You are not logged in.
So I have just switched from a 2060 SUPER to an RX 7800 XT, and for some reason:
rebooting from the 6.7.0-rc3-1-mainline kernel into any other kernel makes the GPU fail to initialize,
rebooting from any other kernel works fine,
cold booting into any kernel works fine.
When the GPU fails to initialize, the following can be found in journalctl -b, the card won't output any display while still lighting up, lspci still lists the card but takes a long time, vulkaninfo reports not seeing any GPU at all.
Nov 29 21:57:34 orbital kernel: amdgpu 0000:09:00.0: [drm:amdgpu_ring_test_helper [amdgpu]] *ERROR* ring mes_kiq_3.1.0 test failed (-110)
Nov 29 21:57:34 orbital kernel: [drm:amdgpu_gfx_enable_kcq [amdgpu]] *ERROR* KCQ enable failed
Nov 29 21:57:34 orbital kernel: [drm:amdgpu_device_init [amdgpu]] *ERROR* hw_init of IP block <gfx_v11_0> failed -110
Nov 29 21:57:34 orbital kernel: amdgpu 0000:09:00.0: amdgpu: amdgpu_device_ip_init failed
Nov 29 21:57:34 orbital kernel: amdgpu 0000:09:00.0: amdgpu: Fatal error during GPU init
Things I've tried:
adding
amdgpu.aspm=0 amdgpu.runpm=0 amdgpu.bapm=0 pcie_aspm=off
kernel parameters, doesn't work
adding MODULES=(amdgpu) in /etc/mkinitcpio.conf, doesn't work
How can I fix/troubleshoot this problem further?
Thank you,
flippette
Last edited by flippette (2023-12-11 18:19:49)
Offline
I am experiencing the same exact behavior on Debian (unstable). Actually I had problems with every kernel I have tried, in pre 6.7 rc kernels either halt/reboot don't work, or suspend doesn't work properly. My main OS is Debian, but I have tried others (mainly Arch, Fedora), and in each I have problems on my 7800 XT, except for Kubuntu, there everything works as its supposed to. But I don't like Ubuntu, so screw it. The kernel that runs the best for me is Liquorix where halt/reboot works every time, and suspend mostly works.
I had high hopes for 6.7 since there are some AMD GPU improvements, and halt/reboot AND suspend works, but I ran in exact same problem as you described. I tried a precompiled 6.7 rc3 (from Siduction) and compiled my own. I hope this gets resolved in the next few rcs, since I want the computer to work as intended.
Offline
I just compiled 6.7.0-rc4 (so 4th release candidate) and sadly the same exact behavior as rc3.
Offline
6.7.0-rc5 seems to exhibit the same behavior.
Offline
Still the same for rc6, at least the one I compiled.
Offline
cold booting into any kernel works fine.
Is this also valid for the 6.7 kernels ?
Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.
clean chroot building not flexible enough ?
Try clean chroot manager by graysky
Offline
For me it is. Cold booting works every time regardless of the kernel.
Offline
Someone filed an issue on the upstream drm/amd GitLab repo.
Offline
Someone filed an issue on the upstream drm/amd GitLab repo.
Need any help with the bisection requested by upstream?
Offline
flippette wrote:Someone filed an issue on the upstream drm/amd GitLab repo.
Need any help with the bisection requested by upstream?
I'm not very familiar with the kernel myself, so I'd definitely appreciate some pointers here.
Offline
@flippette did you ever try 6.7-rc1 and can recall if that had the issue? This may help reduce the number of revisions needed to be bisected. Is the issue still present in 6.7-rc7?
Offline
I can confirm the issue is still present in rc7, I never tried rc1.
