[SOLVED] mesa-19.3.4-x86_64-1 - crash of the X11 server on a box with AM3 Athlon x4 605e running -current and latest Plasma5

LuckyCyborg · 02-13-2020, 11:22 PM

I had to rollback to the previous version: mesa-19.3.3-x86_64-1

The reason is that the latest Mesa package crashes the X11 server leaving me in console.

Please note that's used the on-board Radeon HD4250 and the system have 8GB DDR3 1333MHz.

Code:

Current Operating System: Linux darkstar.example.org 5.4.19 #1 SMP Tue Feb 11 13:39:28 CST 2020 x86_64
Kernel command line: auto BOOT_IMAGE=Slackware64 ro root=UUID=9a83ce56-19db-4a1d-839c-5273ec4306c8 resume=/dev/sdb3 scsi_mod.use_blk_mq=1 zswap.enabled=1 zswap.compressor=lzo zswap.max_pool_percent=20 zswap.zpool=z3fold rootflags=defaults,noatime,discard,commit=60
Build Date: 14 January 2020  12:43:20PM

Current version of pixman: 0.38.4
        Before reporting problems, check http://wiki.x.org
        to make sure that you have the latest version.
Markers: (--) probed, (**) from config file, (==) default setting,
        (++) from command line, (!!) notice, (II) informational,
        (WW) warning, (EE) error, (NI) not implemented, (??) unknown.
(==) Log file: "/var/log/Xorg.0.log", Time: Fri Feb 14 06:51:17 2020
(==) Using system config directory "/usr/share/X11/xorg.conf.d"
(II) [KMS] Kernel modesetting enabled.
(EE)
(EE) Backtrace:
(EE) 0: /usr/libexec/Xorg (OsLookupColor+0x138) [0x58af68]
(EE) 1: /lib64/libpthread.so.0 (funlockfile+0x50) [0x7fe3961556df]
(EE) 2: /usr/lib64/xorg/modules/dri/r600_dri.so (nouveau_drm_screen_create+0x19f830) [0x7fe394063150]
(EE) 3: /usr/lib64/xorg/modules/dri/r600_dri.so (nouveau_drm_screen_create+0x1b14dc) [0x7fe3940867bc]
(EE) 4: /usr/lib64/xorg/modules/dri/r600_dri.so (nouveau_drm_screen_create+0x19d31d) [0x7fe39405e1cd]
(EE) 5: /usr/lib64/xorg/modules/dri/r600_dri.so (nouveau_drm_screen_create+0x19e449) [0x7fe39405fba9]
(EE) unw_get_proc_name failed: no unwind info found [-10]
(EE) 6: /usr/lib64/xorg/modules/dri/r600_dri.so (?+0x0) [0x7fe393aaa450]
(EE) 7: /lib64/ld-linux-x86-64.so.2 (call_init.part.0+0x9a) [0x7fe3966a646a]
(EE) 8: /lib64/ld-linux-x86-64.so.2 (_dl_init+0x79) [0x7fe3966a6569]
(EE) 9: /lib64/ld-linux-x86-64.so.2 (dl_open_worker+0x3e3) [0x7fe3966aa653]
(EE) 10: /lib64/libc.so.6 (_dl_catch_exception+0x5e) [0x7fe3960b087e]
(EE) 11: /lib64/ld-linux-x86-64.so.2 (_dl_open+0xa6) [0x7fe3966a9c76]
(EE) 12: /lib64/libdl.so.2 (dlopen_doit+0x58) [0x7fe396167248]
(EE) 13: /lib64/libc.so.6 (_dl_catch_exception+0x5e) [0x7fe3960b087e]
(EE) 14: /lib64/libc.so.6 (_dl_catch_error+0x1f) [0x7fe3960b08ef]
(EE) 15: /lib64/libdl.so.2 (_dlerror_run+0x65) [0x7fe396167965]
(EE) 16: /lib64/libdl.so.2 (dlopen+0x36) [0x7fe3961672c6]
(EE) 17: /usr/lib64/libgbm.so.1 (gbm_format_get_name+0x268b) [0x7fe3952ce8cb]
(EE) 18: /usr/lib64/libgbm.so.1 (gbm_format_get_name+0x15e5) [0x7fe3952cc895]
(EE) 19: /usr/lib64/libgbm.so.1 (gbm_format_get_name+0x1a68) [0x7fe3952cd058]
(EE) 20: /usr/lib64/libgbm.so.1 (gbm_create_device+0x44) [0x7fe3952c9624]
(EE) unw_get_proc_name failed: no unwind info found [-10]
(EE) 21: /usr/lib64/xorg/modules/drivers/radeon_drv.so (?+0x0) [0x7fe3953320a0]
(EE) unw_get_proc_name failed: no unwind info found [-10]
(EE) 22: /usr/lib64/xorg/modules/drivers/radeon_drv.so (?+0x0) [0x7fe395321b40]
(EE) 23: /usr/libexec/Xorg (InitOutput+0xa14) [0x47e304]
(EE) 24: /usr/libexec/Xorg (InitFonts+0x1c4) [0x442f64]
(EE) 25: /lib64/libc.so.6 (__libc_start_main+0xeb) [0x7fe395f80e5b]
(EE) 26: /usr/libexec/Xorg (_start+0x2a) [0x42deea]
(EE)
(EE) Illegal instruction at address 0x7fe393ec3920
(EE)
Fatal server error:
(EE) Caught signal 4 (Illegal instruction). Server aborting
(EE)
(EE)
Please consult the The X.Org Foundation support
         at http://wiki.x.org
 for help.
(EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
(EE)
(EE) Server terminated with error (1). Closing log file.

