X segfaults on start - possibly libpthread/glibc related.

skip · 2011-02-21 22:16:18

I restarted my laptop this evening and when I ran startx X segfaulted straight away.

This is a Thinkpad R500 with a Radeon GPU.

I noticed that the log mentions libpthread. That is owned by glibc and I remember that glibc was upgraded on the 20th Feb. I tried to downgrade to a previous release but the issue persisted. Everything else seems to be running fine. What can I do here?

I managed to redirect X startup message to a log file and transfer it to my old wintel computer (where I am know). Is there any other output you need for diagnostics?

X.Org X Server 1.9.4
Release Date: 2011-02-04
X Protocol Version 11, Revision 0
Build Operating System: Linux 2.6.37-ARCH x86_64 
Current Operating System: Linux ages 2.6.30-ARCH #1 SMP PREEMPT Mon Aug 17 16:06:45 CEST 2009 x86_64
Kernel command line: root=/dev/disk/by-uuid/302159b4-f56d-4f44-a121-4cfe4a8462cd resume=/dev/sda2 ro
Build Date: 04 February 2011  09:38:18PM
 
Current version of pixman: 0.20.2
    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: Mon Feb 21 23:38:44 2011
(==) Using config directory: "/etc/X11/xorg.conf.d"
(EE) Failed to load module "fbdev" (module does not exist, 0)
(II) [KMS] drm report modesetting isn't supported.
RADEONDPEncoderService: 1 19
  XRANDR name: DVI-0
  Connector: DVI-I
  DFP1: INTERNAL_UNIPHY
  DDC reg: 0x7e60
  XRANDR name: LVDS
  Connector: LVDS
  LCD1: INTERNAL_KLDSCP_LVTMA
  DDC reg: 0x7e40
  XRANDR name: DisplayPort-0
  Connector: DisplayPort
  DFP2: INTERNAL_UNIPHY
  DDC reg: 0x7e20
  XRANDR name: VGA-0
  Connector: VGA
  CRT1: INTERNAL_KLDSCP_DAC1
  DDC reg: 0x7e50
Unhandled monitor type 0
finished output detect: 0
finished output detect: 1
RADEONDPEncoderService: 1 19
failed to get auxch 0030 40 30 02 after 0 retries
failed to get auxch 0050 40 30 02 after 0 retries
failed to get auxch 0050 40 30 02 after 0 retries
failed to get auxch 0050 40 30 02 after 0 retries
failed to get auxch 0050 40 30 02 after 0 retries
finished output detect: 2
Dac detection success
finished output detect: 3
finished all detect
Unhandled monitor type 0
RADEONDPEncoderService: 1 19
failed to get auxch 0030 40 30 02 after 0 retries
failed to get auxch 0050 40 30 02 after 0 retries
failed to get auxch 0050 40 30 02 after 0 retries
failed to get auxch 0050 40 30 02 after 0 retries
failed to get auxch 0050 40 30 02 after 0 retries
Dac detection success
Output DIG0 transmitter setup success
Blank CRTC 0 success
Disable CRTC memreq 0 success
Disable CRTC 0 success
Blank CRTC 1 success
Disable CRTC memreq 1 success
Disable CRTC 1 success
mc fb loc is 00f700f0
Output DIG0 transmitter setup success
Output DIG0 transmitter setup success
Output CRT1 disable success
Blank CRTC 0 success
Disable CRTC memreq 0 success
Disable CRTC 0 success
Blank CRTC 1 success
Disable CRTC memreq 1 success
Disable CRTC 1 success
Output DIG0 transmitter setup success
Blank CRTC 0 success
Disable CRTC memreq 0 success
Disable CRTC 0 success
Set CRTC 0 Source success
Mode 1680x1050 - 1904 1066 10
Picked PLL 0
before 12200
after 12200
best_freq: 122000
best_feedback_div: 122
best_frac_feedback_div: 0
best_ref_div: 3
best_post_div: 9
Set CRTC 0 PLL success
Set CRTC Timing success
Set CRTC 0 Overscan success
Not using RMX
scaler 0 setup success
Set CRTC 0 Source success
crtc 0 YUV disable setup success
Output DIG0 transmitter setup success
Output DIG1 encoder setup success
Output DIG1 encoder setup success
Output DIG0 transmitter setup success
Output DIG0 transmitter setup success
Output DIG0 transmitter setup success
Output DIG0 transmitter setup success
Enable CRTC 0 success
Enable CRTC memreq 0 success
Unblank CRTC 0 success
Output DIG0 transmitter setup success
Output CRT1 disable success
Blank CRTC 1 success
Disable CRTC memreq 1 success
Disable CRTC 1 success

