18:22 gekret005: I have a reproducible issue on my setup with the 7700x gfx 1036 where opening up 5 instances of alacritty times result in full screening them make them white
18:23 gekret005: no idea how to create something meaningfull from this
18:23 gekret005: it also resulted in both my displays flickering white really fast but I haven't recreated that
18:32 agd5f: hakzsam, https://gitlab.freedesktop.org/agd5f/linux/-/commits/gpu_fault_info_ioctl rebased
18:52 Venemo: gekret005: open an issue in the mesa gitlab
19:01 gekret005: Venemo: is there anything I can include or just a testcase to reproduce
19:07 Venemo: gekret005: the bug report template will tell you what you need to provide
22:17 gekret005: how do i dump the vbios properly i tried the echo 1 rom then cat thing but its only 60k compared fo 2m from gpu-z
22:19 gekret005: read that dumping bios from csm mode results in working pci passthrough for the 7900 xtx
22:26 agd5f: gekret005, you can grab it from debugfs /sys/kernel/debug/dri/0/amdgpu_vbios
22:29 agd5f: gekret005, the driver can fetch the vbios from the ROM directly on most dGPUs so it should be accessible even if the VM doesn't pass the ROM BAR through properly
22:34 gekret005: agd5f: that file is also 60k, when i used gpu-z that was in uefi, what is gpu-z doing to get a 2m vbios?
22:35 gekret005: or is it expected that the vbios is only 60k from csm
22:37 gekret005: i cant dump from a vm as the gpu stalls the vm init process
22:40 gekret005: also should there even be a difference between vbios dumped from csm and uefi?
22:53 gekret005: well i can say that the 60k vbios is different from what i get on linux uefi
22:54 agd5f: gekret005, 60K seems more reasonable. The roms are usually between 64-128K at most
23:42 gekret005: unfortunately passing this rom dumped from csm did not allow the gpu to work in the vm at all still