Xorg-Server 1.20.9 crashes with Nvidia 450xx drivers

[stephane]

from testing

• Xorg 1.20.9 + 450xx drivers ( fail on return sleep video) , login ok , xorg failed after
• Xorg 1.20.8 + 450xx drivers ( ok )
• Xorg 1.20.9 + 440xx drivers ( fail on return sleep video) , login ok , xorg failed after

sudo downgrade xorg-server 

Paquets disponibles :

-   1)  xorg-server    1.20.4  1  x86_64  (distant)
-   2)  xorg-server    1.20.5  1  x86_64  (distant)
-   3)  xorg-server    1.20.5  2  x86_64  (distant)
    4)  xorg-server    1.20.5  3  x86_64  (distant)
    5)  xorg-server    1.20.5  4  x86_64  (distant)
-   6)  xorg-server    1.20.6  1  x86_64  (distant)
-   7)  xorg-server    1.20.6  2  x86_64  (distant)
-   8)  xorg-server    1.20.6  3  x86_64  (distant)
-   9)  xorg-server    1.20.7  1  x86_64  (distant)
-  10)  xorg-server    1.20.8  1  x86_64  (distant)
-  11)  xorg-server    1.20.8  2  x86_64  (distant)
-  12)  xorg-server    1.20.8  3  x86_64  (distant)
-  13)  xorg-server    1.20.8  4  x86_64  (local) ( OK)
+  14)  xorg-server    1.20.9  1  x86_64  (distant) 
+  15)  xorg-server    1.20.9  1  x86_64  (local)  (KO )

[intersectRaven]

I’m experiencing the xorg crashes also. It typically occurs after suspend. I also have multi-monitor setup. After rolling back to 1.20.8 all seems stable again. Also have 450xx drivers.

Aug 28 17:18:35 raven-lpc systemd-coredump[726165]: Process 1337 (Xorg) of user 0 dumped core.

                                                Stack trace of thread 1337:
                                                #0  0x00007f0145bff615 raise (libc.so.6 + 0x3d615)
                                                #1  0x00007f0145be8862 abort (libc.so.6 + 0x26862)
                                                #2  0x0000564899c1138a OsAbort (Xorg + 0x14a38a)
                                                #3  0x0000564899c12e51 FatalError (Xorg + 0x14be51)
                                                #4  0x0000564899c18aa9 n/a (Xorg + 0x151aa9)
                                                #5  0x00007f0145bff6a0 __restore_rt (libc.so.6 + 0x3d6a0)
                                                #6  0x0000564899b45f03 n/a (Xorg + 0x7ef03)
                                                #7  0x0000564899b68faf ChangeWindowAttributes (Xorg + 0xa1faf)
                                                #8  0x0000564899b36a58 n/a (Xorg + 0x6fa58)
                                                #9  0x0000564899b01165 n/a (Xorg + 0x3a165)
                                                #10 0x00007f0145bea152 __libc_start_main (libc.so.6 + 0x28152)
                                                #11 0x0000564899b015ae _start (Xorg + 0x3a5ae)
                                                
                                                Stack trace of thread 1344:
                                                #0  0x00007f0145ab76a2 pthread_cond_wait@@GLIBC_2.3.2 (libpthread.so.0 + 0xf6a2)
                                                #1  0x00007f014027990c n/a (iris_dri.so + 0x4ae90c)
                                                #2  0x00007f0140278308 n/a (iris_dri.so + 0x4ad308)
                                                #3  0x00007f0145ab13e9 start_thread (libpthread.so.0 + 0x93e9)
                                                #4  0x00007f0145cc2293 __clone (libc.so.6 + 0x100293)
                                                
                                                Stack trace of thread 1346:
                                                #0  0x00007f0145ab76a2 pthread_cond_wait@@GLIBC_2.3.2 (libpthread.so.0 + 0xf6a2)
                                                #1  0x00007f014027990c n/a (iris_dri.so + 0x4ae90c)
                                                #2  0x00007f0140278308 n/a (iris_dri.so + 0x4ad308)
                                                #3  0x00007f0145ab13e9 start_thread (libpthread.so.0 + 0x93e9)
                                                #4  0x00007f0145cc2293 __clone (libc.so.6 + 0x100293)
                                                
                                                Stack trace of thread 1345:
                                                #0  0x00007f0145ab76a2 pthread_cond_wait@@GLIBC_2.3.2 (libpthread.so.0 + 0xf6a2)
                                                #1  0x00007f014027990c n/a (iris_dri.so + 0x4ae90c)
                                                #2  0x00007f0140278308 n/a (iris_dri.so + 0x4ad308)
                                                #3  0x00007f0145ab13e9 start_thread (libpthread.so.0 + 0x93e9)
                                                #4  0x00007f0145cc2293 __clone (libc.so.6 + 0x100293)
                                                
                                                Stack trace of thread 1347:
                                                #0  0x00007f0145ab76a2 pthread_cond_wait@@GLIBC_2.3.2 (libpthread.so.0 + 0xf6a2)
                                                #1  0x00007f014027990c n/a (iris_dri.so + 0x4ae90c)
                                                #2  0x00007f0140278308 n/a (iris_dri.so + 0x4ad308)
                                                #3  0x00007f0145ab13e9 start_thread (libpthread.so.0 + 0x93e9)
                                                #4  0x00007f0145cc2293 __clone (libc.so.6 + 0x100293)

Aug 28 17:18:35 raven-lpc systemd[1]: systemd-coredump@2-726164-0.service: Succeeded.

