| Summary: | On PPC, OpenGL programs use incorrect texture colors. | ||
|---|---|---|---|
| Product: | Mesa | Reporter: | GhostlyDeath <ghostlydeath> |
| Component: | Drivers/Gallium/r300 | Assignee: | Default DRI bug account <dri-devel> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | medium | CC: | dand1972, erhard_f, ghostlydeath, herminio.hernandezjr, joseph, lowtech, mathieu.malaterre, mpi, nathanhi, riesebie, stupidyetsmart7, svu, will_schmidt |
| Version: | 11.2 | ||
| Hardware: | PowerPC | ||
| OS: | Linux (All) | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
| Attachments: |
PrBoom
OpenArena NeverBall AdanaxisGPL GLXInfo lspci Xorg log WIP endianness fixes PrBoom with patch Adjust vertex format swizzles to compensate for the vertex fetcher byte swap attachment-31027-0.html attachment-23608-0.html |
||
Created attachment 54299 [details]
OpenArena
Created attachment 54300 [details]
NeverBall
Created attachment 54301 [details]
AdanaxisGPL
Created attachment 54302 [details]
GLXInfo
Created attachment 54303 [details]
lspci
Created attachment 54304 [details]
Xorg log
This only affects textures, whether they were created with image loading or if they were created with PBOs. Anything that does not use textures, appear how they are supposed to appear. Created attachment 54470 [details] [review] WIP endianness fixes Here's some Mesa/Gallium endianness fixes I've been working on. Note that I'm serious about 'WIP'; at the very least, this will need to be split/cleaned up, and e.g. the util changes are just a quick'n'dirty hack which mixes driver specific requirements into driver independent code. Also, this will probably break colours with the r600g driver on big endian, but that should be easy to fix by disabling its use of GPU byte swapping for textures and render targets, and I think that's necessary anyway to work correctly in all cases possible with OpenGL. I'm still seeing some wrong colours e.g. in gnome-shell and assaultcube, either I'm missing some more vital formats, or there could be similar issues with vertex colours. Hello, I tested the patch and most of the colors are correct now. However I spotted some rendering problems. Here's the summary of what I found. Neverball: Everything looks correct now, all the colors are good and nothing seems wrong. Adanaxisgpl: Everything renders correctly even with compiled shaders. Openarena: Most of the colors are correct, HOWEVER there are serious rendering problems now. Most polygons just go *poof* and re-appear and just flicker like crazy. This varies the FPS by alot since some polygons in front of other polygons disappear causing everything behind it to render also. Color wise, textures are the correct colors and such. **Smoke, colored UI text, railguns, and my own character don't render colors correctly. The title screen rendered very badly and had about 1 frame every 5-7 seconds. PrBoom: Colors are correct but it suffers from disappearing polygons that OpenArena suffers from. Some misc OpenGL SuperBible 4th edition tests: * C08/toon: Renders correctly * C08/tunnel: Renders correctly * C11/sphereworld: Renders correctly * C11/thunderbird: Renders correctly * C11/thundergl: Renders correctly (VBOs are not complete garbled messes anymore) * C17/bumpmap: The bumpmap renders correctly now * C18/fbodrawbuffers: Renders colors correctly I'm going to check it out again with the as of current latest git and see if the problems are still around (they may have been fixed already). Created attachment 54516 [details]
PrBoom with patch
Here is what PrBoom looks like (note I used IDBEHOLDL to set everything to full brightness).
From what I can tell, only PrBoom and OpenArena have this problem.
However softpipe does not have polygons disappearing at all, so it would appear to be a r300g problem. PrBoom uses GL_TRIANGLE_STRIPs to draw the walls, so this is probably unrelated to this bug.
And another thing I noticed fixed. The OGLSB4th Edition chapter 16 vertexblend draws the shiny light correctly (before it was slightly garbled, it would be a bright white but with distorted red/white/pink in the middle).
