Kernel 4.9-rc1 fails on shutdown/reboot with a panic. Bisect pointed me to 00ea8cba5ef7b783f11cb1a0b900b7c18d2ce0b6 is the first bad commit commit 00ea8cba5ef7b783f11cb1a0b900b7c18d2ce0b6 Author: Alex Deucher <alexander.deucher@amd.com> Date: Thu Sep 22 14:40:29 2016 -0400 drm/amdgpu: always apply pci shutdown callbacks (v2) We can't properly detect all hypervisors and we need this to properly tear down the hardware. v2: trivial warning fix Signed-off-by: Alex Deucher <alexander.deucher@amd.com> :040000 040000 868a00ff980f95dac82b908df0737342bf40cdc3 02bf2cf4e8797d37d454301afedf2f97735b60ca M drivers After removing this commit, everything works again. System is a Carrizo.
Can you attach the panic messages?
If you can't get them during a reboot/shutdown, you can reproduce it my unloading the module as they use the same code. E.g., modprobe -r amdgpu
I only have a black screen when the panic happens. Rebuild the kernel with amdgpu as module for the unloading module option, but this way I can't remove the driver cause the module is in use.
rmmod -f amdgpu should do the trick
(In reply to Mike Lothian from comment #4) > rmmod -f amdgpu should do the trick This way I can remove the module, but I don't get the panic. The screen goes black and I have to reboot via SysReq Keys.
You can also unbind the console before unloading the module. sudo bash -c "echo 0 > /sys/class/vtconsole/vtcon1/bind" sudo modprobe -r amdgpu
(In reply to Alex Deucher from comment #6) > You can also unbind the console before unloading the module. > > sudo bash -c "echo 0 > /sys/class/vtconsole/vtcon1/bind" First time this resulted in a black screen with a panic, next time the system hung, but reboot via SysReq Keys worked. Always happened on the unbind.
Created attachment 127488 [details] dmesg output after unloading amdgpu.ko I've retried removing the module via ssh session and the second time I was successful. The dmesg output is attached.
Is this fixed somewhere now also? Related to https://bugs.freedesktop.org/show_bug.cgi?id=98200 I guess?
(In reply to Ernst Sjöstrand from comment #9) > Is this fixed somewhere now also? still happens with 4.9-rc2 as well as 4.10-wip
As the dmesg shows a problem in the ACP code, I've just disabled it now and it works. Don't know what's the real problem, but without ACP my system reboot fine.
Created attachment 127736 [details] [review] possible fix This should fix it.
(In reply to Alex Deucher from comment #12) > Created attachment 127736 [details] [review] [review] > possible fix > > This should fix it. Yes, this fixes the panic on reboot/unloading the driver.
-- GitLab Migration Automatic Message -- This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity. You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/drm/amd/issues/106.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.