Looks like the AVX instructions strikes back...

I say that because in another box everything works perfectly, with an Intel i3-2120 and same system setup.

gmgf · 02-14-2020, 03:26 AM

Same problem, here. with radeon.

bathory · 02-14-2020, 03:53 AM

Same here with nvidia 7300
Had to downgrade mesa in order to use X again

3rensho · 02-14-2020, 06:17 AM

Nividia GTX-970, driver 440.59, Slack64-current, Plasma5-pam and everything is working fine in the x world using the new mesa-19.3.4.

linuxxer · 02-14-2020, 08:55 AM

I am also facing the same problem.

Code:

[    76.646] (II) Loading sub module "glamoregl"
[    76.646] (II) LoadModule: "glamoregl"
[    76.646] (II) Loading /usr/lib64/xorg/modules/libglamoregl.so
[    76.669] (II) Module glamoregl: vendor="X.Org Foundation"
[    76.669] 	compiled for 1.20.7, module version = 1.0.1
[    76.669] 	ABI class: X.Org ANSI C Emulation, version 0.4
[    77.296] (EE) 
[    77.296] (EE) Backtrace:
[    77.297] (EE) 0: /usr/libexec/Xorg (OsLookupColor+0x138) [0x58af68]
[    77.299] (EE) 1: /lib64/libpthread.so.0 (funlockfile+0x50) [0x7f5fb0f266df]
[    77.301] (EE) 2: /usr/lib64/xorg/modules/dri/iris_dri.so (nouveau_drm_screen_create+0x19f830) [0x7f5faed83150]
[    77.303] (EE) 3: /usr/lib64/xorg/modules/dri/iris_dri.so (nouveau_drm_screen_create+0x1b14dc) [0x7f5faeda67bc]
[    77.304] (EE) 4: /usr/lib64/xorg/modules/dri/iris_dri.so (nouveau_drm_screen_create+0x19d31d) [0x7f5faed7e1cd]
[    77.306] (EE) 5: /usr/lib64/xorg/modules/dri/iris_dri.so (nouveau_drm_screen_create+0x19e449) [0x7f5faed7fba9]
[    77.308] (EE) unw_get_proc_name failed: no unwind info found [-10]
[    77.308] (EE) 6: /usr/lib64/xorg/modules/dri/iris_dri.so (?+0x0) [0x7f5fae7ca450]
[    77.311] (EE) 7: /lib64/ld-linux-x86-64.so.2 (call_init.part.0+0x9a) [0x7f5fb145c46a]
[    77.313] (EE) 8: /lib64/ld-linux-x86-64.so.2 (_dl_init+0x79) [0x7f5fb145c569]
[    77.316] (EE) 9: /lib64/ld-linux-x86-64.so.2 (dl_open_worker+0x3e3) [0x7f5fb1460653]
[    77.319] (EE) 10: /lib64/libc.so.6 (_dl_catch_exception+0x5e) [0x7f5fb0e8187e]
[    77.321] (EE) 11: /lib64/ld-linux-x86-64.so.2 (_dl_open+0xa6) [0x7f5fb145fc76]
[    77.324] (EE) 12: /lib64/libdl.so.2 (dlopen_doit+0x58) [0x7f5fb0f38248]
[    77.327] (EE) 13: /lib64/libc.so.6 (_dl_catch_exception+0x5e) [0x7f5fb0e8187e]
[    77.329] (EE) 14: /lib64/libc.so.6 (_dl_catch_error+0x1f) [0x7f5fb0e818ef]
[    77.332] (EE) 15: /lib64/libdl.so.2 (_dlerror_run+0x65) [0x7f5fb0f38965]
[    77.334] (EE) 16: /lib64/libdl.so.2 (dlopen+0x36) [0x7f5fb0f382c6]
[    77.336] (EE) 17: /usr/lib64/libgbm.so.1 (gbm_format_get_name+0x268b) [0x7f5fb040a8cb]
[    77.338] (EE) 18: /usr/lib64/libgbm.so.1 (gbm_format_get_name+0x15e5) [0x7f5fb0408895]
[    77.340] (EE) 19: /usr/lib64/libgbm.so.1 (gbm_format_get_name+0x1a68) [0x7f5fb0409058]
[    77.342] (EE) 20: /usr/lib64/libgbm.so.1 (gbm_create_device+0x44) [0x7f5fb0405624]
[    77.344] (EE) 21: /usr/lib64/xorg/modules/libglamoregl.so (glamor_egl_init+0x53) [0x7f5faff021f3]
[    77.346] (EE) unw_get_proc_name failed: no unwind info found [-10]
[    77.346] (EE) 22: /usr/lib64/xorg/modules/drivers/modesetting_drv.so (?+0x0) [0x7f5fb041d780]
[    77.347] (EE) 23: /usr/libexec/Xorg (InitOutput+0xa14) [0x47e304]
[    77.348] (EE) 24: /usr/libexec/Xorg (InitFonts+0x1c4) [0x442f64]
[    77.351] (EE) 25: /lib64/libc.so.6 (__libc_start_main+0xeb) [0x7f5fb0d51e5b]
[    77.351] (EE) 26: /usr/libexec/Xorg (_start+0x2a) [0x42deea]
[    77.352] (EE) 
[    77.352] (EE) Illegal instruction at address 0x7f5faebe3920
[    77.352] (EE) 
Fatal server error:
[    77.352] (EE) Caught signal 4 (Illegal instruction). Server aborting

