LinuxQuestions.org
Welcome to the most active Linux Forum on the web.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Slackware
User Name
Password
Slackware This Forum is for the discussion of Slackware Linux.

Notices


Reply
  Search this Thread
Old 05-12-2018, 11:03 PM   #1
Rod3775
Member
 
Registered: May 2008
Location: Washington State
Distribution: Slackware-current
Posts: 94
Blog Entries: 10

Rep: Reputation: 18
slackware-current xorg-server-1.20.0 with xf86-video-nouveau segfaults


After upgrading to xorg-server-1.20.0 (and upgrading to the -3 build of xf86-video-nouveau-1.0.15 for the new ABI), X crashes after entering data in the login screen. The xsession file does not get successfully created, and .xsession-errors records a segfault.

Reverting to server-1.19.6 and the corresponding -1 build of nouveau-1.0.15 gets things going again. The upgrade was from the 5/10/18 and 5/12/18 -current ChangeLog. I didn't revert anything else.

The Xorg.log for the server-1.20.0 startup mentions using udev for various things, so maybe a udev update is needed? The 1.19.6 log didn't mention udev so much. Or some other incompatibility.

My system uses the xfce4 window manager, and the segfault occurs during its startup. "startxfce4" fails, whether invoked at the command line, or via a "telinit 4" startup. I'll stay reverted until a new set of xfce4 packages is released, then try again.

Last edited by Rod3775; 05-13-2018 at 09:02 AM. Reason: additional info
 
Old 05-13-2018, 01:25 AM   #2
H1p8r10n
Member
 
Registered: Feb 2016
Location: on the border of milky way
Distribution: Slackware
Posts: 149

Rep: Reputation: Disabled
confirmed here with ati driver.

same rollback to make it work
 
Old 05-13-2018, 02:25 AM   #3
H1p8r10n
Member
 
Registered: Feb 2016
Location: on the border of milky way
Distribution: Slackware
Posts: 149

Rep: Reputation: Disabled
for radeon, this config fixes the problem :

Code:
Section "Device"
	Identifier  "Radeon"
	Driver "radeon"
	Option "ShadowPrimary" "on"
	Option "TearFree" "on"
EndSection
Cheers
JP
 
Old 05-13-2018, 04:49 PM   #4
Rod3775
Member
 
Registered: May 2008
Location: Washington State
Distribution: Slackware-current
Posts: 94

Original Poster
Blog Entries: 10

Rep: Reputation: 18
more info on server crash

Further exploration running slackware64-current

I tried different window managers using xwmconfig and startx to get a GUI window
These are the results:

xorg-server: = 1.19.6 ........ 1.20.0
xfce4 ......... OK ............ crashes X
fluxbox ....... OK ............ OK
blackbox ...... OK ............ OK
xfwm2 ......... OK ............ OK
wmaker ........ OK ............ OK
twm ........... OK ............ OK
kde4 .......... not installed / not installed

traceback from startx: (the warnings from xfce4-session are a red herring - they occur in
1.19.6 also, but it runs normally)

xauth: file /root/.serverauth.10487 does not exist


X.Org X Server 1.20.0
X Protocol Version 11, Revision 0
Build Operating System: Slackware 15.0 Slackware Linux Project
Current Operating System: Linux core5.localdomain 4.14.40 #2 SMP Wed May 9 15:56:50 CDT 2018 x86_64
Kernel command line: BOOT_IMAGE=Linux-ssd-m ro root=811 vt.default_utf8=1 acpi_enforce_resources=lax
Build Date: 10 May 2018 01:09:13PM

Current version of pixman: 0.34.0
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.2.log", Time: Sun May 13 14:00:19 2018
(==) Using system config directory "/usr/share/X11/xorg.conf.d"
resize called 1920 1080
/usr/bin/startxfce4: X server already running on display :2
gpg-agent[10531]: WARNING: "--write-env-file" is an obsolete option - it has no effect
gpg-agent: a gpg-agent is already running - not starting a new one

(xfce4-session:10525): xfce4-session-WARNING **: 14:00:19.723: gpg-agent returned no PID in the variables

