XPost: linux.debian.bugs.dist 
 From: uncleruc2075@gmail.com 
  
 Hello again, 
 I have completed the bisection tests as you suggested. Here are the 
 final results: 
 The 6.13.5-1~exp1 kernel is GOOD. The system boots normally, and the 
 kernel panic does not occur. 
 The 6.13~rc6-1~exp1 kernel is BAD. The system fails to boot and 
 exhibits a kernel panic. I've attached the full dmesg log for this 
 test. 
  
 The panic appears to be a stack trace related to the CPU scheduler, 
 but the log also shows activity from the amdgpu driver, indicating the 
 underlying issue is still present. 
  
 This indicates that the fix for the initial amdgpu problem was 
 introduced between the 6.13~rc6 and 6.13.5 kernel versions. I have 
 completed all requested tests and hope this information is helpful in 
 narrowing down the problematic commit. Best regards, Ruslan Semchenko 
  
 €€€€, 20 €€€€€€. 2025€€€€€. €€ 19:01, Salvatore Bonaccorso : 
 > 
 > Control: tags -1 + trixie 
 > Control: fixed -1 6.16.1-1~exp1 
 > Hi, 
 > 
 > On Wed, Aug 20, 2025 at 05:08:05PM +0300, €€€€€€€€€€€€ €€€€€€€€€€€€€€€€ 
 wrote: 
 > > Hello, Uwe, 
 > > Thank you for your response and for pointing out the confusion in my 
 > > previous message. I apologize for the unclear report. I performed two 
 > > separate tests as you requested, and I will provide the full logs for 
 > > each. 
 > > Test 1: Stable Kernel (6.12.41+deb13-amd64) This is the kernel that 
 > > comes with Debian 13 (Trixie). The system fails to boot with a kernel 
 > > panic related to the amdgpu driver, as initially reported. The full 
 > > dmesg log for this test is attached. 
 > > 
 > > Test 2: Experimental Kernel (6.16-amd64) This kernel was installed 
 > > from the 'experimental' repository. The system boots normally, and the 
 > > kernel panic does not occur. The full dmesg log for the successful 
 > > boot is also attached. 
 > > This indicates that the issue appears to be resolved in the 6.16 
 > > kernel. I hope these complete logs help. 
 > 
 > Thank you that is already very helpful to asess this for upper 
 > versions, i.e. we can mark the metadata accordingly. 
 > 
 > Would you be in the position to try to bisect the upstream changes and 
 > see if we can isolate which change upstream did fix the issue? A start 
 > would be to first "bisect" the Debian kernel images version, but is is 
 > not a straighforward line as we know 6.12.41-1 is problematic, and 
 > 6.16.1-1~exp1 is good. 
 > 
 > So we have the following candidates uploaded later on to experimental: 
 > 
 >         6.16.1-1~exp1 
 >         6.16-1~exp1 
 >         6.16~rc7-1~exp1 
 >         6.15.6-1~exp1 
 >         6.15.5-1~exp1 
 >         6.15.4-1~exp1 
 >         6.15.3-1~exp1 
 >         6.15.2-1~exp1 
 >         6.15.1-1~exp1 
 >         6.15-1~exp1 
 >         6.15~rc7-1~exp1 
 >         6.14.6-1~exp1 
 >         6.14.5-1~exp1 
 >         6.14.3-1~exp1 
 >         6.13.11-1~exp1 
 >         6.13.10-1~exp1 
 >         6.13.9-1~exp1 
 >         6.13.8-1~exp1 
 >         6.13.7-1~exp1 
 >         6.13.6-1~exp1 
 >         6.13.5-1~exp1 
 >         6.13.4-1~exp1 
 >         6.13.3-1~exp1 
 >         6.13.2-1~exp1 
 >         6.13~rc7-1~exp1 
 >         6.13~rc6-1~exp1 
 > 
 > You could start by fetching 6.13~rc6-1~exp1 and see if that is bad as 
 > well, before bisecting though check the major version bumps, so 
 > 6.14.3-1~exp1, 6.15~rc7-1~exp1. If we have found the two major 
 > versions which are one one side good and on the other bad, we have 
 > IMHO a stating point to bisect now the upstream versions. 
 > 
 > You can get the Debian packages for the (signed) image via the 
 > snapshots service: 
 > https://snapshot.debian.org/package/linux-signed-amd64/ 
 > 
 > (Remember to remove after the experiment old images again). 
 > 
 > Once we have the two major upstream version on one side bad and on the 
 > other good, we could start bisection, roughly it should go the 
 > following way. Say hypotetical you foun 6.13~rc6-1~exp1 as bad 
 > 6.14.3-1~exp1 as good, then 
 > 
 > 
 >         git clone https://git.kernel.org/pub/scm/linux/kernel/ 
 it/stable/linux-stable.git 
 >         cd linux-stable 
 >         git checkout v6.14.3 
 >         cp /boot/config-$(uname -r) .config 
 >         yes '' | make localmodconfig 
 >         make savedefconfig 
 >         mv defconfig arch/x86/configs/my_defconfig 
 > 
 >         # test 6.14.3 to ensure this is "good" 
 >         make my_defconfig 
 >         make -j $(nproc) bindeb-pkg 
  
 [continued in next message] 
  
 --- SoupGate-Win32 v1.05 
  * Origin: you cannot sedate... all the things you hate (1:229/2)