[exa] XSetClipMask+XPutImage = garbage on an off-screen buffer

[GiuseppeVerde]

Binary package hint: grace

Fonts in grace in Jaunty are atrociously and unusually bad. Picture of Pain attached.

[lspci]
00:00.0 Host bridge [0600]: Intel Corporation Mobile 945GME Express Memory Controller Hub [8086:27ac] (rev 03)
     Subsystem: ASUSTeK Computer Inc. Device [1043:830f]
00:02.0 VGA compatible controller [0300]: Intel Corporation Mobile 945GME Express Integrated Graphics Controller [8086:27ae] (rev 03)
     Subsystem: ASUSTeK Computer Inc. Device [1043:830f]

[Dirteat (dirteat)]

Created an attachment (id=20898)
xorg.conf (in working XAA mode)

[Éric Piel (pieleric)]

I can reproduce it here, with xmgrace: the fonts in the graph are shown as black noisy squares. However, I cannot notice this in any other program. GTK and QT programs work fine. xedit, xmag are fine as well. The widgets of xfontsel are displayed correctly, however, it doens't show any font in the font display (but I haven't checked if in XAA it works).

I also have a 64bits computer with server 1.5.3 (on Mandriva cooker). Driver is intel 2.5.1, with a :
00:02.1 Display controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 0c)

[Michel-tungstengraphics (michel-tungstengraphics)]

Unless this can be reproduced with other drivers, it's probably a driver issue.

[GiuseppeVerde (launchpad-digitasaru)]

Binary package hint: grace

Fonts in grace in Jaunty are atrociously and unsuably bad. Picture of Pain attached.

[Evgeny Stambulchik (fnevgeny)]

You're using an Intel graphics chipset, are you?

http://plasma-gate.weizmann.ac.il/Grace/phpbb/viewtopic.php?t=1547
http://forums.fedoraforum.org/showthread.php?p=1132569

[GiuseppeVerde (launchpad-digitasaru)]

Yep.

[Evgeny Stambulchik (fnevgeny)]

Please attach xdpyinfo output and lspci|grep VGA.

What do you see if you disable the font antialiasing (View->Page Setup)?

[GiuseppeVerde (launchpad-digitasaru)]

lspci: 00:02.0 VGA compatible controller: Intel Corporation Mobile 945GME Express Integrated Graphics Controller (rev 03)
xdpyinfo attached.

Disabling antialiasing has no effect.

[Evgeny Stambulchik (fnevgeny)]

Please try xglyph (from t1lib-bin) like

xglyph /usr/share/grace/fonts/type1/n021003l.pfb

and press on "String", "StringX", "AAString", and "AAStringX" buttons and describe what you see (better attach snapshots).

[Evgeny Stambulchik (fnevgeny)]

Try disabling hardware acceleration in the i915 driver: put

Option "NoAccel" "true"

in the "Driver" section of your /etc/X11/xorg.conf.

[GiuseppeVerde (launchpad-digitasaru)]

Finally got a chance to test that.

With the abovementioned option, horkage is fixed. Typing in the identi.ca "What's up" box seems more responsive. The memscroller screensaver is much, much slower.

[Evgeny Stambulchik (fnevgeny)]

Moved to Intel's driver

[Erik (erik-redhat-bugs)]

Description of problem:
Fonts in the graph area in xmgrace are shown with black boxes surrounding them, and are difficult to read. Occasionally resizing the axes will fix some of the text, or make it worse by rewriting new text over the old text without clearing the background. This seems to only occur on machines with recent integrated Intel graphics cards using the i915 driver, as reported by lsmod. The program is otherwise usable. No data is required to view the bug, it will be obvious immediately upon starting xmgrace.

Version-Release number of selected component (if applicable):
Fedora 10

How reproducible:
Occurs 100% of the time on the appropriate hardware.

Steps to Reproduce:
1. Procure video hardware using the i915 video driver.
2. Install Fedora Core 10
3. sudo yum install grace
4. xmgrace

Actual results:
Fonts in the graph are difficult to read.

Expected results:
The fonts should be clear.

