[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#1108496: linux-image-6.12.30+bpo-amd64: USB audio glitches: retire_capture_urb



Control: tag -1 moreinfo

On Sun, 2025-06-29 at 19:13 +0100, Jamie Forth wrote:
[...]
> There are no differences in default ALSA parameters running under 6.1
> or 6.12 kernels.
> 
> $ cat /proc/asound/USB/pcm0c/sub0/sw_params

You mean "hw_params", right?

>   access: MMAP_INTERLEAVED
>   format: S32_LE
>   subformat: STD
>   channels: 10
>   rate: 48000 (48000/1)
>   period_size: 128
>   buffer_size: 32768
> 
> $ cat /proc/asound/USB/pcm0c/sub0/sw_params
> 
>   tstamp_mode: ENABLE
>   period_step: 1
>   avail_min: 128
>   start_threshold: 9223372036854775807
>   stop_threshold: 32768
>   silence_threshold: 0
>   silence_size: 0
>   boundary: 4611686018427387904

But these are for the capture side (pcm0c), whereas I thought the
problem was with playback (pcm0p)?

[...]
> The additional IRQs listed under the 6.12 kernel don't appear to
> exist, only irq/155-xhci_hcd in this case.

They exist and have been allocated but are not used.  The XHCI PCI
driver can currently allocate multiple IRQs but only installs an IRQ
handler for the first one.  This is a bit strange but should not cause a
problem because it also doesn't configure the hardware to use the
additional IRQs.

[...]
> ** Loaded modules:
[...]
> nvidia_drm(POE)
> nvidia_modeset(POE)
> nvidia(POE)
[...]

I see some reports that the Nvidia proprietary drivers can cause latency
spikes that affect audio.  Can you please test whether this problem
still occurs when using the nouveau driver instead?

Ben.

-- 
Ben Hutchings
73.46% of all statistics are made up.

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: