Re: [PATCH xserver 02/10 v2] glamor: Require that 16bpp pixmap depths match for Render copies.
On 26/09/16 03:53 AM, Eric Anholt wrote: > Michel Dänzerwrites: >> On 23/09/16 09:51 PM, Eric Anholt wrote: >>> >>> We do need some concerted effort on actually fixing our rendering bugs >>> and reenabling the skipped tests. I've spent a while trying to come up >>> with why the remaining rendercheck test fails and come up with nothing >>> yet. >> >> How can others help with this? [...] > and rendercheck/cacomposite/all/a8: FWIW, this test passes with radeonsi. The other tests you mentioned also have failures with radeonsi, but different ones. -- Earthling Michel Dänzer | http://www.amd.com Libre software enthusiast | Mesa and X developer ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH xserver 02/10 v2] glamor: Require that 16bpp pixmap depths match for Render copies.
Michel Dänzerwrites: > On 23/09/16 09:51 PM, Eric Anholt wrote: >> Michel Dänzer writes: >>> On 23/09/16 04:57 PM, Eric Anholt wrote: The copy optimization in d37329cba42fa8e72fe4be8a7be18e512268b5bd replicated a bug from last time we did a copy optimization, and didn't get rendercheck run on it. >>> >>> Actually, I'm pretty sure I did run rendercheck, but didn't notice the >>> regression due to the radeonsi bug affecting the same test. Please >>> consider removing that language from the commit log. >> >> OK. I'll drop the comment about running rendercheck. > > Thanks. > > >> We do need some concerted effort on actually fixing our rendering bugs >> and reenabling the skipped tests. I've spent a while trying to come up >> with why the remaining rendercheck test fails and come up with nothing >> yet. > > How can others help with this? I fixed the remaining glamor-specific case in the scripts on the plane. Now we're down to: # Skip some tests that are failing at the time of importing the script. #"REPORT: min_bounds, rbearing was 0, expecting 2" PIGLIT_ARGS="$PIGLIT_ARGS -x xlistfontswithinfo@3" PIGLIT_ARGS="$PIGLIT_ARGS -x xlistfontswithinfo@4" PIGLIT_ARGS="$PIGLIT_ARGS -x xloadqueryfont@1" PIGLIT_ARGS="$PIGLIT_ARGS -x xqueryfont@1" PIGLIT_ARGS="$PIGLIT_ARGS -x xqueryfont@2" #"REPORT: Incorrect pixel on inside of area at point (0, 0): 0x0 != 0x1" PIGLIT_ARGS="$PIGLIT_ARGS -x xgetimage@7" PIGLIT_ARGS="$PIGLIT_ARGS -x xgetsubimage@7" in the current XTS set across fb and glamor/llvmpipe. XGetImage/7 seems to be BadMatching. I have no idea about the font thing, but it doesn't seem too rendering-related either. So, at this point, I think it's up to us to go bang on our hardware 3D drivers to make sure that they're rendering glamor correctly. Sending out a script to make that easier, even if we're not going to put it in make check. For i965, I'm seeing failures in: rendercheck/blend/over: Beginning blend test on a4r4g4b4 Over blend test error of 15. at (0, 0) -- R G B A got: 0.000 0.000 0.000 0.000 expected: 0.000 0.000 0.000 1.000 src color: 0.00 0.00 0.00 1.00 (a8) dst color: 0.00 0.00 0.00 1.00 src: 1x1R a8, dst: a4r4g4b4 Beginning blend test on a4b4g4r4 Over blend test error of 15. at (0, 0) -- R G B A got: 0.000 0.000 0.000 0.000 expected: 0.000 0.000 0.000 1.000 src color: 0.00 0.00 0.00 1.00 (a8) dst color: 0.00 0.00 0.00 1.00 src: 1x1R a8, dst: a4b4g4r4 and rendercheck/blend/src: Beginning blend test on b8g8r8a8 Src blend test error of 255. at (0, 0) -- R G B A got: 0.000 0.000 0.000 0.000 expected: 0.000 0.000 0.000 1.000 src color: 0.00 0.00 0.00 1.00 (a8) dst color: 0.00 0.00 0.00 1.00 src: 1x1R a8, dst: b8g8r8a8 Beginning blend test on a4r4g4b4 Src blend test error of 15. at (0, 0) -- R G B A got: 0.000 0.000 0.000 0.000 expected: 0.000 0.000 0.000 1.000 src color: 0.00 0.00 0.00 1.00 (a8) dst color: 0.00 0.00 0.00 1.00 src: 1x1R a8, dst: a4r4g4b4 Beginning blend test on a4b4g4r4 Src blend test error of 15. at (0, 0) -- R G B A got: 0.000 0.000 0.000 0.000 expected: 0.000 0.000 0.000 1.000 src color: 0.00 0.00 0.00 1.00 (a8) dst color: 0.00 0.00 0.00 1.00 src: 1x1R a8, dst: a4b4g4r4 and rendercheck/cacomposite/all/a8: Ignoring server-supported format: x2b10g10r10 Beginning composite CA mask test on a8 Saturate CA composite test error of 128. at (0, 0) -- R G B A got: 0.000 0.000 0.000 0.498 expected: 0.000 0.000 0.000 1.000 src color: 0.00 0.00 0.00 1.00 msk color: 0.00 0.00 0.00 0.50 dst color: 0.00 0.00 0.00 1.00 src: 1x1R a8, mask: 1x1R a8, dst: a8 Beginning composite CA mask test on a8r8g8b8 DisjointIn CA composite test error of 255. at (0, 0) -- R G B A got: 0.000 0.000 0.000 0.000 expected: 0.000 0.000 0.000 1.000 src color: 0.00 0.00 0.00 1.00 msk color: 0.00 0.00 0.00 1.00 dst color: 0.00 0.00 0.00 1.00 and rendercheck/triangles: Beginning Dst Triangles test on a8 triangles test error of 255. at (0, 0) -- R G B A got: 0.000 0.000 0.000 0.000 expected: 0.000 0.000 0.000 1.000 ... signature.asc Description: PGP signature ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH xserver 02/10 v2] glamor: Require that 16bpp pixmap depths match for Render copies.
