Bug 1808767

Created attachment 1666701 [details]
00-keyboard.conf (probably irrelevant)

Created attachment 1666702 [details]
Xorg.0.log

Created attachment 1666703 [details]
dmesg (compressed with xz)

The image can be found at https://fedoraproject.org/w/uploads/0/07/F20-alpha-wallpaper-wide.png (https://fedoraproject.org/wiki/F20_Artwork widescreen)

I've looked at the logs and it seems that I was using:
 - xorg-x11-drv-intel-2.99.917-43.20180618.fc31.x86_64
 - xorg-x11-server-Xorg-1.20.6-1.fc31.x86_64
 - mesa-dri-drivers-19.2.8-1.fc31.x86_64
before the upgrade to Fedora 32.

X keeps on crashing more and more often. ristretto-0.10.0-3.fc32.x86_64, pidgin-2.13.0-17.fc32.x86_64 are virt-manager-2.2.1-3.fc32.noarch are causing segfaults too.

I can confirm. Got same issue with same backtrace.

X crashed when using EoG, Firefox, Chrome, Celluloid, mpv, mplayer, Eclipse, Freecad

P.S: Sorry about report issue so briefly. Gnome-abrt always fail to report issue because fail on remote backtrace.

===================================================================
sudo find /sys/module/i915/parameters -type f  -printf '%P: ' -exec cat -A {} ';'
===================================================================
enable_ips: 1$
fastboot: -1$
enable_psr: 1$
huc_firmware_path: (null)$
edp_vswing: 0$
enable_fbc: 1$
disable_power_well: 1$
alpha_support: N$
vbt_firmware: (null)$
enable_hangcheck: Y$
guc_firmware_path: (null)$
reset: 3$
modeset: -1$
lvds_channel_mode: 0$
guc_log_level: -1$
enable_dpcd_backlight: -1$
prefault_disable: N$
vbt_sdvo_panel_type: -1$
enable_dc: 2$
nuclear_pageflip: N$
force_probe: $
panel_use_ssc: -1$
error_capture: Y$
mmio_debug: 0$
verbose_state_checks: Y$
enable_guc: -1$
dmc_firmware_path: (null)$
invert_brightness: 0$
enable_dp_mst: Y$
force_reset_modeset_test: N$
enable_gvt: N$
disable_display: N$
load_detect_test: N$

===================================================================
grep intel ~/.local/share/xorg/Xorg.0.log
===================================================================
[ 28526.711] (II) LoadModule: "intel"
[ 28526.711] (II) Loading /usr/lib64/xorg/modules/drivers/intel_drv.so
[ 28526.711] (II) Module intel: vendor="X.Org Foundation"
[ 28526.712] (II) intel: Driver for Intel(R) Integrated Graphics Chipsets:
[ 28526.712] (II) intel: Driver for Intel(R) HD Graphics
[ 28526.712] (II) intel: Driver for Intel(R) Iris(TM) Graphics
[ 28526.712] (II) intel: Driver for Intel(R) Iris(TM) Pro Graphics
[ 28526.712] (II) intel(0): Using Kernel Mode Setting driver: i915, version 1.6.0 20200114
[ 28526.726] (--) intel(0): Integrated Graphics Chipset: Intel(R) HD Graphics
[ 28526.726] (--) intel(0): CPU: x86-64, sse2, sse3, ssse3, sse4.1, sse4.2, avx, avx2; using a maximum of 6 threads
[ 28526.726] (II) intel(0): Creating default Display subsection in Screen section
[ 28526.727] (==) intel(0): Depth 24, (--) framebuffer bpp 32
[ 28526.727] (==) intel(0): RGB weight 888
[ 28526.727] (==) intel(0): Default visual is TrueColor
[ 28526.727] (**) intel(0): Option "AccelMethod" "sna"
[ 28526.727] (**) intel(0): Option "DRI" "3"
[ 28526.727] (**) intel(0): Option "TearFree" "true"
[ 28526.727] (II) intel(0): Output eDP1 has no monitor section
[ 28526.733] (**) intel(0): Found backlight control interface intel_backlight (type 'raw') for output eDP1
[ 28526.733] (II) intel(0): Enabled output eDP1
[ 28526.733] (II) intel(0): Output DP1 has no monitor section
[ 28526.733] (II) intel(0): Enabled output DP1
[ 28526.733] (II) intel(0): Output DP2 has no monitor section
[ 28526.733] (II) intel(0): Enabled output DP2
[ 28526.733] (II) intel(0): Output DP3 has no monitor section
[ 28526.733] (II) intel(0): Enabled output DP3
[ 28526.733] (--) intel(0): Using a maximum size of 256x256 for hardware cursors
[ 28526.734] (II) intel(0): Output VIRTUAL1 has no monitor section
[ 28526.734] (II) intel(0): Enabled output VIRTUAL1
[ 28526.734] (--) intel(0): Output eDP1 using initial mode 3840x2160 on pipe 0
[ 28526.734] (**) intel(0): TearFree enabled
[ 28526.734] (==) intel(0): Using gamma correction (1.0, 1.0, 1.0)
[ 28526.734] (==) intel(0): DPI set to (96, 96)
[ 28526.746] (II) intel(0): SNA initialized with Coffeelake (gen9) backend
[ 28526.746] (==) intel(0): Backing store enabled
[ 28526.746] (==) intel(0): Silken mouse enabled
[ 28526.746] (II) intel(0): HW Cursor enabled
[ 28526.746] (==) intel(0): DPMS enabled
[ 28526.746] (==) intel(0): Display hotplug detection enabled
[ 28526.746] (II) intel(0): [DRI2] Setup complete
[ 28526.746] (II) intel(0): [DRI2]   DRI driver: i965
[ 28526.746] (II) intel(0): [DRI2]   VDPAU driver: va_gl
[ 28526.747] (II) intel(0): direct rendering: DRI2 DRI3 enabled
[ 28526.747] (II) intel(0): hardware support for Present enabled
[ 28526.797] (II) intel(0): switch to mode 3840x2160 on eDP1 using pipe 0, position (0, 0), rotation normal, reflection none
[ 28526.800] (II) intel(0): Setting screen physical size to 1016 x 571
[ 28527.225] (II) intel(0): EDID vendor "SHP", prod id 5261
[ 28527.225] (II) intel(0): Printing DDC gathered Modelines:
[ 28527.225] (II) intel(0): Modeline "3840x2160"x0.0  533.25  3840 3888 3920 4000  2160 2163 2168 2222 -hsync -vsync (133.3 kHz eP)

Intel(R) Core(TM) i7-8850H CPU @ 2.60GHz

00:02.0 VGA compatible controller: Intel Corporation UHD Graphics 630 (Mobile)

with i915.enable_guc= 1 or 2 change nothing. Still crash.

I've uninstalled xorg-x11-drv-intel in the mean time and, so far I'm not missing it. XFCE, Firefox, eog, ristretto, mpv and other programs work fine.

(In reply to Cristian Ciupitu from comment #10)
> I've uninstalled xorg-x11-drv-intel in the mean time and, so far I'm not
> missing it. XFCE, Firefox, eog, ristretto, mpv and other programs work fine.

Disable SNA also workaround this issue. But also poor performance too.
Latest 5.5 kernel from FC31 also not works.
At the same time I'm trying different firmware with i915 backport and xorg driver combination.
Still bad luck so far.

According test result I already have right now.
I don't think problem comes from kernel module or firmware.
It's time to dig into git. Q_Q

P.S: here is latest backtrace from xorg log:
[  2109.966] (EE) Backtrace:
[  2109.968] (EE) 0: /usr/libexec/Xorg (OsLookupColor+0x139) [0x5577997a2829]
[  2109.968] (EE) 1: /usr/lib64/libpthread.so.0 (funlockfile+0x60) [0x7f4cb312daef]
[  2109.969] (EE) unw_get_proc_name failed: no unwind info found [-10]
[  2109.969] (EE) 2: /usr/lib64/xorg/modules/drivers/intel_drv.so (?+0x0) [0x7f4cb265b990]
[  2109.970] (EE) unw_get_proc_name failed: no unwind info found [-10]
[  2109.970] (EE) 3: /usr/lib64/xorg/modules/drivers/intel_drv.so (?+0x0) [0x7f4cb268d6f0]
[  2109.970] (EE) unw_get_proc_name failed: no unwind info found [-10]
[  2109.970] (EE) 4: /usr/lib64/xorg/modules/drivers/intel_drv.so (?+0x0) [0x7f4cb26ad7f0]
[  2109.971] (EE) 5: /usr/libexec/Xorg (BlockHandler+0x40) [0x557799642480]
[  2109.971] (EE) 6: /usr/libexec/Xorg (WaitForSomething+0x124) [0x55779979be04]
[  2109.972] (EE) 7: /usr/libexec/Xorg (SendErrorToClient+0x117) [0x55779963d7e7]
[  2109.972] (EE) 8: /usr/libexec/Xorg (InitFonts+0x3ac) [0x557799641b2c]
[  2109.973] (EE) 9: /usr/lib64/libc.so.6 (__libc_start_main+0xf2) [0x7f4cb2f76042]
[  2109.973] (EE) 10: /usr/libexec/Xorg (_start+0x2e) [0x55779962ae4e]