Backtrace:
0: /usr/bin/X (xorg_backtrace+0x28) [0x49f718]
1: /usr/bin/X (0x400000+0x60349) [0x460349]
2: /lib/libpthread.so.0 (0x7fa310632000+0xf150) [0x7fa310641150]
Segmentation fault at address (nil)

Fatal server error:
Caught signal 11 (Segmentation fault). Server aborting


Please consult the The X.Org Foundation support 
     at http://wiki.x.org
 for help. 
Please also check the log file at "/var/log/Xorg.0.log" for additional information.

Output DIG0 transmitter setup success
Blank CRTC 0 success
Disable CRTC memreq 0 success
Disable CRTC 0 success
Blank CRTC 1 success
Disable CRTC memreq 1 success
Disable CRTC 1 success
Enable CRTC 0 success
Enable CRTC memreq 0 success
Unblank CRTC 0 success
xinit: giving up
xinit: unable to connect to X server: Connection refused
xinit: unexpected signal 2

falconindy · 2011-02-21 22:24:55

What stands out isn't the backtrace, but this:

Current Operating System: Linux ages 2.6.30-ARCH #1 SMP PREEMPT Mon Aug 17 16:06:45 CEST 2009 x86_64

Why are you running a 2 year old kernel?

skip · 2011-02-21 22:32:58

It's stable and it works so I put kernel26 on my pacman block list.

Also I'm not too clued up on Arch's kernel upgrade procedures. That was the stock kernel and my wifi worked out the box at install time so I blocked it. I don't like compiling kernels myself so having the stock install kernel work was good enough for me...

Would that be the issue though? Everything else seems to work.

karol · 2011-02-21 22:42:34

skip wrote:

It's stable and it works so I put kernel26 on my pacman block list.
Also I'm not too clued up on Arch's kernel upgrade procedures. That was the stock kernel and my wifi worked out the box at install time so I blocked it. I don't like compiling kernels myself so having the stock install kernel work was good enough for me...
Would that be the issue though? Everything else seems to work.

I don't know anything about the wifi drivers but Arch is a binary distro and you install the kernel packages just like any other - they are already compiled.

skip · 2011-02-21 23:04:07

karolI wrote:

don't know anything about the wifi drivers but Arch is a binary distro and you install the kernel packages just like any other - they are already compiled.

I know that but I didn't want to have to do a custom compile if things didn't work.

Back to the issue at hand though. If I downgrade glibc do I have to restart before I run startx again? I have downgraded from 2.13-4 to 2.13-1 and it sill segfaults when I run startx afterwards.

skip · 2011-02-22 00:54:01

OK I've made some progress.

I generated a new basic Xorg.conf and set it up to use the 'vesa' driver instead of 'radeon'.

Now X starts but I have to use twm because my old WM (i3) was using some fonts that could no longer be found.

Could the issue be the radeon driver then? Would it cause issues with glibc/libpthread? Is it looking for some kernel feature that is not in 2.6.30 but is in later kernels?

skip · 2011-02-22 07:31:48

I downgraded the following libraries: glibc libdrm libgl ati-dri and now everything works. Now this is not ideal (especially glibc being downgraded). Is this an issue from running a 2.6.30 kernel?

falconindy · 2011-02-22 13:50:01

skip wrote:

I downgraded the following libraries: glibc libdrm libgl ati-dri and now everything works. Now this is not ideal (especially glibc being downgraded). Is this an issue from running a 2.6.30 kernel?

In short, yes.

nicky7 · 2011-02-23 05:22:44

I seemed have to encountered this with 'nomodeset' as a kernel parameter, removing it allows X to start for me.

steffinger · 2011-02-23 15:10:08

trusted user wrote:

skip wrote:
I downgraded the following libraries: glibc libdrm libgl ati-dri and now everything works. Now this is not ideal (especially glibc being downgraded). Is this an issue from running a 2.6.30 kernel?
In short, yes.

