Last few days we started to see this SEGV during login: [ 35.257] (EE) [ 35.257] (EE) Backtrace: [ 35.257] (EE) 0: /usr/bin/X (xorg_backtrace+0x5b) [0x56145d5b41eb] [ 35.257] (EE) 1: /usr/bin/X (0x56145d47a000+0x13de55) [0x56145d5b7e55] [ 35.257] (EE) 2: /lib64/libc.so.6 (0x7fa38481e000+0x3d770) [0x7fa38485b770] [ 35.257] (EE) 3: /usr/bin/X (DRI2Authenticate+0x28) [0x56145d632d78] [ 35.257] (EE) 4: /usr/bin/X (0x56145d47a000+0x1b9ae2) [0x56145d633ae2] [ 35.257] (EE) 5: /usr/bin/X (0x56145d47a000+0x72c25) [0x56145d4ecc25] [ 35.257] (EE) 6: /usr/bin/X (0x56145d47a000+0x76934) [0x56145d4f0934] [ 35.257] (EE) 7: /lib64/libc.so.6 (0x7fa38481e000+0x2919a) [0x7fa38484719a] [ 35.257] (EE) 8: /lib64/libc.so.6 (__libc_start_main+0x78) [0x7fa384847248] [ 35.257] (EE) 9: /usr/bin/X (_start+0x21) [0x56145d4b4a91] [ 35.257] (EE) [ 35.257] (EE) Segmentation fault at address 0x1 [ 35.257] (EE) on different machines and GPUs, in both MATE and KDE We have not updated anything significant but gcc 9 to 11 Rebuilt several Xorg pkgs:, xorg-serv, all X drivers, libX11 etc. but didn't help. Any clues welcome.
Please enable debug symbols for the x11-base/xorg-server package. You should then be able to obtain a more useful backtrace. Also, always provide emerge --info.
BT from gdb (gdb) bt #0 DRI2Authenticate (client=client@entry=0x5588c132b270, pScreen=0x5588c0700770, magic=1) at /var/tmp/portage/x11-base/xorg-server-21.1.3-r1/xorg-server-21.1.3/hw/xfree86/dri2/dri2.c:1368 #1 0x00005588bfce2ae2 in ProcDRI2Authenticate (client=0x5588c132b270) at /var/tmp/portage/x11-base/xorg-server-21.1.3-r1/xorg-server-21.1.3/hw/xfree86/dri2/dri2ext.c:156 #2 ProcDRI2Dispatch (client=0x5588c132b270) at /var/tmp/portage/x11-base/xorg-server-21.1.3-r1/xorg-server-21.1.3/hw/xfree86/dri2/dri2ext.c:609 #3 0x00005588bfb9bc25 in Dispatch () at /var/tmp/portage/x11-base/xorg-server-21.1.3-r1/xorg-server-21.1.3/dix/dispatch.c:550 #4 0x00005588bfb9f934 in dix_main (argc=14, argv=0x7fffad3bbdb8, envp=<optimized out>) at /var/tmp/portage/x11-base/xorg-server-21.1.3-r1/xorg-server-21.1.3/dix/main.c:272 #5 0x00007fcaaffde19a in __libc_start_call_main (main=main@entry=0x5588bfb63a60 <main>, argc=argc@entry=14, argv=argv@entry=0x7fffad3bbdb8) at ../sysdeps/nptl/libc_start_call_main.h:58 #6 0x00007fcaaffde248 in __libc_start_main_impl (main=0x5588bfb63a60 <main>, argc=14, argv=0x7fffad3bbdb8, init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>, stack_end=0x7fffad3bbda8) at ../csu/libc-start.c:409 #7 0x00005588bfb63a91 in _start ()
Created attachment 775698 [details] emerge --info
print dri2_client $1 = (DRI2ClientPtr) 0x1 1 is not a good ptr value
Also reported upstream: https://gitlab.freedesktop.org/xorg/xserver/-/issues/1340
Seems like I have the same problem here on at least two different machines, the stack trace looks exactly like the one shown in this thread. x11-base/xorg-server-21.1.7 dies with segfault when opening the remote console of a virtual machine connected to via virt-manager. Does not happen on all machines, but when it happens, it happens reproducibly. All machines, affected and non-affected, are built from same portage snapshot. I can reproduce this via: 1. Have a virtual machine on a remote host, managed by libvirt. Graphics interface in VM is set to QXL and spice server 2. On admin-Terminal, open virt-manager and open the console for the VM 3. Crash. Thrown out of X, need to re-login app-emulation/virt-manager-4.0.0 [gtk] net-misc/spice-gtk-0.41-r1 (gdb) bt #0 0x00007fb2f3f195cc in ?? () from /lib64/libc.so.6 #1 0x00007fb2f3ecd0e2 in raise () from /lib64/libc.so.6 #2 0x00007fb2f3eb7473 in abort () from /lib64/libc.so.6 #3 0x000055b0df2500dc in OsAbort () #4 0x000055b0df2557f5 in ?? () #5 0x000055b0df2567c6 in FatalError () #6 0x000055b0df24d5d9 in ?? () #7 <signal handler called> #8 0x000055b0df2c86a8 in DRI2Authenticate () #9 0x000055b0df2c9412 in ?? () #10 0x000055b0df181c55 in ?? () #11 0x000055b0df185964 in ?? () #12 0x00007fb2f3eb820c in ?? () from /lib64/libc.so.6 #13 0x00007fb2f3eb82bc in __libc_start_main () from /lib64/libc.so.6 #14 0x000055b0df149ac1 in _start ()
I think my problem went away with x11-drivers/xf86-video-qxl-0.1.6 Perhaps you need to rebuild this pkg to avoid any ABI issues ?
net-misc/spice-gtk-0.41-r1 is old There is net-misc/spice-gtk-0.42
x11-drivers/xf86-video-qxl is a driver module for Xorg. This is not involved in the problem: The server(VM), to which I want to connect to, is in a text-console mode only installation, Xorg not even installed. libvirt/qemu is emulating a virtual QXL based graphics card. The clients which are crashing, is (1) a native amd-radeon driver and the (2) a virtualbox based desktop system running the vmware driver.
Upgraded spice-gtk to net-misc/spice-gtk-0.42, but same problem. As soon as I open the virtmanager-console for a VM crash.
Created attachment 895283 [details, diff] Patch that fixes the crash I'm getting the same crash (different trigger, but the same stack trace). I've attached a patch based on one on the upstream bug tracker that fixes it for me.
(In reply to Mark Wagner from comment #11) > Created attachment 895283 [details, diff] [details, diff] > Patch that fixes the crash > > I'm getting the same crash (different trigger, but the same stack trace). > I've attached a patch based on one on the upstream bug tracker that fixes it > for me. URL to upstream bug?
https://gitlab.freedesktop.org/xorg/xserver/-/issues/1534
(In reply to Mark Wagner from comment #11) > I'm getting the same crash (different trigger, but the same stack trace). > I've attached a patch based on one on the upstream bug tracker that fixes it > for me. I have an identical back trace with an x11-wm/xpra-6.0.1 client on a machine using the "Kernel Mode Setting driver: i915, version 1.6.0 20230929", on an old HD Graphics 2500. The patch Mark Wagner attached avoids the crash.
