compiz (decor) - Warn: failed to bind pixmap to texture

Bug Description

[Test Case]
1. Open a terminal and run: tail -f ~/.xsession-errors
2. Open/close lots of windows.
-> Observe the terminal output and verify that no "compiz (decor) - Warn: failed to bind pixmap to texture" messages on the terminal

[Regression Potential]
Broken window decorations in compiz

Original description:

Whenever I run compiz I get lots of these in the log:
compiz (decor) - Warn: failed to bind pixmap to texture

Either this indicates a bug that needs fixing, or this warning is lower priority than it sounds and should be removed...

On 2012年2月10日, Daniel van Vugt wrote:

> Public bug reported:
>
> Whenever I run compiz I get lots of these in the log:
> compiz (decor) - Warn: failed to bind pixmap to texture
>
> Either this indicates a bug that needs fixing, or this warning is lower
> priority than it sounds and should be removed...
>

Indeed, its a synchronization issue between the decorator and the decor
plugin, I'll need to implement some kind of XSync barriers for decoration
updates since there's a race where a pixmap can be destroyed before its
bound.

> ** Affects: compiz-core
> Importance: Undecided
> Status: New
>
> --
> You received this bug notification because you are a member of Compiz
> Maintainers, which is the registrant for Compiz Core.
> https://bugs.launchpad.net/bugs/929989
>
> Title:
> compiz (decor) - Warn: failed to bind pixmap to texture
>
> Status in Compiz Core:
> New
>
> Bug description:
> Whenever I run compiz I get lots of these in the log:
> compiz (decor) - Warn: failed to bind pixmap to texture
>
> Either this indicates a bug that needs fixing, or this warning is
> lower priority than it sounds and should be removed...
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/compiz-core/+bug/929989/+subscriptions
>

Sounds like the kind of graphics glitch I should be able to see... Is this related to the flashing white decorations bug 454218?

On that note... why is *-window-decorator a separate program instead of a plugin? Sounds like there would be fewer issues if it wasn't in a separate process (like Gnome Shell AFAIK).

On 2012年2月10日, Daniel van Vugt wrote:

> On that note... why is *-window-decorator a separate program instead of
> a plugin? Sounds like there would be fewer issues if it wasn't in a
> separate process (like Gnome Shell AFAIK).
>

Partly political reasons, partly performance reasons - generating the
decorations out of process has pretty huge benefits because you can
optimize the pixmap generation (eg instead of drawing on each frame window
you can generate one decoration and use a bit of protocol to paint it
correctly).

I explored this a lot more in
lp:~smspillaz/compiz-core/compiz-core.decorator but we never merged it

> --
> You received this bug notification because you are a member of Compiz
> Maintainers, which is the registrant for Compiz Core.
> https://bugs.launchpad.net/bugs/929989
>
> Title:
> compiz (decor) - Warn: failed to bind pixmap to texture
>
> Status in Compiz Core:
> New
>
> Bug description:
> Whenever I run compiz I get lots of these in the log:
> compiz (decor) - Warn: failed to bind pixmap to texture
>
> Either this indicates a bug that needs fixing, or this warning is
> lower priority than it sounds and should be removed...
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/compiz-core/+bug/929989/+subscriptions
>

Has there been any update on this??? I am also affected by 'failed to bind pixmap to texture'.

it happens quite frequently, and has for a couple of months now.

I'm also seeing this as well.

Fix committed into lp:compiz-core at revision 3131

@Daniel

r3131 does not fix the problem, in fact (at least not with 'vanilla' compiz / no unity) it causes gtk-window-decorator to be very unstable.

