Bug 75818

Summary: Odd color under menu on old hardware on google page with SNA
Product: xorg Reporter: clockley1 <clockley1>
Component: Driver/intelAssignee: Chris Wilson <chris>
Status: RESOLVED FIXED QA Contact: Intel GFX Bugs mailing list <intel-gfx-bugs>
Severity: trivial    
Priority: medium    
Version: unspecified   
Hardware: x86 (IA32)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
Screenshot
none
Screenshot
none
xorg log none

Description clockley1 2014-03-06 02:41:48 UTC
Created attachment 95203 [details]
Screenshot

This only happens when the default depth is 24bits on sna.

Firefox 27.0.1
cairo 1.13.1-0.1.git337ab1f
fedora 20
VGA compatible controller: Intel Corporation 82865G Integrated Graphics Controller (rev 02)
Comment 1 clockley1 2014-03-06 02:42:35 UTC
Created attachment 95204 [details]
Screenshot
Comment 2 Chris Wilson 2014-03-06 08:26:07 UTC
Please attach your Xorg.0.log.
Comment 3 Chris Wilson 2014-03-06 08:29:29 UTC
For reference, the opaque outline around the google popup should be a blur.
Comment 4 Chris Wilson 2014-03-06 10:44:39 UTC
Note, I won't have a chance to look at this till tomorrow, but if you are able to compile your own driver in the meantime that will be useful. If you can compile with ./configure --enable-debug (and perhaps --enable-debug=full) it will help immensely.
Comment 5 clockley1 2014-03-07 02:58:55 UTC
Created attachment 95276 [details]
xorg log

I compiled the driver with --enable-debug=full but it keep crashing whenever I tried to start Firefox.
Comment 6 Chris Wilson 2014-03-07 07:51:11 UTC
The crashes are significant data. One of the steps that --enable-debug does is to enable assertions. Sadly those assertions are only printed out on stderr, and not captured in Xorg.0.log. If you still have the stderr (e.g. /var/log/gdm/:0.log) that would also be very useful.
Comment 7 Chris Wilson 2014-03-08 17:28:01 UTC
The issue exists since acceleration was enabled on 845g -- I was hoping for an easy bisect clue.
Comment 8 Chris Wilson 2014-03-10 12:06:42 UTC
I've isolated it to rendering into an 8-bit (single channel alpha) surface.
Comment 9 Chris Wilson 2014-03-10 14:55:26 UTC
commit 0075c90d3192ba90ff1ae8a7b04bfc3ff4fccda7
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Mon Mar 10 15:17:51 2014 +0000

    sna/gen2: Fix alpha blending with 8bit destination surfaces
    
    On gen2 (like gen3), 8-bit destination surfaces are read into the Green
    channel (and written to from the Green channel). Therefore the expected
    alpha blending must instead be converted to colour blending.
    
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75818
    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.