XPost: linux.debian.kernel
Control: tags -1 + upstream moreinfo
Control: found -1 6.13~rc6-1~exp1
Hi
On Wed, Feb 12, 2025 at 03:56:38PM -0500, Jason McCormick wrote:
Looking through the changes beween 6.1.119 and 6.1.123 I don't spot something obviously suspicious. There were no changes in sound/core/oss.
I ended up writing to Takashi Iwai who authored this patch: ALSA: usb-audio: Notify xrun for low-latency mode - kernel/git/stable/linux.git - Linux
kernel stable tree <https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=linux-6.1.y&id=8c28274b2dccdcf2fe29ef625f3c5f8a09e2b809>
which is where we believe the issue originated. This was applied to the 6.1.y, 6.6.y, and 6.12.y branches along with the latest. Our project was
able to confirm the issue in all three kernels. Takaski confirmed that this created a change in snd_usb_audio that "reveals the buffer underrun on the device that was ignored beforehand". But dealing with the underrun was causing the audio to be flushed with a bad chop. Loading the snd_usb_audio module with lowlatency=0 on systems that encounter the problem resolves the issue. Basically, it seems like fixing this issue in the kernel triggered other downstream audio issues interacting with USB audiot devices. It's
added a new emphasis for us to get off of the last part of OSS once-and-for-all.
Have you asked this public on a mailing list? Can you reference the
current discussion, is Takaski planning to fix the regression?
If not, and given there is a workaround which will work enough and the
fade off snd_pcm_oss, should we close this bug?
Regards,
Salvatore
--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)