I've filed a bug report here;

https://bugs.launchpad.net/compiz-core/+bug/996901

I actually get way more 'failed to bind pixmap to texture' with r3131 and window decorations are pretty much unusable.

cheerz

Fix reverted in r3132. Apparently it was causing instability in some cases (bug 996901)

Fix committed into lp:compiz-core at revision 3137

Fix committed into lp:compiz-core/0.9.7 at revision 3103

This patch causes gtk-window-decorator to be unstable.
I ma running Ubuntu 12.04 (Precise) and I built and tested compiz (1:0.9.7.8-0ubuntu1.1) in the precise-proposed wait queue and narrowed down the patch causing the instability to this one. This patch was cherry-picked for the SRU.

With this patch gtk-window-decorator crashes and leaves me without window decorations. The error in .xsession-errors:

The program 'gtk-window-decorator' received an X Window System error.
This probably reflects a bug in the program.
The error was 'BadWindow (invalid Window parameter)'.
(Details: serial 5433 error_code 3 request_code 20 minor_code 0)
(Note to programmers: normally, X errors are reported asynchronously;
that is, you will receive the error a while after causing it.
To debug your program, run it with the --sync command line
option to change this behavior. You can then get a meaningful
backtrace from your debugger if you break on the gdk_x_error() function.)

The easiest way to reproduce it is using IntelliJ (a Java IDE). Hitting Ctrl+N opens up a small window, and closing it by hitting Esc immediately triggers the gtk-window-decorator crash. I attached a screenshot of IntelliJ with the small popup in the middle.

Thanks Miklos. Please try to capture a stack of gtk-window-decorator crashing:

1. Kill gtk-window-decorator if it's not dead already.
2. Run: gdb /usr/bin/gtk-window-decorator
3. (gdb) run --replace
4. Trigger the crash.
5. Go back to gdb.
6. (gdb) bt
7. Copy and paste the stack info it shows you into a NEW bug.

Hello Daniel, or anyone else affected,

Accepted compiz into precise-proposed. The package will build now and be available at http://launchpad.net/ubuntu/+source/compiz/1:0.9.7.8-0ubuntu1.1 in a few hours and then in the -proposed repository. Please help us by testing this new package. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Your feedback will aid us getting this update out to other Ubuntu users.

If this package fixes the bug for you please change the bug tag from verification-needed to verification-done. If it does not, change the tag to verification-failed. In either case details of your testing will help us make a better decision. Further information regarding the verification process can be found at https://wiki.ubuntu.com/QATeam/PerformingSRUVerification . Thank you in advance!

This bug was fixed in the package compiz - 1:0.9.8+bzr3249-0ubuntu1

---------------
compiz (1:0.9.8+bzr3249-0ubuntu1) quantal-proposed; urgency=low

* New upstream snapshot.
- Fall back to a refresh rate that is more likely to look correct; 60Hz.
(LP: #1009338)
- Benchmark plugin should consume its key binding, and not pass the key to
the underlying window. (LP: #1009320)
- Avoid needless STL operations leading to expensive heap operations.
(LP: #1006335)
- Fix a typo that was causing (LP: #1002606)
(widthIncBorders/heightIncBorders)
- Check if the window is decorated before trying to change its event window
states (which won't exist if not decorated) (LP: #1007754)
- Use the XDamage extension more efficiently (the way it was designed to be
used). This dramatically reduces CPU usage, reduces wakeups, and
increases frame rates. It also solves at least one observed performance
bug (LP: #1007299) and probably several more.
- Avoid constructing and destructing lots of strings on every single event,
which was wasting lots of CPU (LP: #1005569)
- md LINGUAS doesn't exist, it's mnk (Mandinka in ISO 639-3)
- Move grid plugin to google test and don't depend on the plugin for the
test (LP: #1005009)
- Don't read plugin.Initialized and test the value. (LP: #1004848)
- libcompizconfig's install () commands were still using the old includedir
and libdir variables rather than their libcompizconfig_* variants.
(LP: #1005176)
- Execute the cmake files separately to ensure that DESTDIR is respected.
(LP: #1005177)
- Don't set_target_properties on a target that might not exist
(LP: #1005008)
- Don't allow windows which we weren't even tracking as decoratable to
become decorated if they try and change their hints. (LP: #963794)
- Change the mouse pointer while dragging windows in expo. Just like the
ubuntu branches do. (LP: #987647)
- Fix uninitialized memory use (LP: #1004338)
- Fix uninitialized variable (LP: #1004335)
- Delay unbinding of pixmaps until then next rebind (LP: #729979)
(LP: #1002602)
- Don't drop plugins from the list to try and load before you've even tried
to load them. Doing so makes missing plugins silently ignored instead of
an error message (LP: #1002715). It also means valid plugins in more
unusual, but real locations in LD_LIBRARY_PATH will never get loaded
(LP: #1002721).
- If running test cases under a real X server, we don't care if Xvfb is
missing (LP: #994841)
- Don't assume pkg_check_modules always sets _PREFIX (LP: #993608)
- Don't clear selections in ~PrivateScreen because it causes a race between
the existing and the new compiz instances, breaking --replace and
non-replace behaviour. (LP: #988684) (LP: #989545)
- Always paint with infiniteRegion as the clip region if the window is
transformed and always use the supplied region if painting with offset or
on transformed screen. (LP: #987639)
- Add synchronization primitives to the decoration protocol so that there
isn't a r...