Created attachment 55416 [details] [review] Adjust vertex format swizzles to compensate for the vertex fetcher byte swap With this incremental patch, AssaultCube, gnome-shell and PrBoom look good here. Can't test OpenArena right now due to an issue with its current Debian powerpc build. I still get rendering errors with PrBoom with the patch applied. Do you have a complete patch that can be applied to a more recent git revision? By the way, I rebuilt with: # make clean # make distclean # patch -p1 < befixes1.patch # patch -p1 < befixes2.patch # ./configure --prefix=/opt/xorg --with-dri-driverdir=/opt/xorg/lib/dri/ --with-gallium-drivers=r300,r600,swrast --enable-glx-tls --enable-gles1 --enable-gles2 --enable-openvg --enable-xorg --enable-gallium-egl --with-x # make # rm -rf /opt/xorg # make install I'm running into this bug as well, with an iMac G5 containing a Radeon 9600 (rv350) chip. It's been a few months since this bug has seen any activity; is a fix still forthcoming? Is there anything I can do to help? I am not familiar with OpenGL or Mesa's codebase, but I can test on my hardware as needed. And, well, if this bug has been abandoned, I'm willing to spend some time trying to familiarize myself enough with the codebase to, at minimum, attempt an updated patch based on the patches provided previously. (In reply to comment #13) > I'm running into this bug as well, with an iMac G5 containing a Radeon 9600 > (rv350) chip. It's been a few months since this bug has seen any activity; is a > fix still forthcoming? Is there anything I can do to help? I am not familiar > with OpenGL or Mesa's codebase, but I can test on my hardware as needed. And, > well, if this bug has been abandoned, I'm willing to spend some time trying to > familiarize myself enough with the codebase to, at minimum, attempt an updated > patch based on the patches provided previously. Is this the latest code? I have not tested my Powermac G4 in a long time and it uses out of date Mesa (but the colors are correct). (In reply to comment #14) > (In reply to comment #13) > > I'm running into this bug as well, with an iMac G5 containing a Radeon 9600 > > (rv350) chip. It's been a few months since this bug has seen any activity; is a > > fix still forthcoming? Is there anything I can do to help? I am not familiar > > with OpenGL or Mesa's codebase, but I can test on my hardware as needed. And, > > well, if this bug has been abandoned, I'm willing to spend some time trying to > > familiarize myself enough with the codebase to, at minimum, attempt an updated > > patch based on the patches provided previously. > > Is this the latest code? > > I have not tested my Powermac G4 in a long time and it uses out of date Mesa > (but the colors are correct). This is with version 8.0.4-2 as packaged by Debian, which is the most recent package available in the sid (unstable) distribution. So, it's the latest available in Debian. As far as this bug is concerned, no, I don't believe any changes have been made that would have fixed this yet. Have these patches found their way up in the current codebase or is it still necessary to apply them on measa 11.2.x and beyond? I do not believe the patches were ever applied upstream. Last time I tried about two years ago, the patches did not work