Bug 478424

Summary: kernel-2.6.29-0.6.rc0.git2.fc11.x86_64 (and -0.5.rc0) "white screens" compiz....
Product: [Fedora] Fedora Reporter: Tom London <selinux>
Component: kernelAssignee: Kristian Høgsberg <krh>
Status: CLOSED WORKSFORME QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: medium Docs Contact:
Priority: low    
Version: 11CC: drago01, keith.allcock, kernel-maint, krh, masao-takahashi, sangu.fedora, somlo, vedran, vonbrand
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2009-10-26 11:40:15 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
Output of dmesg obtained from alternate console with "white screen"
none
Xorg.0.log generated with "white screen"
none
~/.xsession-errors generated with "white screen" none

Description Tom London 2008-12-29 21:29:42 UTC
Created attachment 327930 [details]
Output of dmesg obtained from alternate console with "white screen"

Description of problem:
Downloaded kernel-2.6.29-0.6.rc0.git2.fc11.x86_64 from koji.

Appears to boot up fine, gdm login promt is OK, and gnome stuff displays fine, until compiz is started.  Then I get the "white screen of indifference": can see/move cursor, but screen is 100% white.

Ctrl-alt-F2 works fine, as does booting up kernel-2.6.28-3.fc11.x86_64.

I'll attach dmesg, .xsession-errors and Xorg.0.log.

I spotted this in /var/log/messages/dmesg:

Xorg:2606 map pfn expected mapping type write-back for e1800000-e1e40000, got uncached-minus
Xorg:2606 freeing invalid memtype e1800000-e1e40000

Haven't seen this with previous kernels.

I did try to revert to previous version of xorg-x11-server-*, but that did not resolve the issue.

It does seem that booting without compiz may work..... I believe there is a command line command to turn off compiz.  I'll try to locate and try.

Version-Release number of selected component (if applicable):
kernel-2.6.29-0.6.rc0.git2.fc11.x86_64

How reproducible:
Every time

Steps to Reproduce:
1.
2.
3.
  
Actual results:


Expected results:


Additional info:

Comment 1 Tom London 2008-12-29 21:30:52 UTC
Created attachment 327931 [details]
Xorg.0.log generated with "white screen"

Comment 2 Tom London 2008-12-29 21:31:26 UTC
Created attachment 327932 [details]
~/.xsession-errors generated with "white screen"

Comment 3 Tom London 2008-12-29 21:32:00 UTC
Oops, forgot to add:

System is Thinkpad X61 with Intel 965 graphics.

Comment 4 Tom London 2008-12-30 03:53:56 UTC
Looking at Xorg.0.log, I see:

(II) AIGLX: Screen 0 is not DRI2 capable
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: open result is 15, (OK)
drmOpenByBusid: Searching for BusID pci:0000:00:02.0
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: open result is 15, (OK)
drmOpenByBusid: drmOpenMinor returns 15
drmOpenByBusid: drmGetBusid reports pci:0000:00:02.0
(EE) AIGLX error: drmMap of framebuffer failed (Resource temporarily unavailable)(EE) AIGLX: reverting to software rendering
(II) AIGLX: Loaded and initialized /usr/lib64/dri/swrast_dri.so
(II) GLX: Initialized DRISWRAST GL provider for screen 0


Equivalent lines from a "working" Xorg.0.log:

(II) AIGLX: Screen 0 is not DRI2 capable
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: open result is 15, (OK)
drmOpenByBusid: Searching for BusID pci:0000:00:02.0
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: open result is 15, (OK)
drmOpenByBusid: drmOpenMinor returns 15
drmOpenByBusid: drmGetBusid reports pci:0000:00:02.0
(II) AIGLX: enabled GLX_SGI_make_current_read
(II) AIGLX: enabled GLX_MESA_copy_sub_buffer
(II) AIGLX: enabled GLX_SGI_swap_control and GLX_MESA_swap_control
(II) AIGLX: enabled GLX_texture_from_pixmap with driver support
(II) AIGLX: Loaded and initialized /usr/lib64/dri/i965_dri.so
(II) GLX: Initialized DRI GL provider for screen 0


Sorry if this is an xorg-x11 issue.....

Comment 5 Tom London 2008-12-30 14:55:08 UTC
Disabling "Desktop Effects" allows system to boot up to gnome/etc.

Comment 6 Tom London 2008-12-30 16:34:37 UTC
Noticed this similar message when running with compiz disabled:

Dec 30 07:49:38 tlondon kernel: gnome-screensav:3900 map pfn expected mapping type write-back for e1800000-e1e40000, got uncached-minus
Dec 30 07:49:38 tlondon kernel: gnome-screensav:3900 freeing invalid memtype e1800000-e1e40000

Looks similar to the one above....

