov 2013 00:40:07 -0800, Joe Perches wrote:
On Fri, 2013-11-15 at 07:56 +0100, Takashi Iwai wrote:
At Thu, 14 Nov 2013 09:52:34 -0800, Joe Perches wrote:
Hrm, the codec proc for the analog audio isn't seen here. Maybe the codec communication was already screwed up at this point.
Can you have a good output by any chance (showing the right contents of /proc/asound/card1/codec#*)?
Using the linux version I've normally used:
$ cat /proc/version Linux version 3.11.0-031100rc2-generic (apw@gomeisa) (gcc version 4.6.3 (Ubuntu/Linaro 4.6.3-1ubuntu5) ) #201307211535 SMP Sun Jul 21 19:45:33 UTC 201
$ cat /proc/asound/card1/codec#0 Codec: Realtek ALC271X
So this codec was detected properly. Judging from your patch description, the verb triggering the stall (0x170500) was about the power up to D0 for AFG node. Did the sound work even after the flood of the messages?
The problem might be also not a stall but a bogus CORB/RIRB pointer. Try the first patch, and give the first "spurious response" message. Now it shows rp and wp values, and we might see a pattern in them to trigger the spurious message (e.g. wp is always same value).
Also, to be sure, try the second patch, too. It disables the access to some vendor-specific verbs.
Hi again.
I'm not ignoring you. I've rebuilt the kernel with these patches and I'm waiting for the messages to show up again.
The only times I've seen them have been after hibernating and resuming.
Ah, that's an important piece.
I've done several hibernate/restart cycles but haven't had the messages emitted in the logs yet.
OK, let me know if you still hit the issue.
thanks,
Takashi