[Bug 107913] New: amdgpu_drv.so segfaults on boot; black screen

[b***@freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=107913

Bug ID: 107913
Summary: amdgpu_drv.so segfaults on boot; black screen
Product: xorg
Version: unspecified
Hardware: x86-64 (AMD64)
OS: Linux (All)
Status: NEW
Severity: normal
Priority: medium
Component: Driver/AMDgpu
Assignee: xorg-driver-***@lists.x.org
Reporter: ***@kai.giebeler.de
QA Contact: xorg-***@lists.x.org

Created attachment 141539
--> https://bugs.freedesktop.org/attachment.cgi?id=141539&action=edit
Xorg.0.log including the segfoult

History (maybe only partly related to the problem):
I followed this guide to update my drivers:
https://github.com/ValveSoftware/Proton/blob/proton_3.7/PREREQS.md

sudo add-apt-repository ppa:paulo-miguel-dias/mesa
sudo apt dist-upgrade
sudo apt install mesa-vulkan-drivers mesa-vulkan-drivers:i386

which worked fine for some days. I updated my system as normal (via apt) and
after a reboot the machine came up with a black screen.

I found a segfault in my Xorg.0.log (attached). Blacklisting the module
"amdgpu" lets me boot into the desktop using only the integrated Intel GPU.

Some information that are hopefully helpful:

$ lspci -k
00:02.0 VGA compatible controller: Intel Corporation HD Graphics 620 (rev 02)
Subsystem: Dell HD Graphics 620
Kernel driver in use: i915
Kernel modules: i915
01:00.0 Display controller: Advanced Micro Devices, Inc. [AMD/ATI] Topaz XT
[Radeon R7 M260/M265 / M340/M360 / M440/M445] (rev c3)
Subsystem: Dell Radeon R7 M445
Kernel modules: amdgpu

$ uname -a
Linux notebook-kai 4.15.0-34-generic #37-Ubuntu SMP Mon Aug 27 15:21:48 UTC
2018 x86_64 x86_64 x86_64 GNU/Linux

My PC is a notebook with an integrated Intel GPU and a dedicated AMD GPU.
OS is Linux Mint 19 Cinnamon
CPU is Intel© Core™ i7-7500U CPU @ 2.70GHz × 2

What else is needed?
Can I somehow re-install or downgrade the amdgpu-drivers?

[b***@freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=107913

Alex Deucher <***@gmail.com> changed:

What |Removed |Added
----------------------------------------------------------------------------
Attachment #141539|text/x-log |text/plain
mime type| |

[b***@freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=107913

--- Comment #1 from Michel DÀnzer <***@daenzer.net> ---
Please also attach the corresponding dmesg output.

Please install the xserver-xorg-video-amdgpu-dbg package, attach a new Xorg log
file from running with that installed, and either get a backtrace with gdb (per
https://www.x.org/wiki/Development/Documentation/ServerDebugging/), or provide
the output of the following commands:

addr2line -e /usr/lib/xorg/modules/drivers/amdgpu_drv.so
0x7f9e326d3000+0x17305
addr2line -e /usr/lib/xorg/modules/drivers/amdgpu_drv.so 0x7f9e326d3000+0xedfd

For the time being, assuming there are no display connectors on the AMD GPU
(that you need to use), you can avoid the problem by putting this in
/etc/X11/xorg.conf:

Section "ServerFlags"
Option "AutoAddGPU" "off"
EndSection

[b***@freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=107913

--- Comment #2 from Michel DÀnzer <***@daenzer.net> ---
I was able to reproduce the crash (it only happens when the kernel driver
doesn't expose any display outputs),
https://patchwork.freedesktop.org/patch/249254/ fixes it.

[b***@freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=107913

Michel DÀnzer <***@daenzer.net> changed:

What |Removed |Added
----------------------------------------------------------------------------
Resolution|--- |FIXED
Status|NEW |RESOLVED

--- Comment #3 from Michel DÀnzer <***@daenzer.net> ---
Thanks for the report, fixed in Git master:

commit 6572be49b713a26eca14f16e1854cabf28101288
Author: Michel DÀnzer <***@amd.com>
Date: Thu Sep 13 11:44:21 2018 +0200

Bail from drmmode_cm_init if there's no CRTC

[b***@freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=107913

--- Comment #4 from ***@kai.giebeler.de ---
Thanks a lot for the fix, that was pretty quick!

Is any of the requested information still relevant? Maybe for documentation
purpose?

[b***@freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=107913

--- Comment #5 from Michel DÀnzer <***@daenzer.net> ---
(In reply to freedesktop from comment #4)

Post by b***@freedesktop.org
Is any of the requested information still relevant? Maybe for documentation
purpose?

Thanks for asking, but it's not needed anymore.