| View previous topic :: View next topic |
| Author |
Message |
n05ph3r42
Tux's lil' helper
Joined: 11 Jul 2016
Posts: 135
|
 Posted: Wed Jun 15, 2022 9:38 pm Post subject: Xorg libc segfault. |
 |
|
HIYA.
Im using amd64 stable.
Also amdgpu which requires LLVM (RX580 polaris10).
After recent update (@Sun12.06.22) encountered troubles with rust-1.60 and llvm-14.0.4 stack (compiler-rt-sanitizers AFAIK).
Noticed some mess with llvm slots. (14 and 13 mixed up resulting lld-14 used with parts of llvm-13 stack)
Also linux-fiirmware was updated to latest.
After this firefox segfaulted.
I rebooted and journey begins.
Black screen when display-manager starts.
X halted hard, kbd not working (no ttys via alt+%num%, numlock not working, etc).
But sometimes it seems to shutdown properly via ACPI poweroff hw button (i can see shudtown log by OpenRC).
First, i decided switch to safe CFLAGS.
(Before i was using lto + fipa-pta and other experimental\exotic flags, but it was OK for a long time until this happen;+portage.env for few pkg's which requiring safe flags (mostly w/o lto) )
Switched to safe CPU specific flags from
| Code: | | gcc -### -march=native -x c - |
+ few other safe flags:
| Code: | -O2 -pipe \
-fstack-clash-protection -fstack-protector-strong --param ssp-buffer-size=32 \
-fomit-frame-pointer \
-fasynchronous-unwind-tables \
-feliminate-unused-debug-types -fallocation-dce \
-ftree-vectorize \ |
This helped not much, because latest stable binutils-2.34 is not emerged succesfully even with those safe cflags.
Same as rust-1.60 - still failing, and LLVM stack 13 messsed with 14 completely (more components of toolchain become broken)
Then i tried to emerge @system from chroot with safeflags.
stable binutils not emerged. changed accept_keywords for binutils to ~amd64. emerged ok.
emerged gcc-11.3 with safe flags ok.
@system emerged ok.
re emerged @world with safe cflags. looks like LLVM-14.0.4 stack emerged ok, LLVM 13 with 14 at slots.
rebooted
nothing changed
disabled display-manager service
re-emerged sddm and enabled haveged service at boot to increase entropy to exclude sddm failure
this doesnt helped much (nothing happened)
noticed some of lib.c.so segfaults in dmesg. forgot about it.
noticed messages at bugtrack about brocken firmware.
deleted linux firmware from rootfs,masked >=sys-kernel/linux-firmware-22020310
re emerged firmware-20220209
recompiled kernel and mesa
nothig helped.
re-emerged x11-server + emerge @x11-module-rebuild
still nothing
tried startx from non-root user.
noticed again lib.c.so segfault and glibc error during startx
switched to ~amd64 for glibc. accept_keywords ~amd64 for =glibc-2.35-r4
re-emerged glibc
nothing
noticed news about switching to python3_10.
switched safe way - ok.
nothing.
masked all the LLVM-14.0.4 stack (clang, sanitizers, gold, common etc all of them) and removed it.
checked that there is no LLVM-14.0.4 components left, only LLVM-13 via equery.
re-emerged all deps,
re-emerged mesa, Xorg, x11-modules,
nothing helps.
Now i dont know what to do.
startx log is here http://0x0.st/ouuQ.log
emerge --info is here.http://0x0.st/ouu1.bin
(i swiched back to my tested "unsafe" flags, because safe flags did nothing in this case)
the only way i see now is to rebuild everything from scratch on new disk (anyway i palnned to migrate to bigger nvme disk soon)
Any suggestions?
Last edited by n05ph3r42 on Wed Jun 15, 2022 9:49 pm; edited 1 time in total |
|
| Back to top |
|
 |
n05ph3r42
Tux's lil' helper
Joined: 11 Jul 2016
Posts: 135
|
| Posted: Wed Jun 15, 2022 9:44 pm Post subject: Re: Xorg libc segfault. |
|
|
beware its http-only
|
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Desktop Environments