Offline
https://drive.google.com/file/d/1Y_svoI … sp=sharing linux-mainline-6.7rc1-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/1S_2Cfm … sp=sharing linux-mainline-headers-6.6.r8649.g7d461b291e65-1-x86_64.pkg.tar.zst
Assuming rc1 linked above has the issue:
$ git bisect start
status: waiting for both good and bad commits
$ git bisect bad v6.7-rc1
status: waiting for good commit(s), bad commit known
$ git bisect good v6.6
Bisecting: 7553 revisions left to test after this (roughly 13 steps)
[7d461b291e65938f15f56fe58da2303b07578a76] Merge tag 'drm-next-2023-10-31-1' of git://anongit.freedesktop.org/drm/drm
https://drive.google.com/file/d/18oXvAF … sp=sharing linux-mainline-6.6.r8649.g7d461b291e65-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/1t1bGgv … sp=sharing linux-mainline-headers-6.7rc1-1-x86_64.pkg.tar.zst
Offline
6.7-rc1 does exhibit the behavior, while 6.6.r8649.g7d461b291e65 does not.
I'm using the packages loqs provided.
Offline
$ git bisect good
Bisecting: 3769 revisions left to test after this (roughly 12 steps)
[2a80532c0745e140852e6b579bbe8371332bb45d] Merge tag 'printk-for-6.7' of git://git.kernel.org/pub/scm/linux/kernel/git/printk/linux
https://drive.google.com/file/d/1VNsgyq … sp=sharing linux-mainline-6.6.r12433.g2a80532c0745-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/11ieQ6d … sp=sharing linux-mainline-headers-6.6.r12433.g2a80532c0745-1-x86_64.pkg.tar.zst
Offline
6.6.r12433.g2a80532c0745 works fine.
Offline
$ git bisect good
Bisecting: 1889 revisions left to test after this (roughly 11 steps)
[e70703890b2586bc3567365d391c260d23fb7a94] Merge tag 'topic/nvidia-gsp-2023-11-03' of git://anongit.freedesktop.org/drm/drm
https://drive.google.com/file/d/1XBSrpy … sp=sharing linux-mainline-6.6.r14313.ge70703890b25-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/1Tx-Rsv … sp=sharing linux-mainline-headers-6.6.r14313.ge70703890b25-1-x86_64.pkg.tar.zst
Offline
6.6.r14313.ge70703890b25 works fine.
Offline
$ git bisect good
Bisecting: 936 revisions left to test after this (roughly 10 steps)
[b8dd631fcabe2656c8d3751ad4836131d51fb63b] Merge tag 'rproc-v6.7' of git://git.kernel.org/pub/scm/linux/kernel/git/remoteproc/linux
https://drive.google.com/file/d/1lGFW4w … sp=sharing linux-mainline-6.6.r15266.gb8dd631fcabe-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/1tPbdzy … sp=sharing linux-mainline-headers-6.6.r15266.gb8dd631fcabe-1-x86_64.pkg.tar.zst
Offline
6.6.r15266.gb8dd631fcabe works fine.
Offline
$ git bisect good
Bisecting: 444 revisions left to test after this (roughly 9 steps)
[a12deb44f9734dc25970c266249b272e44d3d1b5] Merge tag 'input-for-v6.7-rc0' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
https://drive.google.com/file/d/10nEYV8 … sp=sharing linux-mainline-6.6.r15758.ga12deb44f973-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/1kKQgKn … sp=sharing linux-mainline-headers-6.6.r15758.ga12deb44f973-1-x86_64.pkg.tar.zst
Offline
6.6.r15758.ga12deb44f973 doesn't work.
Offline
$ git bisect bad
Bisecting: 263 revisions left to test after this (roughly 8 steps)
[6bc986ab839c844e78a2333a02e55f02c9e57935] Merge tag 'nfs-for-6.7-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs
https://drive.google.com/file/d/1fTOvJn … sp=sharing linux-mainline-6.6.r15494.g6bc986ab839c-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/12OkdXo … sp=sharing linux-mainline-headers-6.6.r15494.g6bc986ab839c-1-x86_64.pkg.tar.zst
Offline
6.6.r15494.g6bc986ab839c doesn't work.
Offline
$ git bisect bad
Bisecting: 128 revisions left to test after this (roughly 7 steps)
[305230142ae0637213bf6e04f6d9f10bbcb74af8] Merge tag 'pm-6.7-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
https://drive.google.com/file/d/10fofFu … sp=sharing linux-mainline-6.6.r15365.g305230142ae0-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/1BqyIun … sp=sharing linux-mainline-headers-6.6.r15365.g305230142ae0-1-x86_64.pkg.tar.zst
Offline