Created attachment 38751 [details] i965.png System Environment: -------------------------- Arch: i386 Platform: Capella OSD: Fedora release 13 (Goddard) Libdrm: (master)2.4.21-21-g7ec9a1effa4f551897f91f3b017723a8adf011d9 Mesa: (7.9)d46b7c9f6f91f80b7349067dd5be52f0af736ec1 Xserver: (master)xorg-server-1.9.0 Xf86_video_intel: (master)2.12.0-87-g08c2caca48323d6d5701dcef3486f850619d7905 Cairo: (master)cb0bc64c16b3a38cbf0c622830c18ac9ea6e2ffe Libva: (master)6772bdb4406edaf55da2e3604003c9eafff31c36 Kernel: (master)9fe6206f400646a2322096b56c59891d530e8d51 Bug detailed description: --------------------------- With Mesa(7.9)d46b7c9f6f91f80b7349067dd5be52f0af736ec1, Mesa demo spriteblast render incorrectly using i965. But using swrast, it render correctly. Reproduce steps: ---------------- 1.xinit& 2. Mesa_demos/src/demos/spriteblast
Created attachment 38752 [details] swrast.png
mesa demos pointblast also have the same problem
Bumping this! Using this setup: Intel Corporation Arrandale Integrated Graphics Controller (rev 02) xorg-server-1.9.2.902 libdrm, mesa + xf86-video-intel git master Both spriteblast and pointblast render incorrectly when run with the hardware driver. Switching to the gallium software renderer fixes both cases. I guess it also introduces this problem to wine: http://bugs.winehq.org/attachment.cgi?id=32783 This is a screenshot of running FEAR (demo) on wine-1.3.10, which clearly shows that sprite rendering is broken. This also affects some other effects in the game, e.g. muzzle flash, etc.
It should also be noted that this seems to be (at least) partially tied to GLSL compilation. In wine the issues disappear when disabling the use of GLSL (then only ARB_fp and ARB_vp are used).
Tobias, that app isn't using point sprites. It's not even using points.
Created attachment 44992 [details] result from GPU Caps Viewer normal screenshot of the problem
The problem is maybe related to this problem, running OpenGL demo 'GLSL point sprite particles' in "GPU Caps Viewer" using WINE results incorrect rendering. System: Ubuntu 10.10 Linux 2.6.36-02063604-generic Mesa3D 7.10.1 (earlier Mesas' also have this problem) WINE 1.3.16 GPU Caps Viewer v1.10.0 Intel GMA 4500 - 256MB the result is bundled in the attachment Thanks.
*** Bug 42080 has been marked as a duplicate of this bug. ***
Is this specific to Calpella platform, or has it been fixed? I don't see this on my Sandybridge using plain Fedora 16.
It appears pass on our Sandy Bridge. Chao, can you confirm it on our Calpella?
I test this on Calpella platform(Fedora 13) and do not see this issue, So maybe this isuue has been fixed.
This bug has been present from the beginning of any Mesa version for Intel GMA 4500 chipset. I am not sure for other Intel chipset though. Still present in git, Mesa 7.11.1, 7.10.3, back to 7.7.1 Gallium software rendering from Mesa git now almost renders this properly. export LIBGL_ALWAYS_SOFTWARE=1 same for GLSL point sprite particles from GPU Caps Viewer renders almost perfectly by gallium software rendering.
OK, I can reproduce it on GM45 platform now which environment as follows: OSD: Fedora release 13 (Goddard) Kernel_version: 2.6.31.12 Libdrm: (master)2.4.26-18-gcc088f1721eaa5f8f1ba1932723882f92e34c39a Mesa: (7.11)bf7b347c1041edc77472c117e0ffcfbb71d5c13b Xserver: (server-1.11-branch)xorg-server-1.11.1
This seems only happening with GM45. Recently tested on Intel GMA 945 with GLSL enabled from driconf renders fine but not with GM45 with Mesa 8 stable and 8.1-devel from git. Also on Sandy Bridge it works fine. Therefore the bug is releted to only GM45 GL driver code. Cheers
Also with Gallium software renderer it renders fine with Mesa 8.1-devel from git
still present in the git
Patch is on the list
commit 50a5d5dea0c21886bc3445c0ad0928b03e64ab10 Author: Eric Anholt <eric@anholt.net> Date: Wed Feb 20 14:00:47 2013 -0800 i965: Fix the W value of deprecated pointcoords on pre-gen6.
Bo, please verify.
it is fixed in master, but there is one more similar bug like I mentioned before with GPU Caps Viewer.
It has been fixed on my side.
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.