against the latest mesa at the time. Patch "fdo43698-vtxfmt-endianness.diff" applies against 11.2.2 via patch -p1, the others do not. Are things working any better with mesa from git without any patches? Oded has done a lot of fixes for r600g and Marek committed some fixes for r300g. Compiled and using Mesa 11.3.0-devel (git-59156b2) now on my PowerMac G5 7,3 with an AGP Radeon 9600, rv350 class card. Tested so far: AssaultCube, Extreme TuxRacer, RedEclipse, Sauerbraten. All installed from Ubuntu MATE 16.04 repositories. Garbled textures: RedEclipse, Sauerbraten. Extreme TuxRacer runs and is actually playable! However the colors are wrong and the backgrounnd sky is flickering constantly. AssaultCube crashes after showing the "Loading..." window. See https://bugs.launchpad.net/ubuntu/+source/assaultcube/+bug/1581805 Extreme TuxRacer and Sauerbraten have problems while changing the resolution. Both get REALLY sluggish then, hardware acceleration does not seem to work any longer. The console output starts showing "radeon 000:f0:10.0: ring 0 stalled for more than xxxx msec" messages during these resolution change problems in Extreme TuxRacer. Created attachment 123755 [details] attachment-31027-0.html -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 Micheal I tried to apply the patch to the git mesa source and I got these errors. Am I doing something wrong? rican-linux@lubuntu-powerbook:~/Build/mesa/src$ patch -p2 --dry-run < ~/Downloads/wip.patch checking file gallium/auxiliary/util/u_index_modify.c Hunk #1 FAILED at 35. Hunk #2 FAILED at 44. Hunk #3 FAILED at 66. Hunk #4 FAILED at 88. Hunk #5 FAILED at 97. Hunk #6 FAILED at 118. 6 out of 6 hunks FAILED checking file gallium/drivers/r300/r300_render_translate.c Hunk #1 FAILED at 51. 1 out of 1 hunk FAILED checking file mesa/main/colormac.h Hunk #1 succeeded at 65 with fuzz 2 (offset -27 lines). Hunk #2 FAILED at 119. 1 out of 2 hunks FAILED checking file mesa/main/compiler.h Hunk #1 FAILED at 226. 1 out of 1 hunk FAILED can't find file to patch at input line 184 Perhaps you used the wrong -p or --strip option? The text leading up to this was: -------------------------- |diff --git a/src/mesa/main/depthstencil.c b/src/mesa/main/depthstencil.c |index af5c12f..67553a4 100644 |--- a/src/mesa/main/depthstencil.c |+++ b/src/mesa/main/depthstencil.c -------------------------- File to patch: Skip this patch? [y] Skipping patch. 21 out of 21 hunks ignored can't find file to patch at input line 432 Perhaps you used the wrong -p or --strip option? The text leading up to this was: -------------------------- |diff --git a/src/mesa/main/format_pack.c b/src/mesa/main/format_pack.c |index 390b494..b04663f 100644 |--- a/src/mesa/main/format_pack.c |+++ b/src/mesa/main/format_pack.c -------------------------- File to patch: Skip this patch? [y] Skipping patch. 3 out of 3 hunks ignored can't find file to patch at input line 484 Perhaps you used the wrong -p or --strip option? The text leading up to this was: -------------------------- |diff --git a/src/mesa/main/format_unpack.c b/src/mesa/main/format_unpack.c |index 4f23f3d..ebd17fb 100644 |--- a/src/mesa/main/format_unpack.c |+++ b/src/mesa/main/format_unpack.c -------------------------- File to patch: Skip this patch? [y] Skipping patch. 