Additional info:
Screen shot and hardware parameters of at least two individuals experiencing this problem are available here:
http://forums.fedoraforum.org/showthread.php?t=207920

[Erik (erik-redhat-bugs)]

Mistake above, the Version-Release should be:

Version-Release number of selected component (if applicable):
grace-5.1.22-2.fc10.i386

[Geir Ove Myhr (gomyhr)]

This doesn't look like the typical intel driver bug, but nonetheless could you attach the following information which is required for all intel driver bugs (please upload as separate files):
- /var/log/Xorg.0.log
- /etc/X11/xorg.conf
- the output of `lspci -vvnn`

[GiuseppeVerde (launchpad-digitasaru)]

Requested attachments attached.

[Bryce Harrington (bryce)]

See please https://wiki.ubuntu.com/X/Troubleshooting/HugeFonts

In particular, check if "542x203 millimeters" seems a reasonable measurement for your monitor's physical size. If not, your monitor may be reporting incorrect information. (IIRC we've seen this at least once before on AUO displays). In this case please provide 'xrandr --verbose' and 'get-edid | parse-edid' output as described on the above page.

[José (jos-redhat-bugs)]

I am trying to replicate the problem. The question is that I don't have any machine here with an intel graphics card. :-(
Without further evidence I suspect that this a bug on the i915 driver.

So for the moment this a note of "work in progress"...

[Erik (erik-redhat-bugs)]

Would it be helpful if I submitted a bug report for the i915 driver? (I don't know if I would do that here or somewhere else, like kernel.org.) Or would it be helpful to test other motif apps? The only two I have and am aware of are OpenDX and Nedit, and in neither of them have I observed this issue. Is there maybe another app I could test?

[Matěj (matj-redhat-bugs)]

Thanks for the bug report. We have reviewed the information you have provided above, and there is some additional information we require that will be helpful in our diagnosis of this issue.

Please attach your X server config file (/etc/X11/xorg.conf, if available) and X server log file (/var/log/Xorg.*.log) to the bug report as individual uncompressed file attachments using the bugzilla file attachment link below.

Could you please also try to run without any /etc/X11/xorg.conf (if you have one) whatsoever and let X11 autodetect your display and video card? Attach to this bug /var/log/Xorg.0.log from this attempt as well, please.

We will review this issue again once you've had a chance to attach this information.

Thanks in advance.

[RudraB (rudrab-redhat-bugs)]

I am also facing the same problem. i don't have any xorg.conf so grace is not biased to i915

[Matěj (matj-redhat-bugs)]

Yes, and could we get the files I asked for from both of you, please?

[Andrew (andrew-redhat-bugs)]

Created attachment 333628
Xorg log with no xorg.conf

I'm hitting the same problem.

[RudraB (rudrab-redhat-bugs)]

Created attachment 333633
/var/log/Xorg.0.log file when no xorg.conf exist

[RudraB (rudrab-redhat-bugs)]

May I request to increase the priority of this bug as we people in research fied depends heavily on xmgrace and also many of my friends hesitating to fc10 because of this problem.

[Erik (erik-redhat-bugs)]

Created attachment 333778
Xorg.*.log files

I have attached my Xorg.log files in the gzipped tar archive. I have no xorg.conf file.

[Andrea (andrea-redhat-bugs)]

I have the Intel i915 graphic card, and I confirm the same problem. Did someone tryed to investigate if the problem presists even in Fedora 11 ? I will supply later my xorg.0.log file.

[Andrea (andrea-redhat-bugs)]

Created attachment 333992
My xorg log file in a standard configuration

[Andrea (andrea-redhat-bugs)]

Created attachment 333993
Grace at start

This is how Grace appears when it starts: the fonts corruption is evident.

[GiuseppeVerde (launchpad-digitasaru)]

1) The problem is not huge fonts, it's *corrupted* fonts.
2) "542x203 millimeters" is not being auto-detected. From xorg.0.log (already attached to this bug):
(II) intel(0): Setting screen physical size to 195 x 113
From xdpyinfo (already attached to this bug)
  dimensions: 1024x600 pixels (195x113 millimeters)
From xrandr --verbose:
LVDS connected 1024x600+0+0 (0x3d) normal (normal left inverted right x axis y axis) 195mm x 113mm

Maybe you accidentally put this on the wrong bug?