Comment 7 Sachin Garg 2008-12-30 18:11:32 UTC
I am also having the similar issue with GM965 and i386. But no issues with metacity composite manager "metacity -c". White screen is only with compiz

Comment 8 Tom London 2008-12-31 00:16:04 UTC
For completeness:

Also see the following (similar) messages:

Dec 30 15:59:28 tlondon kernel: glxgears:3589 map pfn expected mapping type write-back for e1800000-e1e40000, got uncached-minus
Dec 30 15:59:28 tlondon kernel: glxgears:3589 freeing invalid memtype e1800000-e1e40000

Comment 9 Sachin Garg 2008-12-31 00:56:12 UTC
when booted with 2.6.29-0.6.rc0.git2.fc11.i686 (output of glxinfo)
-----------------------------------------------
OpenGL vendor string: Mesa Project
OpenGL renderer string: Software Rasterizer
OpenGL version string: 2.1 Mesa 7.3-devel
OpenGL shading language version string: 1.10


when booted with 2.6.28-3.fc11.i686 (output of glxinfo)
------------------------------------

OpenGL vendor string: Tungsten Graphics, Inc
OpenGL renderer string: Mesa DRI Intel(R) 965GM GEM 20080716 x86/MMX/SSE2
OpenGL version string: 2.1 Mesa 7.3-devel
OpenGL shading language version string: 1.10

Comment 10 sangu 2008-12-31 01:57:41 UTC
in NVidia proprietary  driver
$dmesg 
[skip]
Xorg:2576 map pfn expected mapping type write-combining for d8000000-d88e8000, g
ot uncached-minus
Xorg:2576 freeing invalid memtype d8000000-d88e8000
Xorg:2589 map pfn expected mapping type write-combining for d8000000-d88e8000, g
ot uncached-minus
Xorg:2589 freeing invalid memtype d8000000-d88e8000
eth0: no link during initialization.
ADDRCONF(NETDEV_UP): eth0: link is not ready
...

Comment 11 Horst H. von Brand 2009-01-02 21:33:12 UTC
(In reply to comment #5)
> Disabling "Desktop Effects" allows system to boot up to gnome/etc.

I have "Desktop Effects" disabled here, white screen.

kernel-2.6.29-0.6.rc0.git2.fc11.x86_64
kernel-2.6.29-0.7.rc0.git3.fc11.x86_64

Sorry, just noticed that System --> Preferences --> Look and Feel --> Desktop Effects always offers to enable, never to disable...

Comment 12 Horst H. von Brand 2009-01-03 13:56:17 UTC
After really setting Desktop Effects off, and rebooting into kernel-2.6.29-0.7.rc0.git3.fc11.x86_64, after a short while X hangs completely. No mouse movement, ctrl-alt-BS or ctrl-alt-F2 do nothing. Had to turn off (notebook, no reset button).

Comment 13 Tom London 2009-01-03 16:02:28 UTC
Yeah, I've seen this too.

When this happened, I tried to ssh into the laptop from another system, but this failed.

Suspect the system is "really, really dead".

On subsequent boot, I could find no evidence or spew from the crash: I checked /var/log/messages, /var/log/Xorg.0.log, ~/.xsession-errors, etc.

Comment 14 drago01 2009-01-04 17:21:16 UTC
Does passing "nopat" to the kernel help?

Comment 15 Tom London 2009-01-04 17:40:53 UTC
Helps a bit....

Booting with "nopat": I no longer get the the Xorg messages about "invalid mapping type" in /var/log/messages, and glxgears runs much faster (about 930 fps vs. about 250).

But compiz won't start:
compiz (core) - Error: Could not acquire compositing manager selection on screen 0 display ":0.0"
compiz (core) - Fatal: No manageable screens found on display :0.0
gtk-window-decorator: Screen 0 on display ":0.0" already has a decoration manager; try using the --replace option to replace the current decoration manager.
compiz (core) - Error: Could not acquire compositing manager selection on screen 0 display ":0.0"
compiz (core) - Fatal: No manageable screens found on display :0.0

But it no longer "white screens".....

Comment 16 Sachin Garg 2009-01-04 17:52:43 UTC
thnx .. "nopat" works for me.

Tom,

check if you have metacity compositing enabled. if it is please disable it and try

/apps/metacity/general/compositing_manager

Comment 17 Sachin Garg 2009-01-04 17:55:19 UTC
glxinfo output is again following

OpenGL vendor string: Tungsten Graphics, Inc
OpenGL renderer string: Mesa DRI Intel(R) 965GM GEM 20080716 x86/MMX/SSE2
OpenGL version string: 2.1 Mesa 7.3-devel
OpenGL shading language version string: 1.10

Comment 18 Tom London 2009-01-05 02:35:24 UTC
(In reply to comment #16)
> thnx .. "nopat" works for me.
> 
> Tom,
> 
> check if you have metacity compositing enabled. if it is please disable it and
> try
> 
> /apps/metacity/general/compositing_manager

Yeah, that fixed it for me.