25 out of 25 hunks ignored checking file mesa/main/formats.c Hunk #1 FAILED at 312. Hunk #2 FAILED at 330. Hunk #3 FAILED at 348. Hunk #4 FAILED at 402. Hunk #5 FAILED at 429. Hunk #6 FAILED at 438. Hunk #7 FAILED at 456. Hunk #8 FAILED at 474. Hunk #9 FAILED at 638. Hunk #10 FAILED at 1314. Hunk #11 FAILED at 2062. Hunk #12 FAILED at 2093. Hunk #13 FAILED at 2121. Hunk #14 FAILED at 2131. Hunk #15 FAILED at 2141. Hunk #16 FAILED at 2185. Hunk #17 succeeded at 1186 (offset -1075 lines). Hunk #18 FAILED at 2598. Hunk #19 FAILED at 2620. Hunk #20 FAILED at 2668. Hunk #21 FAILED at 2778. 20 out of 21 hunks FAILED checking file mesa/main/formats.h Hunk #1 FAILED at 92. Hunk #2 FAILED at 146. Hunk #3 FAILED at 242. 3 out of 3 hunks FAILED checking file mesa/main/texformat.c Hunk #1 FAILED at 103. Hunk #2 FAILED at 135. Hunk #3 FAILED at 149. Hunk #4 FAILED at 163. Hunk #5 FAILED at 195. Hunk #6 FAILED at 213. Hunk #7 FAILED at 322. Hunk #8 FAILED at 333. Hunk #9 FAILED at 344. Hunk #10 FAILED at 355. Hunk #11 FAILED at 367. Hunk #12 FAILED at 759. Hunk #13 FAILED at 786. 13 out of 13 hunks FAILED checking file mesa/main/texstore.c Hunk #1 succeeded at 142 (offset -837 lines). Hunk #2 FAILED at 1126. Hunk #3 succeeded at 287 (offset -859 lines). Hunk #4 FAILED at 1166. Hunk #5 succeeded at 1271 with fuzz 2 (offset -870 lines). Hunk #6 FAILED at 2282. Hunk #7 FAILED at 2297. Hunk #8 FAILED at 2402. Hunk #9 FAILED at 2472. Hunk #10 FAILED at 2514. Hunk #11 FAILED at 2529. Hunk #12 FAILED at 2573. Hunk #13 FAILED at 3328. Hunk #14 FAILED at 3547. Hunk #15 FAILED at 3567. Hunk #16 FAILED at 3598. Hunk #17 FAILED at 4514. Hunk #18 FAILED at 4550. Hunk #19 FAILED at 4576. 16 out of 19 hunks FAILED checking file mesa/state_tracker/st_cb_clear.c Hunk #1 FAILED at 389. Hunk #2 FAILED at 445. 2 out of 2 hunks FAILED checking file mesa/state_tracker/st_format.c Hunk #1 succeeded at 53 with fuzz 1 (offset -94 lines). Hunk #2 FAILED at 203. Hunk #3 succeeded at 177 with fuzz 2 (offset -88 lines). Hunk #4 FAILED at 299. Hunk #5 FAILED at 334. Hunk #6 FAILED at 475. Hunk #7 FAILED at 526. Hunk #8 succeeded at 660 with fuzz 2 (offset 67 lines). Hunk #9 succeeded at 697 with fuzz 2 (offset 66 lines). Hunk #10 FAILED at 666. Hunk #11 FAILED at 804. Hunk #12 FAILED at 1570. Hunk #13 FAILED at 1586. Hunk #14 FAILED at 1602. 10 out of 14 hunks FAILED rican-linux@lubuntu-powerbook:~/Build/mesa/src$ -----BEGIN PGP SIGNATURE----- Version: Mailvelope v1.4.0 Comment: https://www.mailvelope.com wsFcBAEBCAAQBQJXN//JCRBZfgHS2LQxKAAAdnwP/RacNj1k5rQWFJ50RxSC DoFeJCxPXspbhXiNXQpCON87ZKWXpphwwXS+/a/BPfi/rfQyLyaK6W34Gmv+ 9eMzTOeteLSzo9H5vHs+GUj48bJsitrsXTZ8yjxrExYdKsqXk5AfOjKVBtOk waic8mN4k+JughYpTVLvVnqtbslDm9ov4iP+h5u/hF/oIERnzzcKA2UhBHxn G6zociCVFlWWE/4/tCaWg9CJk0DOkez0Hk83jdMfaiYK4yCseqIktctlA3TM XI0npSG3zUi0etMwNcmcz6gMQToTYjqoeJTk0ZIi4BeVDgUW858F0OqSkwMm g3a3+GysP495puqo+UdrJOGVAWzGbPEq8dS8l4k37BHA80znl+KO8jS4TR5V l7fB3lvrC/h/CP/788vESIDNtQeXH+Zy87EHkoX+7x9+Jg2R7CbAYjacp5rx qwXEj+CDOW44sqmGldEqgT7gmxhK1EuVuvSqaue5G0XQVTfse1ogzQOz0zGk GSXuNFeoJLysVX27pDElJ1igY7wX/hliz1Vjcv1Ps9FlwF2Y43FS/1uXagg9 hiCQ6E+AgFVSiOWQIXHFmUArOM2r44lQF+Fh2vkAqPdp0AgOkGwEATDdJRdK K4PaXEBodQnXsNLZ5z24KVIpImP+hz76nriGQUE3iS1TVT6E9tXGG7h/6Ssa 6jFM =rjrD -----END PGP SIGNATURE----- On Sat, May 14, 2016 at 