(xfce4-session:10525): xfce4-session-WARNING **: 14:00:19.725: xfsm_manager_load_session: Something wrong with /root/.cache/sessions/xfce4-session-core5.localdomain:2, Does it exist? Permissions issue?
(EE)
(EE) Backtrace:
(EE) 0: /usr/libexec/Xorg (OsLookupColor+0x139) [0x58a319]
(EE) 1: /lib64/libpthread.so.0 (funlockfile+0x50) [0x7f91d23083cf]
(EE) 2: /usr/libexec/Xorg (miRenderColorToPixel+0xe) [0x4fef4e]
(EE) 3: /usr/lib64/xorg/modules/libexa.so (exaEnableDisableFBAccess+0x1fa6) [0x7f91cdc6f3e6]
(EE) 4: /usr/libexec/Xorg (DamageRegionAppend+0x676) [0x511406]
(EE) 5: /usr/libexec/Xorg (AddTraps+0x36d5) [0x5095d5]
(EE) 6: /usr/libexec/Xorg (SendErrorToClient+0x348) [0x43fa48]
(EE) 7: /usr/libexec/Xorg (InitFonts+0x3b8) [0x443928]
(EE) 8: /lib64/libc.so.6 (__libc_start_main+0xe7) [0x7f91d1f2dba7]
(EE) 9: /usr/libexec/Xorg (_start+0x2a) [0x42e5ba]
(EE)
(EE) Segmentation fault at address 0x8
(EE)
Fatal server error:
(EE) Caught signal 11 (Segmentation fault). 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.2.log" for additional information.
(EE)
(II) AIGLX: Suspending AIGLX clients for VT switch
(EE) Server terminated with error (1). Closing log file.
... messages from unhappy X clients deleted ....

So the segfault is in the server, probably unrelated to the particular video driver, but is triggered by calls made from the xfce4-session client.
Sorry I couldn't try kde4, but I'll bet it works, and nobody tested xfce4 before posting the upgrade...
 
Old 05-13-2018, 07:34 PM   #5
orbea
Senior Member
 
Registered: Feb 2015
Distribution: Slackware64-current
Posts: 1,950

Rep: Reputation: Disabled
Can you try sharing your backtrace with xorg developers? One place you can do that is in the irc channel #xorg-devel @ freenode.
 
Old 05-15-2018, 08:00 PM   #6
Rod3775
Member
 
Registered: May 2008
Location: Washington State
Distribution: Slackware-current
Posts: 94

Original Poster
Blog Entries: 10

Rep: Reputation: 18
workaround for the Xorg crash

This bug turns out to be caused by a fatal interaction between the "compositor" option of xfwm4 and xorg-server-1.20.0. Finding it was more challenging because xfwm4 is so old that it uses deprecated (and perhaps no longer functional) gdk-2 functions to write back to the screen. It won't compile with --enable-debug=full for this reason. So I got the source and tweaked the SlackBuild to fix THAT, and got a trace in xfwm4. As a workaround I added --disable-compositor to the configure call, and all is fine. Since the -current version of xfwm4 does NOT crash xorg-server-1.19.6, the newer server is probably too gullible about whatever strange stuff xfwm4 is sending it. So the Xorg folks need to be more suspicious, and xfwm4-4.12.4 needs to be cleaned up. I may look at that later, but for now the new server is usable with xfce4.

Last edited by Rod3775; 05-16-2018 at 09:42 AM. Reason: typo in version number
 
2 members found this post helpful.
Old 05-16-2018, 06:50 AM   #7
H1p8r10n
Member
 
Registered: Feb 2016
Location: on the border of milky way
Distribution: Slackware
Posts: 149

Rep: Reputation: Disabled
Hi Folks,

It may be a bit more complex... (related to intel graphics ?)

Here (on Radeon) I run stock Slackware XFWM4 on Xorg 1.20, with compositor enabled, with good 2D and 3D GL performance..

The only problem I had to fix is disabling WebGL in Firefox.