On 23/09/16 09:51 PM, Eric Anholt wrote: > Michel Dänzerwrites: >> On 23/09/16 04:57 PM, Eric Anholt wrote: >>> The copy optimization in d37329cba42fa8e72fe4be8a7be18e512268b5bd >>> replicated a bug from last time we did a copy optimization, and didn't >>> get rendercheck run on it. >> >> Actually, I'm pretty sure I did run rendercheck, but didn't notice the >> regression due to the radeonsi bug affecting the same test. Please >> consider removing that language from the commit log. > > OK. I'll drop the comment about running rendercheck. Thanks. > We do need some concerted effort on actually fixing our rendering bugs > and reenabling the skipped tests. I've spent a while trying to come up > with why the remaining rendercheck test fails and come up with nothing > yet. How can others help with this? >>> This is effectively a re-cherry-pick of >>> 510c8605641803f1f5b5d2de6d3bb422b148e0e7. >>> >>> Fixes rendercheck -t blend -o src -f x4r4g4b4,x3r4g4b4 >>> >>> v2: Drop excessive src->depth == dst->depth check that snuck in. >>> >>> Signed-off-by: Eric Anholt >>> --- >>> >>> Michel, I didn't apply your review becuase you said "second clause", >>> but I removed the first of the two. >> >> Is there any case where the drawable depths don't match, but a copy does >> what's expected? > > I asked keithp, and he agreed that CopyArea may not be used to copy > between depths. Render is totally defined across depths, though. Right, which is why I thought the first clause was necessary and the second one redundant. (I suspect some cases such as copying between depth 24 and 32 may work out correctly, but it's probably better to use a whitelist approach for those rather than a blacklist for all cases which don't work correctly) -- Earthling Michel Dänzer | http://www.amd.com Libre software enthusiast | Mesa and X developer signature.asc Description: OpenPGP digital signature ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH xserver 02/10 v2] glamor: Require that 16bpp pixmap depths match for Render copies.
Michel Dänzerwrites: > On 23/09/16 04:57 PM, Eric Anholt wrote: >> The copy optimization in d37329cba42fa8e72fe4be8a7be18e512268b5bd >> replicated a bug from last time we did a copy optimization, and didn't >> get rendercheck run on it. > > Actually, I'm pretty sure I did run rendercheck, but didn't notice the > regression due to the radeonsi bug affecting the same test. Please > consider removing that language from the commit log. OK. I'll drop the comment about running rendercheck. We do need some concerted effort on actually fixing our rendering bugs and reenabling the skipped tests. I've spent a while trying to come up with why the remaining rendercheck test fails and come up with nothing yet. >> This is effectively a re-cherry-pick of >> 510c8605641803f1f5b5d2de6d3bb422b148e0e7. >> >> Fixes rendercheck -t blend -o src -f x4r4g4b4,x3r4g4b4 >> >> v2: Drop excessive src->depth == dst->depth check that snuck in. >> >> Signed-off-by: Eric Anholt >> --- >> >> Michel, I didn't apply your review becuase you said "second clause", >> but I removed the first of the two. > > Is there any case where the drawable depths don't match, but a copy does > what's expected? I asked keithp, and he agreed that CopyArea may not be used to copy between depths. Render is totally defined across depths, though. signature.asc Description: PGP signature ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH xserver 02/10 v2] glamor: Require that 16bpp pixmap depths match for Render copies.
On 23/09/16 04:57 PM, Eric Anholt wrote: > The copy optimization in d37329cba42fa8e72fe4be8a7be18e512268b5bd > replicated a bug from last time we did a copy optimization, and didn't > get rendercheck run on it. Actually, I'm pretty sure I did run rendercheck, but didn't notice the regression due to the radeonsi bug affecting the same test. Please consider removing that language from the commit log. > This is effectively a re-cherry-pick of > 510c8605641803f1f5b5d2de6d3bb422b148e0e7. > > Fixes rendercheck -t blend -o src -f x4r4g4b4,x3r4g4b4 > > v2: Drop excessive src->depth == dst->depth check that snuck in. > > Signed-off-by: Eric Anholt> --- > > Michel, I didn't apply your review becuase you said "second clause", > but I removed the first of the two. Is there any case where the drawable depths don't match, but a copy does what's expected? -- Earthling Michel Dänzer | http://www.amd.com Libre software enthusiast | Mesa and X developer ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel