Please test: Blur on y-axis also when only x-axis is scaled billiear
Hi, I've just found something which looks like a bug in the Radeon/RadeonHD EXA code. When scaling a picture in x-axis and billinear filtering is requested, the y-axis is blurred too. However the version I tested against was the one shipped with Ubuntu-9.04, and I don't have a RadeonHD powered machine myself to investigate further with latest code. Could somebody with the latest drivers please run the testcase attached (gcc -lX11 -lXrender testcase.c -o test). The expected result is a single-pixel wide line which fades out at the ends, the pathological one is a line smeared also in the y-directions. If it fails, it would be great if you could send me a screenshot :) Thank you in advance, Clemens
On Jun 08, 09 06:14:42 -0400, Clemens Eisserer wrote:
I've just found something which looks like a bug in the Radeon/RadeonHD EXA code. When scaling a picture in x-axis and billinear filtering is requested, the y-axis is blurred too.
Can you send a screenshot as well?
Thanks
Matthias
--
Matthias Hopf
Hi Matthias,
The first (top) line was rendered with pixman (NoAccel), the second
line with exa on my HD3850 / Ubuntu 9.04 live cd.
This causes problems with themes that cache pre-rendered elements in
pixmaps and use them scaled later on. (E.g. button outline is blurred
instead of sharp 1px line).
Thanks, Clemens
2009/6/8 Matthias Hopf
On Jun 08, 09 06:14:42 -0400, Clemens Eisserer wrote:
I've just found something which looks like a bug in the Radeon/RadeonHD EXA code. When scaling a picture in x-axis and billinear filtering is requested, the y-axis is blurred too.
Can you send a screenshot as well?
Thanks
Matthias
-- Matthias Hopf
__ __ __ Maxfeldstr. 5 / 90409 Nuernberg (_ | | (_ |__ mat@mshopf.de Phone +49-911-74053-715 __) |_| __) |__ R & D www.mshopf.de -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org
On Jun 08, 09 07:49:24 -0400, Clemens Eisserer wrote:
The first (top) line was rendered with pixman (NoAccel), the second line with exa on my HD3850 / Ubuntu 9.04 live cd.
This indeed looks very wrong.
Alex, any ideas?
Matthias
--
Matthias Hopf
Hi again,
Could somebody please re-test on a recent driver build to see wether
the problem still persists?
Furthermore should I open a bug-report about this issue?
Thanks again, Clemens
2009/6/8, Matthias Hopf
On Jun 08, 09 07:49:24 -0400, Clemens Eisserer wrote:
The first (top) line was rendered with pixman (NoAccel), the second line with exa on my HD3850 / Ubuntu 9.04 live cd.
This indeed looks very wrong. Alex, any ideas?
Matthias
-- Matthias Hopf
__ __ __ Maxfeldstr. 5 / 90409 Nuernberg (_ | | (_ |__ mat@mshopf.de Phone +49-911-74053-715 __) |_| __) |__ R & D www.mshopf.de
-- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org
On Wed, Jun 10, 2009 at 6:18 AM, Clemens Eisserer
Hi again,
Could somebody please re-test on a recent driver build to see wether the problem still persists?
Still appears to be doing it.
Furthermore should I open a bug-report about this issue?
Please. Alex
Thanks again, Clemens
2009/6/8, Matthias Hopf
: On Jun 08, 09 07:49:24 -0400, Clemens Eisserer wrote:
The first (top) line was rendered with pixman (NoAccel), the second line with exa on my HD3850 / Ubuntu 9.04 live cd.
This indeed looks very wrong. Alex, any ideas?
Matthias
-- Matthias Hopf
__ __ __ Maxfeldstr. 5 / 90409 Nuernberg (_ | | (_ |__ mat@mshopf.de Phone +49-911-74053-715 __) |_| __) |__ R & D www.mshopf.de -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org
-- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org
2009/6/10 Clemens Eisserer
Could somebody please re-test on a recent driver build to see wether the problem still persists? Furthermore should I open a bug-report about this issue?
Seems to be working as expected on RV770. -- Yang Zhao http://yangman.ca
Hi Yang,
Seems to be working as expected on RV770. The screenshot seems broken. The line is smeared over 2 pixel rows in y direction, whereas it should be only one.
Thanks for testing, Clemens -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org
2009/6/10 Clemens Eisserer
Seems to be working as expected on RV770. The screenshot seems broken. The line is smeared over 2 pixel rows in y direction, whereas it should be only one.
Ah, it's only supposed to be a 1px wide line. I'll see if I have some time to look at the code driving this. -- Yang Zhao http://yangman.ca -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org
Hi Yang,
Ah, it's only supposed to be a 1px wide line. I'll see if I have some time to look at the code driving this.
That would be really cool :) It causes artifacts like: https://bugs.freedesktop.org/attachment.cgi?id=19758 (should be) https://bugs.freedesktop.org/attachment.cgi?id=19759 (is) Interestingly <= i945 suffers from exactly the same problem, but >= i965 doesn't have this problem. The corresponding intel bug has been reported at: https://bugs.freedesktop.org/show_bug.cgi?id=16917 (and entered the typical open-but-nobody-cares-state) Thanks, Clemens -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org
2009/6/8 Clemens Eisserer
I've just found something which looks like a bug in the Radeon/RadeonHD EXA code. When scaling a picture in x-axis and billinear filtering is requested, the y-axis is blurred too.
This should be fixed by 645a0a0a that was committed today. Cheers, -- Yang Zhao http://yangman.ca -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org
participants (4)
-
Alex Deucher
-
Clemens Eisserer
-
Matthias Hopf
-
Yang Zhao