All is perfect with radeon Xorg config I posted above, and disabling WebGL in Firefox

My 2 cents
JP
 
Old 05-16-2018, 09:50 AM   #8
Rod3775
Member
 
Registered: May 2008
Location: Washington State
Distribution: Slackware-current
Posts: 94

Original Poster
Blog Entries: 10

Rep: Reputation: 18
Xorg radeon result

JP's experience suggests that there is an interaction in the server that depends on which graphics driver is being used. This is not surprising, given that each video card provides different support for compositing. My card is an nVidia NVA-8 using the current nouveau driver. Perhaps xfwm4 is the trigger for the crash, but not ultimately the culprit. I'm going to try to determine what the specific X request is that triggers the crash. More later.
 
Old 05-16-2018, 10:55 AM   #9
allend
LQ 5k Club
 
Registered: Oct 2003
Location: Melbourne
Distribution: Slackware64-15.0
Posts: 6,357

Rep: Reputation: 2739Reputation: 2739Reputation: 2739Reputation: 2739Reputation: 2739Reputation: 2739Reputation: 2739Reputation: 2739Reputation: 2739Reputation: 2739Reputation: 2739
Just want to confirm that on my 64bit -current system (as at 14 May 2018) using the nouveau driver I also see the segfault reported in post #4. I can confirm that recompiling xfwm4 with the --disable-compositor option enabled allows xfce to start.
Thanks to Rod3775 for the heavy lifting.
 
1 members found this post helpful.
Old 05-16-2018, 11:59 AM   #10
H1p8r10n
Member
 
Registered: Feb 2016
Location: on the border of milky way
Distribution: Slackware
Posts: 149

Rep: Reputation: Disabled
I think that maybe the nouveau driver has some config options that could solve the problem in a similar way has I did for radeon.

I will try to upgrade one of the other machines I have here with Nvidia graphics to see if I find something (next WE)

All the best
JP
 
2 members found this post helpful.
Old 05-19-2018, 05:29 AM   #11
Aymeric3
LQ Newbie
 
Registered: Apr 2015
Posts: 15

Rep: Reputation: Disabled
Hello,

Same problem here: segfault. Should I just do a rollback of xorg until the nouveau driver works again?

Thanks.
 
Old 05-19-2018, 08:24 AM   #12
orbea
Senior Member
 
Registered: Feb 2015
Distribution: Slackware64-current
Posts: 1,950

Rep: Reputation: Disabled
Someone who can reproduce this should report this to the xorg developers.
 
Old 05-19-2018, 08:37 PM   #13
allend
LQ 5k Club
 
Registered: Oct 2003
Location: Melbourne
Distribution: Slackware64-15.0
Posts: 6,357

Rep: Reputation: 2739Reputation: 2739Reputation: 2739Reputation: 2739Reputation: 2739Reputation: 2739Reputation: 2739Reputation: 2739Reputation: 2739Reputation: 2739Reputation: 2739
At xorg they put it on how mesa has been built. https://bugs.freedesktop.org/show_bug.cgi?id=106410
 
Old 05-19-2018, 09:17 PM   #14
Rod3775
Member
 
Registered: May 2008
Location: Washington State
Distribution: Slackware-current
Posts: 94

Original Poster
Blog Entries: 10

Rep: Reputation: 18
[SOLVED] patch for xorg-server crash

The bug that crashed xorg-server-1.20.0 was actually found in 2009, but the fix obviously went away somewhere along the line:

commit 1088073b11ed488c0df45af3867b900ef93c6fe1
Author: Michel Dänzer <daenzer@vmware.com>
Date: Fri Oct 9 11:31:44 2009 +0200

EXA: Fix exaTryDriverSolidFill() for solid source pictures.

Solid pictures have a NULL pFormat field, but their format is always
PICT_a8r8g8b8.