Gordie · 02-14-2020, 09:05 AM

I just rolled back mesa to recover from the same problem. My old laptop has first gen i7 and current64 with plasma5.
NO nvidia

Chuck56 · 02-14-2020, 09:08 AM

X11/KDE is up and running fine also.

AMD A8-7600 Radeon R7, ASUS A88XM-E motherboard, amdgpu,radeon UDEV loaded modules, 64-current, ktown, no PAM (yet)

volkerdi · 02-14-2020, 12:30 PM

Quote:

Originally Posted by LuckyCyborg

I had to rollback to the previous version: mesa-19.3.3-x86_64-1

The reason is that the latest Mesa package crashes the X11 server leaving me in console.

Works on my i7 with Intel graphics, but crashes on the Phenom II with Radeon HD 4200. Same with mesa-20.0.0-rc3. Guess I better bisect it.

ZhaoLin1457 · 02-14-2020, 12:51 PM

Quote:

Originally Posted by volkerdi

Works on my i7 with Intel graphics, but crashes on the Phenom II with Radeon HD 4200. Same with mesa-20.0.0-rc3. Guess I better bisect it.

If it helps, I can confirm the X server crash with the processor: Athlon x2 7850 (AM2+ socket), memory: 4GB DDR2 800 and discrete video card: Radeon HD4350, which uses the "radeon" driver.

Please note that suspecting too a build with too high CPU instructions set, I did a (local) build of mesa package, but even the package built on my own system ended crashing the X server, like the original and I had to return to the previous official package: 19.3.3

volkerdi · 02-14-2020, 01:10 PM

And the winner is:

Code:

4656d2d30a8c176d81f31827302cd082eff59de3 is the first bad commit
commit 4656d2d30a8c176d81f31827302cd082eff59de3
Author: Vinson Lee <vlee@freedesktop.org>
Date:   Sat Feb 1 01:12:32 2020 -0800

    swr: Fix GCC 4.9 checks.

    Fixes: f0a22956be48 ("swr/rast: _mm*_undefined_* implementations for gcc<4.9")
    Fixes: e21fc2c62527 ("swr/rast: non-regex knob fallback code for gcc < 4.9")
    Signed-off-by: Vinson Lee <vlee@freedesktop.org>
    Reviewed-by: Jan Zielinski <jan.zielinski@intel.com>
    (cherry picked from commit deb2bbf57ec1d0660dd85b7080bf5ebeb10e8768)

 .pick_status.json                                                  | 2 +-
 src/gallium/drivers/swr/rasterizer/codegen/templates/gen_knobs.cpp | 2 +-
 src/gallium/drivers/swr/rasterizer/common/os.h                     | 2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)

I have verified that reverting this patch fixes mesa 19.3.4.

tramtrist · 02-14-2020, 03:59 PM

Thank you Pat

linuxxer · 02-15-2020, 01:12 AM

Quote:

Originally Posted by volkerdi

I have verified that reverting this patch fixes mesa 19.3.4.

Thank you Pat this update solve my issue.

gmgf · 02-15-2020, 01:33 AM

Thank, Pat, solved here also.

LuckyCyborg · 02-15-2020, 03:13 AM

Yes, Mr. Volkerding found the culprit and the patched package works perfectly here. Thank you!

ZhaoLin1457 · 02-15-2020, 04:51 AM

Solved also here, thanks!