Attaching xrandr output

[GiuseppeVerde (launchpad-digitasaru)]

http://bugs.freedesktop.org/show_bug.cgi?id=19167

[José (jos-redhat-bugs)]

OK, I can attest that this problem does not show up with other cards. I have tested ATI, nvidia and Sis. That explains why I have not noticed the problem. :-(

I am changing this to xorg-x11-drv-intel since this only happen with intel cards.

[Patrick (patrick-redhat-bugs)]

I do experience the same kind of problems with a Q965 (Intel) graphic card by using gnome. In addition using KDE, it just aborts.
Switching to versa driver fixes the gnome issue but not the KDE one.

[Andrea (andrea-redhat-bugs)]

On a x86_64 with Intel graphics no problem.

00:02.0 VGA compatible controller: Intel Corporation 82Q963/Q965 Integrated Graphics Controller (rev 02)

Linux ntmastellone.cira.it 2.6.27.12-170.2.5.fc10.x86_64 #1 SMP Wed Jan 21 01:33:24 EST 2009 x86_64 x86_64 x86_64 GNU/Linux

grace-5.1.22-2.fc10.x86_64

[Evgeny Stambulchik (fnevgeny)]

I don't know about other applications (?) mentioned in the original report, but the problem with xmgrace is not a font issue at all. See latest comments to <https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/325169>. Hopefully the simple demo attached will help you tracking the bug down.

[Michel Dänzer (michel-daenzer)]

Created an attachment (id=25473)
Take client clip type into account for pixmap migration

> Hopefully the simple demo attached will help you tracking the bug down.

Yes, simple demos to reproduce bugs are usually very helpful.

Does this xserver patch fix the demo/xmgrace with EXA?

[Evgeny Stambulchik (fnevgeny)]

See a (relatively) short standalone code demonstrating the bug. If the problem is not seen immediately, try resizing the window a few times.

Notes:

1. The crucial point is the XSetClipMask() call. If you disable it (change if(1) to if(0) in the line above that), the problem goes away.
2. This is also specific to rendering to an off-screen buffer; direct drawing on win is fine (but Grace uses buffered output).
3. It's the same with the 2.4.1 driver backported <https://wiki.ubuntu.com/ReinhardTartler/X/RevertingIntelDriverTo2.4>, so the bug must have been there for long, but only now triggered by a change in xorg at a higher level.
4. With 2.7.0 compiled from the sources the bug is still there. However, if I select "AccelMethod" "UXA" (new in 2.7.0) everything is fine. "AccelMethod" "XAA" hard locks my box (a P1610 lifebook) with any version of the driver.

[Evgeny Stambulchik (fnevgeny)]

Yes it does, thanks - great & fast work! Can you please explain why only Intel hardware is affected?

[Evgeny Stambulchik (fnevgeny)]

See <https://bugs.freedesktop.org/show_bug.cgi?id=18950#c5>. The patch provided works for me; so it sounds like xserver-xorg-core should be fixed.

[Michel Dänzer (michel-daenzer)]

(In reply to comment #6)
> Yes it does, thanks - great & fast work!

Well, it's not the final fix yet - there are other places which need the same treatment.

> Can you please explain why only Intel hardware is affected?

Doesn't look like that's the case after all.

[Evgeny Stambulchik (fnevgeny)]

(In reply to comment #7)

> Doesn't look like that's the case after all.

Yeah, the patch is generic; but can you reproduce the bug with any other driver? All reports I've seen so far are from Intel-hardware owners:

<https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/325169>
<https://bugzilla.redhat.com/show_bug.cgi?id=484273>
<http://plasma-gate.weizmann.ac.il/Grace/phpbb/viewtopic.php?t=1547>
<http://plasma-gate.weizmann.ac.il/Grace/phpbb/viewtopic.php?t=1557>
<http://plasma-gate.weizmann.ac.il/Grace/phpbb/viewtopic.php?t=1657>

[Evgeny Stambulchik (fnevgeny)]

OK, there is now an evidence the bug affects other hardware, too: <https://bugzilla.redhat.com/show_bug.cgi?id=493131>

[Evgeny Stambulchik (fnevgeny)]

*** Bug 19167 has been marked as a duplicate of this bug. ***

[Evgeny Stambulchik (fnevgeny)]

i386 is also affected.
NetBSD is also affected.