Signed-off-by: Michel Dänzer <daenzer@vmware.com>
Signed-off-by: Keith Packard <keithp@keithp.com>

*******************************************************

My version of the fix (not in exaTryDriverSolidFill, but in the routine it calls):


--- /tmp/x11-build/xorg-server-1.20.0/render/mipict.c 2018-05-10 09:32:37.000000000 -0700
+++ /usr/local/slackbuilds/x11-build/xorg-server-1.20.0/render/mipict.c 2018-05-19 16:04:52.420306670 -0700
@@ -393,8 +393,10 @@
{
CARD32 r, g, b, a;
miIndexedPtr pIndexed;
-
- switch (format->type) {
+ if(! format)
+ *pixel = (color->alpha >> 8 << 24) | (color->red >> 8 << 16) | (color->green & 0xff00) | (color->blue >> 8);
+ else
+ switch (format->type) {
case PictTypeDirect:
r = color->red >> (16 - Ones(format->direct.redMask));
g = color->green >> (16 - Ones(format->direct.greenMask));

*********************************

I guess that changing the ChangeLog doesn't mean the the code stays fixed...
 
2 members found this post helpful.
Old 05-19-2018, 10:23 PM   #15
orbea
Senior Member
 
Registered: Feb 2015
Distribution: Slackware64-current
Posts: 1,950

Rep: Reputation: Disabled
Quote:
Originally Posted by allend View Post
At xorg they put it on how mesa has been built. https://bugs.freedesktop.org/show_bug.cgi?id=106410
This is not the same crash as in the op.

Quote:
Originally Posted by Rod3775 View Post
The bug that crashed xorg-server-1.20.0 was actually found in 2009, but the fix obviously went away somewhere along the line:

commit 1088073b11ed488c0df45af3867b900ef93c6fe1
Author: Michel Dänzer <daenzer@vmware.com>
Date: Fri Oct 9 11:31:44 2009 +0200

EXA: Fix exaTryDriverSolidFill() for solid source pictures.

Solid pictures have a NULL pFormat field, but their format is always
PICT_a8r8g8b8.

Signed-off-by: Michel Dänzer <daenzer@vmware.com>
Signed-off-by: Keith Packard <keithp@keithp.com>

*******************************************************

My version of the fix (not in exaTryDriverSolidFill, but in the routine it calls):


--- /tmp/x11-build/xorg-server-1.20.0/render/mipict.c 2018-05-10 09:32:37.000000000 -0700
+++ /usr/local/slackbuilds/x11-build/xorg-server-1.20.0/render/mipict.c 2018-05-19 16:04:52.420306670 -0700
@@ -393,8 +393,10 @@
{
CARD32 r, g, b, a;
miIndexedPtr pIndexed;
-
- switch (format->type) {
+ if(! format)
+ *pixel = (color->alpha >> 8 << 24) | (color->red >> 8 << 16) | (color->green & 0xff00) | (color->blue >> 8);
+ else
+ switch (format->type) {
case PictTypeDirect:
r = color->red >> (16 - Ones(format->direct.redMask));
g = color->green >> (16 - Ones(format->direct.greenMask));

*********************************

I guess that changing the ChangeLog doesn't mean the the code stays fixed...
Given the date I think its safe to call this a new bug. Assuming the original fix still doesn't work it should be reported to the xorg developers. There is a possibility they could provide a patch to solve this. They have always done this when I have shown them xorg backtraces.
 
  


Reply

Tags
slackware-current, xorg update


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off



Similar Threads
Thread Thread Starter Forum Replies Last Post
[SOLVED] Problem starting Xorg Slackware-Current with nouveau Melke Slackware 14 05-04-2015 04:11 PM
[SOLVED] Xorg / nouveau failure on fresh -current install astrogeek Slackware 5 03-27-2013 04:05 AM
/testing libdrm and xf86 nouveau drivers in Slackware 14.0 H_TeXMeX_H Slackware 15 10-11-2012 06:18 AM
xf86-video-nouveau as a tarball: will it be included in Slackware 14? Didier Spaier Slackware 13 09-29-2012 04:30 AM
xf86-video-nouveau on Slackware sahko Slackware 0 08-06-2009 10:29 AM

LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Slackware

All times are GMT -5. The time now is 02:29 AM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Open Source Consulting | Domain Registration