[michaldybczak]

I am on testing, have updated xorg, switched to 450xx drivers, then switched to Nvidia mode. All works, I can’t detect crashes or dump cores like described here. However, I tried this in single monitor setup. Maybe when I plug in another monitor, things will change.

[intersectRaven]

That gives me an idea. Maybe PRIME synchronization got broken again by the 1.20.9 update. It got fixed with the 450xx drivers so I never had to turn it off again when I updated to those drivers. Maybe if I disable PRIME sync it will stabilize again. Will test and get back to this.

[intersectRaven]

Nope. After 3 suspends xorg crashed again. Rolling back to 1.20.8 for now.

*Seems this also occurs on arch and someone there has already posted the issue at upstream as issue 1073. Guess we’ll have to hold off on 1.20.9 for those who’re experiencing this issue.

[pheiduck]

I have the same on a AMD only System:

[Testing Update] 2020-08-26 - Firefox 80, CoreCTRL, Xorg-Server, AMDVLK, Python, Haskell

Xorg Server is crashing every day one time. Xorg Logs [ 286.585] X.Org X Server 1.20.9 X Protocol Version 11, Revision 0 [ 286.585] Build Operating System: Linux Manjaro Linux [ 286.585] Current Operating System: Linux Workstation 5.7.17-2-MANJARO #1 SMP PREEMPT Sat Aug 22 14:58:17 UTC 2020 x86_64 [ 286.585] Kernel command line: BOOT_IMAGE=/@/boot/vmlinuz-5.7-x86_64 root=UUID=19a71f22-36d1-49b6-a6f4-90c2b499789e rw rootflags=subvol=@ bootsplash.bootfile=bootsplash-themes/manjaro/boots…

So It should work again fine if I downgrade to 1.20.8?

[intersectRaven]

Yes. Just downgrade xorg-server and xorg-server-common. I’m currently bisecting the xorg-server source to find the commit responsible for this so maybe we can have the Manjaro devs revert it for our Manjaro package while upstream solves it for the next official release.

[pheiduck]

Or simply ignore the packages, for next updates. Until this is fixed.

Bug is reported by Arch Linux:
https://bugs.archlinux.org/task/67774?project=1&string=xorg-server

[stephane]

downgrade only xorg-server , xorg-server-common seems to be ok ( unless fails appears )

[intersectRaven]

Just an update, issue has been cited as duplicate and closed since the commit that messed it has been found and slated for the 1.20.10 release. Maybe my corroboration with regards to that commit confirmed that it essentially was the same issue with a previous one? For now, either wait for 1.20.10 or downgrade temporarily to 1.20.8 for those who updated to 1.20.9.

[stephane]

@philm , check version Xorg 1.20.10

[intersectRaven]

There’s no 1.20.10 yet. It’s just tagged as to be fixed for 1.20.10 since they seem to have a fix already that does not involve reverting the commit. Just an oversight which needs to be addressed in another patch.

[intersectRaven]

As of today, the commit fixing this issue has been merged to master. Now we’ll have to wait for either 1.20.10 or for Arch/Manjaro devs to cherry-pick the commit (919f1f46fc67dae93b2b3f278fcbfc77af34ec58) to our 1.20.9 packages.

[intersectRaven]

Tagging @philm since Arch has updated their xorg with the patch fixing this regression with their 1.20.9-2 release.

[BS86]

its already in unstable …
Installed the update 50 minutes ago.

[cges30901]

xorg-server 1.20.9-2 is already in Testing Branch. Now I no longer have crashes.

[stephane]

it’s ok with 1.20.9-2

[Rick]

Hi @stephane.

I am also having the xorg problem and I am trying to downgrade to anything below 1.20.8-4 (or upgrade to 1.20.9-2), but when I run sudo downgrade xorg-server I only have the option to reinstall the same version I am running now (1.20.8-4).

I have discovered that this is because downgrade cannot see the zst files in the remote servers. However, by your post I can see that somehow you can pull all versions. Can you tell me what is the secret?

Cheers

[pobrn]

Hmmm… downgrade should support zst since v6.2.4 at least, now it’s at v8.1.1. As a temporary solution, you could try manually downloading from the Arch Linux Archive.

[Rick]

Hi @pobrn.

Yeah, it is a little strange why my downgrade cannot see the other xorg-server versions.

Your advice is sound. I will try it and report back.

Thanks!