Verification done. Looks good on quantal today.

@Daniel:

I can easily crash the proposed compiz version on several machines on Precise but I can't get any stacktrace. This is what I get:

mjuhasz@desktop:~$ gdb /usr/bin/gtk-window-decorator
GNU gdb (Ubuntu/Linaro 7.4-2012.04-0ubuntu2) 7.4-2012.04
Copyright (C) 2012 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law. Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu".
For bug reporting instructions, please see:
<http://bugs.launchpad.net/gdb-linaro/>...
Reading symbols from /usr/bin/gtk-window-decorator...(no debugging symbols found)...done.
(gdb) run --replace
Starting program: /usr/bin/gtk-window-decorator --replace
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
[New Thread 0x7fffef1a9700 (LWP 7631)]
[New Thread 0x7fffee9a8700 (LWP 7632)]

The program 'gtk-window-decorator' received an X Window System error.
This probably reflects a bug in the program.
The error was 'BadWindow (invalid Window parameter)'.
(Details: serial 10112 error_code 3 request_code 20 minor_code 0)
(Note to programmers: normally, X errors are reported asynchronously;
that is, you will receive the error a while after causing it.
To debug your program, run it with the --sync command line
option to change this behavior. You can then get a meaningful
backtrace from your debugger if you break on the gdk_x_error() function.)
[Thread 0x7fffee9a8700 (LWP 7632) exited]
[Thread 0x7fffef1a9700 (LWP 7631) exited]
[Inferior 1 (process 7628) exited with code 01]
(gdb)
(gdb) bt
No stack.
(gdb)

Miklos, please log a new bug with the details.

Regression (Broken window decorations in compiz) reported here: LP: #1019337

Causing a regression does not mean this bug is not fixed. It is fixed.

Yes, this bug is fixed, the warning is gone from the terminal. The warning however never bothered me but the constant crashes with the proposed package do.

I read in the linked guide: https://wiki.ubuntu.com/QATeam/PerformingSRUVerification
, that "If you encounter a regression in a package that has been uploaded to *-proposed, please:":
step 5: Set the verification-failed tag on the corresponding SRU bug report.

I'm sorry if I misunderstood something.

On the pending-sru page all bug links are green related to version 0.9.7.8-0ubuntu1.1. Will the package in -proposed be pushed into -updates with the regression I reported (LP: #1019337)? The window decorator is confirmed to crash with IntelliJ and PhpStorm as well and probably all JetBrains products since they share the same codebase. These are IDEs used in corporate environments where stability is a major concern.

Then mark one of the bugs verification-failed with a link to your bug. Then at
least it'll be clear there's an issue that needs to be evaluated.

