I'm not sure that this is a kernel issue anymore. Perhaps it is pipewire or similar. I need to make more trouble shooting. I noticed, by mistake, that audio always work in GDM in all kernels.
On Sun, 2025-08-03 at 21:04 +0200, Salvatore Bonaccorso wrote: > Hi, > > On Sun, Aug 03, 2025 at 07:31:12PM +0200, [email protected] wrote: > > Perhaps I did something wrong but here is the file > > > > On Sat, 2025-08-02 at 20:24 +0200, Salvatore Bonaccorso wrote: > > > git bisect log > > > # bad: [259f4977409c87a980fa2227b7c76a2fe3fb8c2f] Linux 6.12.38 > > # good: [783cd2c3dca8b6c434e955b84c20c8940588dc68] Linux 6.12.35 > > git bisect start 'v6.12.38' 'v6.12.35' > > # good: [75705b44e0b9aaa74f4c163d93d388bcba9e386a] vsock/vmci: > > Clear the vmci transport packet properly when initializing it > > git bisect good 75705b44e0b9aaa74f4c163d93d388bcba9e386a > > # good: [b77a5ecb3d3baac85fcbf841b27b5ef7024a533f] bonding: Mark > > active offloaded xfrm_states > > git bisect good b77a5ecb3d3baac85fcbf841b27b5ef7024a533f > > # good: [c584b9b62c0cfdb064b91395e96c6ba78f955dae] > > platform/x86/amd/pmc: Add PCSpecialist Lafite Pro V 14M to 8042 > > quirks list > > git bisect good c584b9b62c0cfdb064b91395e96c6ba78f955dae > > # good: [b68e355a613260127c77c9ffd762ac7acbbcee77] usb: cdnsp: do > > not disable slot for disabled slot > > git bisect good b68e355a613260127c77c9ffd762ac7acbbcee77 > > # good: [b11397bf9ade076873605f8ce903afa9e9650548] platform/x86: > > think-lmi: Create ksets consecutively > > git bisect good b11397bf9ade076873605f8ce903afa9e9650548 > > # good: [0720e436e594d330bc10851889d90f6b48a2b32d] x86/bugs: Rename > > MDS machinery to something more generic > > git bisect good 0720e436e594d330bc10851889d90f6b48a2b32d > > # good: [331cfdd27429ed7a64923123b2482e85238979fa] > > x86/microcode/AMD: Add TSA microcode SHAs > > git bisect good 331cfdd27429ed7a64923123b2482e85238979fa > > # good: [fbad404f04d758c52bae79ca20d0e7fe5fef91d3] Linux 6.12.37 > > git bisect good fbad404f04d758c52bae79ca20d0e7fe5fef91d3 > > # good: [faac2abe895d200615a91acb63a709feb3dac1c2] x86/CPU/AMD: > > Properly check the TSA microcode > > git bisect good faac2abe895d200615a91acb63a709feb3dac1c2 > > # first bad commit: [259f4977409c87a980fa2227b7c76a2fe3fb8c2f] > > Linux 6.12.38 > > Yes that is very odd and does not really make sense. Can you exclude > that you no bisect classification was wrong here, because technically > your last good commit to the fist bad commit, the release of 6.12.38 > is a "no-change" update only releasing the version with the prvious > commit. > > Can you confirm if 6.12.38 as self-compiled is as well just ok? But > the kernel-image is not? That would indicate we might have a Debian > specific regression in the changes from upstream 6.12.38 to the > Debian > packaged version. > > So please confirm if 6.12.38 as self compiled as per above is as well > "good". > > Regards, > Salvatore