Not sure how that got set, since I've been using compiz for quite a while....

So is "nopat" a fix or a workaround?  Is this something I should be adding to my grub.conf?

Comment 19 drago01 2009-01-05 06:43:14 UTC
(In reply to comment #18)
> (In reply to comment #16)
> > thnx .. "nopat" works for me.
> > 
> > Tom,
> > 
> > check if you have metacity compositing enabled. if it is please disable it and
> > try
> > 
> > /apps/metacity/general/compositing_manager
> 
> Yeah, that fixed it for me.
> 
> Not sure how that got set, since I've been using compiz for quite a while....
> 
> So is "nopat" a fix or a workaround?  Is this something I should be adding to
> my grub.conf?

Its a workaround. 
This is a regression in the upstream kernel. A upstream report (lkml) wouldn't hurt.

Comment 20 Horst H. von Brand 2009-01-05 13:10:15 UTC
(In reply to comment #12)
> After really setting Desktop Effects off, and rebooting into
> kernel-2.6.29-0.7.rc0.git3.fc11.x86_64, after a short while X hangs completely.
> No mouse movement, ctrl-alt-BS or ctrl-alt-F2 do nothing. Had to turn off
> (notebook, no reset button).

Yesterday it did work for quite a while with kernel-2.6.29-0.9.rc0.git4.fc11.x86_64, today it hang in GDM after entering the password.

Comment 21 Tom London 2009-01-05 15:13:49 UTC
(In reply to comment #20)
> (In reply to comment #12)
> > After really setting Desktop Effects off, and rebooting into
> > kernel-2.6.29-0.7.rc0.git3.fc11.x86_64, after a short while X hangs completely.
> > No mouse movement, ctrl-alt-BS or ctrl-alt-F2 do nothing. Had to turn off
> > (notebook, no reset button).
> 
> Yesterday it did work for quite a while with
> kernel-2.6.29-0.9.rc0.git4.fc11.x86_64, today it hang in GDM after entering the
> password.

I've had random "hard freezes" (one or two) too.

Not sure this is the same problem or another issue.

When I reboot, I see no evidence or error logging, so I'm at a loss of what to report.

Comment 22 Horst H. von Brand 2009-01-08 16:38:27 UTC
With kernel-2.6.29-0.12.rc0.git7.fc11.x86_64 no mouse works at all (USB mouse or touchpad).

Comment 23 Tom London 2009-01-08 17:33:09 UTC
I've noticed that on some boots, HAL fails to start properly, and when this happens, I get gdm login screen, but no mouse or keyboard.

I haven't been able to track this down further, but rebooting "fixes this" until it happens again.....

Comment 24 Masao Takahashi 2009-01-09 06:46:14 UTC
I have experienced this problem.
Disabling PAT using "menuconfig" resolves this problem.
But compiz does not work. 
 The previous stable kernel-2.6.28 is good.

Comment 25 Tom London 2009-01-14 22:43:47 UTC
compiz "works for me" (without whitescreen!) with kernel-2.6.29-0.35.rc1.git4.fc11.x86_64 without "nopat".

Running Thinkpad X200 (with Intel Corporation Cantiga Integrated Graphics Controller) without xorg.conf.

Comment 26 keith 2009-02-14 11:19:06 UTC
I have been seeing the white screen (where everything locks) using 2.6.29-0.112.rc4.git3.fc11.x86_64.

This happens regardless of driver (NVidia GeForce 7000 series); I have tried, nv, nouveau and proprietary.

If I use the Desktop effects applet (as opposed to commpand line 'compiz --replace') it will not lock so hard, at this time in dmesg I can see:

compiz[3805]: segfault at 3ff782ebd0 ip 0000003ff782ebd0 sp 00007fffa1933328 error 14 in libxml2.so.2.7.3[3ff8400000+152000]

Comment 27 keith 2009-02-14 11:28:05 UTC
nopat stops the white screen, but compiz still fails .. can't find anything logged regards to why

Comment 28 Bug Zapper 2009-06-09 10:32:46 UTC
This bug appears to have been reported against 'rawhide' during the Fedora 11 development cycle.
Changing version to '11'.

More information and reason for this action is here:
http://fedoraproject.org/wiki/BugZappers/HouseKeeping

Comment 29 Vedran Miletić 2009-10-26 11:40:15 UTC
Closing per comment #25.

From keith, what you are seeing is a separate bug and has nothing to do with this one. However:
1) nouveau and nv don't support 3D (and therefore compiz),
2) nvidia proprietary driver is not supported by Fedora.

And, from my experience, this could be an example of broken hardware.