On 2019-01-01 17:17, Programmingkid wrote: > >> On Jan 1, 2019, at 7:27 AM, Zoltán Kővágó <dirty.ice...@gmail.com> wrote: >> >> On 2019-01-01 05:24, Programmingkid wrote: >>>> On 2018-12-29 01:49, Programmingkid wrote: >> [snip] >>>>> >>>>> I tried ac97 with a Windows 2000 guest in qemu-system-i386 - same demonic >>>>> sound. >>>>> With the above configuration but with an es1370 sound card I heard the >>>>> same sound. I kept seeing "es1370: warning: non looping mode" being >>>>> printed in the terminal. This problem might need its own patch. >>>>> Using the sb16 sound card I heard the same unsettling sound. >>>>> >>>>> I could not test the HDA driver due to problems with my Windows 7 VM. >>>>> >>>> >>>> >>>> I have a feeling that bf870a0cf5e2c2dd7438e65473b4fca1fb0ca5d1 >>>> (coreaudio: port to the new audio backend api) is the commit which >>>> breaks it, but without a mac I don't know how could I debug it. The way >>>> it initializes coreaudio changed a bit, that can cause problems. >>>> Could you please test whether this commit is the culprit or not? >>>> >>>> Regards, >>>> Zoltan >>> >>> >>> Hi how do you want me to test your theory? Revert the patch? If I did I >>> don't think QEMU would still compile. >> >> No, don't revert it, that would generate too many conflicts. Just >> >> git checkout bf870a0c > > This causes the very disturbing sound to play. > >> >> and >> >> git checkout bf870a0c~ > > This also causes the very disturbing sound to play. > >> >> should give you two versions that you can compare. >> >> Regards, >> Zoltan > > Thank you. >
That's worrying, since before that commit I didn't really do any big refactoring on coreaudio itself, and without a Mac I have no idea how could I debug it. Could you at least try a git bisect and figure out which commit broke it? Thanks, Zoltan