I'm not so sure if the old kernel is really the problem for the segfault. I got the same GPU (ATI Mobility Radeon HD3470) on my Asus F5SL AP177d and I run the newest Kernel and X only stopped segfaulting after I downgraded ati-dri and libgl. I would even warn you not to update your kernel to 2.6.37 if you use KMS. 2.6.37 is not doing well with my grafic chip and before i turned off KMS I had a snowstorm on my display that I couldn't even find out where it came from.

By the way: @skip: can you suspend your laptop? i got the feeling that my GPU is blocking this. (OK this might not belong here but I really want to find out which part of my computer is causing all the problems.)

skip · 2011-02-25 12:51:56

nicky7 wrote:

I seemed have to encountered this with 'nomodeset' as a kernel parameter, removing it allows X to start for me.

I don't have nomodeset in /boot/grub/menu.lst

skip · 2011-02-25 12:54:40

steffinger wrote:

I'm not so sure if the old kernel is really the problem for the segfault. I got the same GPU (ATI Mobility Radeon HD3470) on my Asus F5SL AP177d and I run the newest Kernel and X only stopped segfaulting after I downgraded ati-dri and libgl. I would even warn you not to update your kernel to 2.6.37 if you use KMS. 2.6.37 is not doing well with my grafic chip and before i turned off KMS I had a snowstorm on my display that I couldn't even find out where it came from.
By the way: @skip: can you suspend your laptop? i got the feeling that my GPU is blocking this. (OK this might not belong here but I really want to find out which part of my computer is causing all the problems.)

I don't use KMS as far as I recall.

If X is working I can suspend/resume fine using pm-suspend.

toad · 2011-02-25 21:15:44

A mate has got an R500 and he got the black screen of death, too. Adding the nomodeset command to the kernel grub line he was able to boot into runlevel 3.

Log says seg fault, but we haven't got any further than that yet.

Has anybody tried changing drivers?

We are just updating...

EDIT:

SOLVED - on my mate's Lenovo R500 we got X working by installing the catalyst driver

Oh, that was for a radeon HD 3400 card, he was running on the radeon driver beforehand.

Last edited by toad (2011-02-25 22:29:40)

waquo · 2011-02-26 12:22:04

I had the same problem this morning with X segfaulting on startup. (Radeon HD 4250 onboard, kernel 2.6.37.1-1)

Fixed by downgrading adi-dri (and libgl with it) or by setting radeon.modeset=1 in grub's menu.lst to turn on KMS.
Apparently the error occurs only with latest extra/ati-dri (7.10.0.git20110215-1) and KMS off.

I didn't use KMS before, as it didn't work properly when I tried it a few months back, now it seems to work fine.

waquo · 2011-02-26 12:38:12

I posted a bug-report here: https://bugs.archlinux.org/task/23056
Vote for it!

Last edited by waquo (2011-02-26 12:40:26)

Arch Linux

#1 2011-02-21 22:16:18

X segfaults on start - possibly libpthread/glibc related.

#2 2011-02-21 22:24:55

Re: X segfaults on start - possibly libpthread/glibc related.

#3 2011-02-21 22:32:58

Re: X segfaults on start - possibly libpthread/glibc related.

#4 2011-02-21 22:42:34

Re: X segfaults on start - possibly libpthread/glibc related.

#5 2011-02-21 23:04:07

Re: X segfaults on start - possibly libpthread/glibc related.

#6 2011-02-22 00:54:01

Re: X segfaults on start - possibly libpthread/glibc related.

#7 2011-02-22 07:31:48

Re: X segfaults on start - possibly libpthread/glibc related.

#8 2011-02-22 13:50:01

Re: X segfaults on start - possibly libpthread/glibc related.

#9 2011-02-23 05:22:44

Re: X segfaults on start - possibly libpthread/glibc related.

#10 2011-02-23 15:10:08

Re: X segfaults on start - possibly libpthread/glibc related.

#11 2011-02-25 12:51:56

Re: X segfaults on start - possibly libpthread/glibc related.

#12 2011-02-25 12:54:40

Re: X segfaults on start - possibly libpthread/glibc related.

#13 2011-02-25 21:15:44

Re: X segfaults on start - possibly libpthread/glibc related.

#14 2011-02-26 12:22:04

Re: X segfaults on start - possibly libpthread/glibc related.

#15 2011-02-26 12:38:12

Re: X segfaults on start - possibly libpthread/glibc related.

Board footer