On Tue, Jan 03, 2023 at 02:39:49AM +0000, Daniel Lee wrote: > > >Fix: > For some reason, the issue seems to disappear completely if the kernel > is compiled with -DUAUDIO_DEBUG. I discovered this by fetching the > latest sources from cvs and building a custom kernel config with > option UAUDIO_DEBUG. I installed the new kernel without building > the userland (so the userland is from the snapshot while the kernel > is built from cvs HEAD, probably not a good idea.) The contents of > the config file are: > > include "arch/amd64/conf/GENERIC.MP" > > options UAUDIO_DEBUG > > except there is a blank line at the start, per config(8). >
Thank you for the report. Using cvs HEAD for testing uaudio(4) makes sense. Just to be sure that I understand it: does removing "options UAUDIO_DEBUG" and rebuilding the kernel results in a kernel that triggers the problem? I'm asking because it's unclear to me if the problem is fixed by switching to cvs HEAD or by adding UAUDIO_DEBUG.