Hi, These are git bisect log:
$ git bisect bad d42b44736ea29fa6d0c3cb9c75569314134b7732 is the first bad commit commit d42b44736ea29fa6d0c3cb9c75569314134b7732 (HEAD) Author: Himal Prasad Ghimiray <[email protected]> Date: Mon Oct 14 13:25:44 2024 +0530 drm/xe/gt: Update handling of xe_force_wake_get return [ Upstream commit 30d105577a3319094f8ae5ff1ceea670f1931487 ] Regards, Asho Asho Yeh - 信佑 <[email protected]> 於 2025年7月30日 週三 上午12:02寫道： > I will try "git bisect" to find which commit cause this bug. > > Asho Yeh - 信佑 <[email protected]> 於 2025年7月29日 週二 下午11:51寫道： > >> Hi, >> >> I downloaded 6.12.34, 6.12.35, 6.12.38, 6.12.40 from kernel.org. >> >> For kernel versions earlier than 6.12.34, during boot, dmesg shows the >> following: >> >> " >> xe 0000:03:00.0: [drm] Using GuC firmware from xe/bmg_guc_70.bin version >> 70.40.2 >> xe 0000:03:00.0: [drm] Using GuC firmware from xe/bmg_guc_70.bin version >> 70.40.2 >> xe 0000:03:00.0: [drm] Using HuC firmware from xe/bmg_huc.bin version 8.2.10 >> xe 0000:03:00.0: [drm] Found BATTLEMAGE (device ID e20b) display version >> 14.01 stepping B0 >> >> " >> >> For kernel versions 6.12.35 and above, during boot, dmesg shows: >> "xe 0000:03:00.0: probe with driver xe failed with error -110" >> >> (Note: To check this, need to SSH into the system from another computer >> to run dmesg.) >> >> This is what I’ve observed, though I’m not sure if it matches the >> information from Salvatore. >> As for which commit caused this, I’m not sure. I manually compared the >> source code across versions and created the patch mentioned in my previous >> mail. >> >> Regards, >> >> Asho >> Salvatore Bonaccorso <[email protected]> 於 2025年7月29日 週二 下午11:23寫道： >> >>> Hi, >>> >>> On Tue, Jul 29, 2025 at 11:15:50PM +0800, Asho Yeh - 信佑 wrote: >>> > Hi, >>> > >>> > I made a patch with B580 support. >>> > Tested with kernel-6.12.35-deb13-amd64 and 6.12.38-deb13-amd64. >>> > >>> > I am not drm or gpu expert. Just digged the source code and found some >>> > differences. >>> > Hope this can help. >>> >>> have you seen my earlier question? >>> >>> As you were able to test upstream version, can you please bisect the >>> changes between 6.12.33 and 6.12.35 to identify the breaking ocmmit >>> and then report it back here? >>> >>> Regards, >>> Salvatore >>> >>