[Michel Dänzer (michel-daenzer)]

Fix pushed to Git master and nominated for server-1.6-branch.

[Evgeny Stambulchik (fnevgeny)]

Fix exists upstream. Please apply.

[keram (marek-sierka)]

Adding

 Option "EXAOptimizeMigration" "true"
 Option "MigrationHeuristic" "greedy"

to the "Device" section of my /etc/X11/xorg.conf solved the issue for me:

Section "Device"
 Identifier "Configured Video Device"
 Option "EXAOptimizeMigration" "true"
 Option "MigrationHeuristic" "greedy"
EndSection

[lspci]
00:02.0 VGA compatible controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 0c)
00:02.1 Display controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 0c)

[Evgeny Stambulchik (fnevgeny)]

The patch: <http://cgit.freedesktop.org/xorg/xserver/commit/?id=850675d4de4373e5df95507dbf2cd9affaaf54bc>

[James (james-redhat-bugs)]

I have a similar problem, though I'm using a different card and it happens only on certain characters in thunderbird/firefox

00:02.0 VGA compatible controller: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller (rev 03)

[James (james-redhat-bugs)]

(In reply to comment #17)
> I have a similar problem, though I'm using a different card and it happens only
> on certain characters in thunderbird/firefox
>
> 00:02.0 VGA compatible controller: Intel Corporation Mobile 945GM/GMS,
> 943/940GML Express Integrated Graphics Controller (rev 03)

Ah yes, and I should mention I'm using the same kernel drivers and I'm using the intel driver for xorg, with no configuration file.

And this is on Fedora 11, interestingly, I didn't see the problem in F10

[Andrew (andrew-redhat-bugs)]

Finally came across the patch to fix this, please apply it and commit to a new release for F11. I have tested it and it does indeed fix the problem.

http://cgit.freedesktop.org/xorg/xserver/commit/?id=850675d4de4373e5df95507dbf2cd9affaaf54bc

Its a relatively minor patch and fixes a reasonably serious issue.

[Alex Valavanis (valavanisalex)]

This has been fixed in karmic. I think the change to uxa acceleration sorted the problem out.

[Geir Ove Myhr (gomyhr)]

The patch http://cgit.freedesktop.org/xorg/xserver/commit/?id=850675d4de4373e5df95507dbf2cd9affaaf54bc is not included in Karmic, since it is using xorg-server 1.6.3 + selected patches, but will we in 1.7.0 and newer. However, since this bug is in the EXA migration code, this bug seems to affect only intel hardware and the 2.8.x intel drivers only use UXA, this should not be a problem in Karmic. I'm therefore closing this bug, but please reopen if it's still a problem.

[Bug (bug-redhat-bugs)]

This message is a reminder that Fedora 10 is nearing its end of life.
Approximately 30 (thirty) days from now Fedora will stop maintaining
and issuing updates for Fedora 10. It is Fedora's policy to close all
bug reports from releases that are no longer maintained. At that time
this bug will be closed as WONTFIX if it remains open with a Fedora
'version' of '10'.

Package Maintainer: If you wish for this bug to remain open because you
plan to fix it in a currently maintained version, simply change the 'version'
to a later Fedora version prior to Fedora 10's end of life.

Bug Reporter: Thank you for reporting this issue and we are sorry that
we may not be able to fix it before Fedora 10 is end of life. If you
would still like to see this bug fixed and are able to reproduce it
against a later version of Fedora please change the 'version' of this
bug to the applicable version. If you are unable to change the version,
please add a comment here and someone will do it for you.

Although we aim to fix as many bugs as possible during every release's
lifetime, sometimes those efforts are overtaken by events. Often a
more recent Fedora release includes newer upstream software that fixes
bugs or makes them obsolete.

The process we are following is described here:
http://fedoraproject.org/wiki/BugZappers/HouseKeeping

[Bug (bug-redhat-bugs)]

Fedora 10 changed to end-of-life (EOL) status on 2009-12-17. Fedora 10 is
no longer maintained, which means that it will not receive any further
security or bug fix updates. As a result we are closing this bug.

If you can reproduce this bug against a currently maintained version of
Fedora please feel free to reopen this bug against that version.

Thank you for reporting this bug and we are sorry it could not be fixed.