El 5/1/23 a las 8:07, Alexandre Ratchov escribió:
Your device seems to stop because the azalia pci host stops (probably
interrupts stop and a reboot is needed), while this diff is about the
Yes,exactly. A quick review using vmstat -i show that, azalia interrupts
is frozen, only hard reboot resolve this temporaly.
azalia codec (i.e. about handling the "analog" parts). So, I'm very
surprised this fixes the random audio hangs (though any strange
interaction between the host and the codec can't be excluded).
For the patch I investigate various things:
1.- Using sndiod -mplay option the audio stuck randomly. This workaround
is very use in other cases, but in mine don´t work.
2.- Disabled MSI signaling. In the Hudson audio problem, disable MSI
signaling solve problem. AMD Summit Ridge/Raven Ridge HD Audio need this
fix for work correctly, but in my case the audio don´t work.
3.- I tried using PCIE SNOOP aproximation (used in AMD Summit
Ridge/Raven Ridge HD Audio) but again, the audio don't work.
4.- Investigating ALC892 datasheet, I can see that ALC888 and ALC892 are
"mostly" same codecs in specs. ALC888 use same code:
* this->qrks |= AZ_QRK_WID_CDIN_1C | AZ_QRK_WID_BEEP_1D; *
In this case I use a conditional to apply this code and it works, there
is no audio problem with this patch. The only problem I see is that it
is very specific to my hardware case (subid), but since the hardware and
its configuration can change, enabling the option globally to the codec
may end up affecting the audio of other configurations.
Are there other changes that may have fixed audio? Any BIOS tweaks?
Suspend-resume cycles? Any devices disabled meanwhile?
No, BIOS config is same, not changes or updates. Suspend-resume is a
feature that I'm not use. Same device, nothing disabled.
Could you test with and without the diff applied in exactly the same
conditions?
I test the patch using this approx:
1.- Using mpd for listen audio only. Without patch audio stuck in a
short and randomly time (few minutes, generally less a hour)
2.- Using various apps streaming audio (Chrome, Firefox and MPD,
precisely) in same time, the audio stuck very quickly without patch.
3.- Recording, audio stuck in few minutes (using Audacity).
In all case, the patch resolve problem for me, various hours
playing/recording and no audio stuck.
These hangs are a longstanding problem (and hard to debug because of
the randomness of the hangs), so I'm very intested in understanding
what's causing it. Thanks
Yeah, I can see this, in my investigation I see reports from 2018, and
in various case the issue is very strange.
*** Note: Sorry for private, I'm remember how use the mailing list ***
--
Dios en su cielo, todo bien en la Tierra