I'm on Fedora 32 (not 31) with the latest kernel, mesa etc, i.e.
kernel-5.6.0-0.rc5.git0.2.fc32.x86_64
mesa-dri-drivers-20.0.1-1.fc32.x86_64
xorg-x11-server-common-1.20.7-1.fc32.x86_64

(In reply to Cristian Ciupitu from comment #12)
> I'm on Fedora 32 (not 31) with the latest kernel, mesa etc, i.e.
> kernel-5.6.0-0.rc5.git0.2.fc32.x86_64
> mesa-dri-drivers-20.0.1-1.fc32.x86_64
> xorg-x11-server-common-1.20.7-1.fc32.x86_64

I know. Me too. I force install FC31 kernel and compile upstream kernel by my self just for verification.

Because on my laptop I also got iwlwifi unable to load firmware issue.

iwlwifi:9260: missing PCI subdevice entries for 0x2526 (0x04010, 0x4018 and 0x401C)(https://bugzilla.kernel.org/show_bug.cgi?id=206661)

Could we change subject to
"Intel i915 family GPU with SNA 2D acceleration, X crashes when running some programs"
or less confusion wording ?

Same here. Unfortunately I cannot uninstall xorg-x11-drv-intel, because without it:
[  3476.926] Require OpenGL version 2.1 or later.
[  3476.926] (EE) modeset(0): Failed to initialize glamor at ScreenInit() time.
[  3476.926] (EE)
Fatal server error:
[  3476.926] (EE) AddScreen/ScreenInit failed for driver 0

*** Bug 1820815 has been marked as a duplicate of this bug. ***

Bug 1818972 and bug 1808767, which are likely duplicates, work around this issue by downgrading to xorg-x11-drv-intel-2.99.917-43.20180618.fc31 from https://koji.fedoraproject.org/koji/buildinfo?buildID=1343048.

Created attachment 1677088 [details]
Fix double-free crashes

The attached patch seems to fix the crashes for me on Intel GMA 3150. It applies to the git head or the snapshot in the current rpm. Please test while I post it to intel-gfx for review.

Created attachment 1677568 [details]
log with full debug

Useful for xf86-video-intel maintainers.

FEDORA-2020-81d6a5b194 has been submitted as an update to Fedora 32. https://bodhi.fedoraproject.org/updates/FEDORA-2020-81d6a5b194

Build 45.20200205 fixes the segfaults for me but the assertion, which is probably related, is still there. I would not be surprise if the issue surfaces again with another gcc version.

Proposed as a Freeze Exception for 32-final by Fedora user frantisekz using the blocker tracking app because:

 FEDORA-2020-81d6a5b194 is needed to fix booting into Xorg sessions withou nomodeset parameter on Intel Gen <= 4 GPUs.

It wouldn't be ideal to fix this as post release update, change seems to pose low risk as it is just a rebuild and affects only old generations of hardware. New intel GPU generations use generic modesetting driver.

FEDORA-2020-81d6a5b194 has been pushed to the Fedora 32 testing repository.
In short time you'll be able to install the update with the following command:
`sudo dnf upgrade --enablerepo=updates-testing --advisory=FEDORA-2020-81d6a5b194`
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2020-81d6a5b194

See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates.

+1 FE for me. Note this can also affect systems with cards that wouldn't usually use intel but which have it specified in a config file. They might hit this on upgrade, so that's another reason for it to go stable.

Discussed during the 2020-04-13 blocker review meeting: [0]

The decision to classify this bug as an "AcceptedFreezeException" was made as it is a noticeable issue that cannot be fixed with an update.

[0] https://meetbot.fedoraproject.org/fedora-blocker-review/2020-04-13/f32-blocker-review.2020-04-13-16.04.txt

FEDORA-2020-81d6a5b194 has been pushed to the Fedora 32 stable repository.
If problem still persists, please make note of it in this bug report.

As we had multiple indications in the update feedback that the update fixed the bug, closing this.

The needinfo request[s] on this closed bug have been removed as they have been unresolved for 1000 days