6:26 AM, <bugzilla-daemon@freedesktop.org> wrote: > *Comment # 20 <https://bugs.freedesktop.org/show_bug.cgi?id=43698#c20> on > bug 43698 <https://bugs.freedesktop.org/show_bug.cgi?id=43698> from > erhard_f@mailbox.org <erhard_f@mailbox.org> * > > Compiled and using Mesa 11.3.0-devel (git-59156b2) now on my PowerMac G5 7,3 > with an AGP Radeon 9600, rv350 class card. > > Tested so far: AssaultCube, Extreme TuxRacer, RedEclipse, Sauerbraten. All > installed from Ubuntu MATE 16.04 repositories. > > Garbled textures: RedEclipse, Sauerbraten. > > Extreme TuxRacer runs and is actually playable! However the colors are wrong > and the backgrounnd sky is flickering constantly. > > AssaultCube crashes after showing the "Loading..." window. Seehttps://bugs.launchpad.net/ubuntu/+source/assaultcube/+bug/1581805 > > Extreme TuxRacer and Sauerbraten have problems while changing the resolution. > Both get REALLY sluggish then, hardware acceleration does not seem to work any > longer. The console output starts showing "radeon 000:f0:10.0: ring 0 stalled > for more than xxxx msec" messages during these resolution change problems in > Extreme TuxRacer. > > ------------------------------ > You are receiving this mail because: > > - You are on the CC list for the bug. > > The patch won't apply, because it's too old. Created attachment 124066 [details] attachment-23608-0.html ok thanks! On Tue, May 24, 2016 at 2:17 PM, <bugzilla-daemon@freedesktop.org> wrote: > *Comment # 22 <https://bugs.freedesktop.org/show_bug.cgi?id=43698#c22> on > bug 43698 <https://bugs.freedesktop.org/show_bug.cgi?id=43698> from Marek > Olšák <maraeo@gmail.com> * > > The patch won't apply, because it's too old. > > ------------------------------ > You are receiving this mail because: > > - You are on the CC list for the bug. > > Comment on attachment 123755 [details] attachment-31027-0.html ><div dir="ltr"><pre>-----BEGIN PGP SIGNED MESSAGE----- >Hash: SHA256 >Micheal I tried to apply the patch to the git mesa source and I got these errors. Am I doing something wrong? > >rican-linux@lubuntu-powerbook:~/Build/mesa/src$ patch -p2 --dry-run < ~/Downloads/wip.patch >checking file gallium/auxiliary/util/u_index_modify.c >Hunk #1 FAILED at 35. >Hunk #2 FAILED at 44. >Hunk #3 FAILED at 66. >Hunk #4 FAILED at 88. >Hunk #5 FAILED at 97. >Hunk #6 FAILED at 118. >6 out of 6 hunks FAILED >checking file gallium/drivers/r300/r300_render_translate.c >Hunk #1 FAILED at 51. >1 out of 1 hunk FAILED >checking file mesa/main/colormac.h >Hunk #1 succeeded at 65 with fuzz 2 (offset -27 lines). >Hunk #2 FAILED at 119. >1 out of 2 hunks FAILED >checking file mesa/main/compiler.h >Hunk #1 FAILED at 226. >1 out of 1 hunk FAILED >can't find file to patch at input line 184 >Perhaps you used the wrong -p or --strip option? >The text leading up to this was: >-------------------------- >|diff --git a/src/mesa/main/depthstencil.c b/src/mesa/main/depthstencil.c >|index af5c12f..67553a4 100644 >|--- a/src/mesa/main/depthstencil.c >|+++ b/src/mesa/main/depthstencil.c >-------------------------- >File to patch: >Skip this patch? [y] >Skipping patch. >21 out of 21 hunks ignored >can't find file to patch at input line 432 >Perhaps you used the wrong -p or --strip option? >The text leading up to this was: >-------------------------- >|diff --git a/src/mesa/main/format_pack.c b/src/mesa/main/format_pack.c >|index 390b494..b04663f 100644 >|--- a/src/mesa/main/format_pack.c >|+++ b/src/mesa/main/format_pack.c >-------------------------- >File to patch: >Skip this patch? [y] >Skipping patch. >3 out of 3 hunks ignored >can't find file to patch at input line 484 >Perhaps you used the wrong -p or --strip option? >The text leading up to this was: >-------------------------- >|diff --git a/src/mesa/main/format_unpack.c b/src/mesa/main/format_unpack.c >|index 4f23f3d..ebd17fb 100644 >|--- a/src/mesa/main/format_unpack.c >|+++ b/src/mesa/main/format_unpack.c >-------------------------- >File to patch: >Skip this patch? [y] >Skipping patch. >25 out of 25 hunks ignored >checking file mesa/main/formats.c >Hunk #1 FAILED at 312. >Hunk #2 FAILED at 330. >Hunk #3 FAILED at 348. >Hunk #4 FAILED at 402. >Hunk #5 FAILED at 429. >Hunk #6 FAILED at 438. >Hunk #7 FAILED at 456. >Hunk #8 FAILED at 474. >Hunk #9 FAILED at 638. >Hunk #10 FAILED at 1314. >Hunk #11 FAILED at 2062. >Hunk #12 FAILED at 2093. >Hunk #13 FAILED at 2121. >Hunk #14 FAILED at 2131. >Hunk #15 FAILED at 2141. >Hunk #16 FAILED at 2185. >Hunk #17 succeeded at 1186 (offset -1075 lines). >Hunk #18 FAILED at 2598. >Hunk #19 FAILED at 2620. >Hunk #20 FAILED at 2668. >Hunk #21 FAILED at 2778. >20 out of 21 hunks FAILED >checking file mesa/main/formats.h >Hunk #1 FAILED at 92. >Hunk #2 FAILED at 146. >Hunk #3 FAILED at 242. >3 out of 3 hunks FAILED >checking file mesa/main/texformat.c >Hunk #1 FAILED at 103. >Hunk #2 FAILED at 135. >Hunk #3 FAILED at 149. >Hunk #4 FAILED at 163. >Hunk #5 FAILED at 195. >Hunk #6 FAILED at 213. >Hunk #7 FAILED at 322. >Hunk #8 FAILED at 333. >Hunk #9 FAILED at 344. >Hunk #10 FAILED at 355. >Hunk #11 FAILED at 367. >Hunk #12 FAILED at 759. >Hunk #13 FAILED at 786. >13 out of 13 hunks FAILED >checking file mesa/main/texstore.c >Hunk #1 succeeded at 142 (offset -837 lines). >Hunk #2 FAILED at 1126. >Hunk #3 succeeded at 287 (offset -859 lines). >Hunk #4 FAILED at 1166. >Hunk #5 succeeded at 1271 with fuzz 2 (offset -870 lines). >Hunk #6 FAILED at 2282. >Hunk #7 FAILED at 2297. >Hunk #8 FAILED at 2402. >Hunk #9 FAILED at 2472. >Hunk #10 FAILED at 2514. >Hunk #11 FAILED at 2529. >Hunk #12 FAILED at 2573. >Hunk #13 FAILED at 3328. >Hunk #14 FAILED at 3547. >Hunk #15 FAILED at 3567. >Hunk #16 FAILED at 3598. >Hunk #17 FAILED at 4514. >Hunk #18 FAILED at 4550. >Hunk #19 FAILED at 4576. >16 out of 19 hunks FAILED >checking file mesa/state_tracker/st_cb_clear.c >Hunk #1 FAILED at 389. >Hunk #2 FAILED at 445. >2 out of 2 hunks FAILED >checking file mesa/state_tracker/st_format.c >Hunk #1 succeeded at 53 with fuzz 1 (offset -94 lines). >Hunk #2 FAILED at 203. >Hunk #3 succeeded at 177 with fuzz 2 (offset -88 lines). >Hunk #4 FAILED at 299. >Hunk #5 FAILED at 334. >Hunk #6 FAILED at 475. >Hunk #7 FAILED at 526. >Hunk #8 succeeded at 660 with fuzz 2 (offset 67 lines). >Hunk #9 succeeded at 697 with fuzz 2 (offset 66 lines). >Hunk #10 FAILED at 666. >Hunk #11 FAILED at 804. >Hunk #12 FAILED at 1570. >Hunk #13 FAILED at 1586. >Hunk #14 FAILED at 1602. >10 out of 14 hunks FAILED >rican-linux@lubuntu-powerbook:~/Build/mesa/src$ > > > >-----BEGIN PGP SIGNATURE----- >Version: Mailvelope v1.4.0 >Comment: <a href="https://www.mailvelope.com">https://www.mailvelope.com</a> > >wsFcBAEBCAAQBQJXN//JCRBZfgHS2LQxKAAAdnwP/RacNj1k5rQWFJ50RxSC >DoFeJCxPXspbhXiNXQpCON87ZKWXpphwwXS+/a/BPfi/rfQyLyaK6W34Gmv+ >9eMzTOeteLSzo9H5vHs+GUj48bJsitrsXTZ8yjxrExYdKsqXk5AfOjKVBtOk >waic8mN4k+JughYpTVLvVnqtbslDm9ov4iP+h5u/hF/oIERnzzcKA2UhBHxn >G6zociCVFlWWE/4/tCaWg9CJk0DOkez0Hk83jdMfaiYK4yCseqIktctlA3TM >XI0npSG3zUi0etMwNcmcz6gMQToTYjqoeJTk0ZIi4BeVDgUW858F0OqSkwMm >g3a3+GysP495puqo+UdrJOGVAWzGbPEq8dS8l4k37BHA80znl+KO8jS4TR5V >l7fB3lvrC/h/CP/788vESIDNtQeXH+Zy87EHkoX+7x9+Jg2R7CbAYjacp5rx >qwXEj+CDOW44sqmGldEqgT7gmxhK1EuVuvSqaue5G0XQVTfse1ogzQOz0zGk >GSXuNFeoJLysVX27pDElJ1igY7wX/hliz1Vjcv1Ps9FlwF2Y43FS/1uXagg9 >hiCQ6E+AgFVSiOWQIXHFmUArOM2r44lQF+Fh2vkAqPdp0AgOkGwEATDdJRdK >K4PaXEBodQnXsNLZ5z24KVIpImP+hz76nriGQUE3iS1TVT6E9tXGG7h/6Ssa >6jFM >=rjrD >-----END PGP SIGNATURE----- ><pre></pre></pre></div><div class="gmail_extra"><br><div class="gmail_quote">On Sat, May 14, 2016 at 6:26 AM, <span dir="ltr"><<a href="mailto:bugzilla-daemon@freedesktop.org" target="_blank">bugzilla-daemon@freedesktop.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> > > > > <div> > <p> > </p><div> > <b><a title="NEW - On PPC, OpenGL programs use incorrect texture colors." href="https://bugs.freedesktop.org/show_bug.cgi?id=43698#c20" target="_blank">Comment # 20</a> > on <a title="NEW - On PPC, OpenGL programs use incorrect texture colors." href="https://bugs.freedesktop.org/show_bug.cgi?id=43698" target="_blank">bug 43698</a> > from <span><a href="mailto:erhard_f@mailbox.org" title="erhard_f@mailbox.org" target="_blank">erhard_f@mailbox.org</a> ></span></b> > <pre>Compiled and using Mesa 11.3.0-devel (git-59156b2) now on my PowerMac G5 7,3 >with an AGP Radeon 9600, rv350 class card. > >Tested so far: AssaultCube, Extreme TuxRacer, RedEclipse, Sauerbraten. All >installed from Ubuntu MATE 16.04 repositories. > >Garbled textures: RedEclipse, Sauerbraten. > >Extreme TuxRacer runs and is actually playable! However the colors are wrong >and the backgrounnd sky is flickering constantly. > >AssaultCube crashes after showing the "Loading..." window. See ><a href="https://bugs.launchpad.net/ubuntu/+source/assaultcube/+bug/1581805" target="_blank">https://bugs.launchpad.net/ubuntu/+source/assaultcube/+bug/1581805</a> > >Extreme TuxRacer and Sauerbraten have problems while changing the resolution. >Both get REALLY sluggish then, hardware acceleration does not seem to work any >longer. The console output starts showing "radeon 000:f0:10.0: ring 0 stalled >for more than xxxx msec" messages during these resolution change problems in >Extreme TuxRacer.</pre> > </div> > <p></p><span class=""> > > > <hr> > <span>You are receiving this mail because:</span> > > <ul> > <li>You are on the CC list for the bug.</li> > </ul> > </span></div> ></blockquote></div><br></div> Fix for this bug result in regression: https://bugs.freedesktop.org/show_bug.cgi?id=98869 |
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.
Created attachment 54298 [details] PrBoom With git b1a8b7b0196c73bcfe488cbfc9e9fcd1d7ce7d9b. Texture colors are incorrect, they appear to be ABGR instead of RGBA, thus blue becomes green, red becomes alpha, etc. The only thing that is not affected is glxgears, but that does not use any textures.