I've set bug #993608 as verification-failed to prevent the copy to -updates

@Daniel:
> Causing a regression does not mean this bug is not fixed. It is fixed.

right, but failing verification is the standard way to block a SRU to avoid it moving to -updates with a regression

@Mikklos: having a stacktrace of the segfault by running gtk-window-decorator --sync would be useful

This bug was fixed in the package compiz - 1:0.9.7.8-0ubuntu1.1

---------------
compiz (1:0.9.7.8-0ubuntu1.1) precise-proposed; urgency=low

[ Łukasz 'sil2100' Zemczak ]
* Cherry-pick upstream patches:
- CMake Error at FindCompiz.cmake:84 (include): include could not find load
file: CompizDefaults (LP: #993608)
- compiz (decor) - Warn: failed to bind pixmap to texture (LP: #929989)
- gtk-window-decorator crashed with SIGFPE in
_decor_blend_horz_border_picture() (LP: #963794)
- [callgrind] compiz spends ~25% of its time constructing/destructing
strings in PrivateScreen::handleActionEvent (LP: #1005569)
- gtk-window-decorator crashed with SIGSEGV in meta_get_decoration_geometry
(LP: #1007754)
- composite refresh rate falls back to 50Hz, which is wrong in most cases
(LP: #1009338)
- [callgrind] compiz spends ~7% of its time inserting into and destructing
the events list in PrivateScreen::processEvents() (LP: #1006335)

[ Didier Roche ]
* debian/control:
- change Vcs-Bzr to point to precise branches
-- Lukasz 'sil2100' Zemczak <email address hidden> 2012年6月20日 15:02:31 +0100

Had a bit of a race condition here. This will get reverted out of precise-
updates shortly.

@Sebastien:
> having a stacktrace of the segfault by running gtk-window-decorator --sync would be useful

Sure, no problem. See LP: #1019337

thanks, not that you might need the libgtk dbg package to have the function resolution working

Stacktrace with the dbg packages installed.

Funny thing. Why Ubuntu is always lying about their bugs? fix is commited ? of course not, I'm experiencing the same bug for al long time on the newest version of compiz.

Committed means the fix is in code and has been accepted for future release. A fix is not available to users until it says Fix Released.
[https://wiki.ubuntu.com/Bugs/Status]

Hello Daniel, or anyone else affected,

Accepted compiz into precise-proposed. The package will build now and be available at http://launchpad.net/ubuntu/+source/compiz/1:0.9.7.8-0ubuntu1.3 in a few hours, and then in the -proposed repository.

Please help us by testing this new package. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Your feedback will aid us getting this update out to other Ubuntu users.

If this package fixes the bug for you, please change the bug tag from verification-needed to verification-done. If it does not, change the tag to verification-failed. In either case, details of your testing will help us make a better decision.

Further information regarding the verification process can be found at https://wiki.ubuntu.com/QATeam/PerformingSRUVerification . Thank you in advance!

The proposed package fixed the issue.

Hello Daniel, or anyone else affected,

Accepted compiz into precise-proposed. The package will build now and be available at http://launchpad.net/ubuntu/+source/compiz/1:0.9.7.8-0ubuntu1.4 in a few hours, and then in the -proposed repository.

Please help us by testing this new package. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Your feedback will aid us getting this update out to other Ubuntu users.

If this package fixes the bug for you, please change the bug tag from verification-needed to verification-done. If it does not, change the tag to verification-failed. In either case, details of your testing will help us make a better decision.

Further information regarding the verification process can be found at https://wiki.ubuntu.com/QATeam/PerformingSRUVerification . Thank you in advance!

Still fixed in 1.4, and no regressions spotted.

The verification of this Stable Release Update has completed successfully and the package has now been released to -updates. Subsequently, the Ubuntu Stable Release Updates Team is being unsubscribed and will not receive messages about this bug report. In the event that you encounter a regression using the package from -updates please report a new bug using ubuntu-bug and tag the bug report regression-update so we can easily find any regresssions.