On 10/11/2018 12:12 PM, Kenneth Graunke wrote: > On Thursday, October 11, 2018 11:58:40 AM PDT Kenneth Graunke wrote: >> On Tuesday, October 2, 2018 9:16:01 AM PDT asimiklit.w...@gmail.com wrote: >>> From: Andrii Simiklit <andrii.simik...@globallogic.com> >>> >>> I guess that when we calculating the width0, height0, depth0 >>> to use for function 'intel_miptree_create' we need to consider >>> the 'base level' like it is done in the 'intel_miptree_create_for_teximage' >>> function. >>> >>> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107987 >>> Signed-off-by: Andrii Simiklit <andrii.simik...@globallogic.com> >>> --- >>> .../drivers/dri/i965/intel_tex_validate.c | 26 ++++++++++++++++++- >>> 1 file changed, 25 insertions(+), 1 deletion(-) >> >> I believe this patch is correct - we're assembling things into a new >> miptree, which we start at level 0 - so we need the sizes for level 0. >> >> Alternatively, we might be able to pass validate_first_level instead >> of 0 when calling intel_miptree_create, to make one that's only good >> up until the new base...and have to re-assemble it the next time they >> change the base. It would save memory potentially. But more copies. >> I don't have a strong preference which is better. >> >> Please do make a Piglit or dEQP test for this. >> >> Reviewed-by: Kenneth Graunke <kenn...@whitecape.org> > > Sorry, withdrawing my review. :( Chris Forbes pointed out on IRC that > your reproducer case is backwards: > > miplevel 0 - 1x1 > miplevel 1 - 2x2 > miplevel 2 - 4x4 > > That's upside down. A proper miptree would have the base be largest:
It's upside down, but it's perfectly valid in two scenarios. 1. You set the level selection (not the LOD clamps) to a single level. GL_TEXTURE_BASE_LEVEL == GL_TEXTURE_MAX_LEVEL. Based on the Chromium bug that Tapani referenced in the fd.o bug, I suspect the original WebGL test is changing at least GL_TEXTURE_BASE_LEVEL. 2. You're not using a mipmap min filter. It looks like his test case uses GL_LINEAR, so it should be fine. It's a real corner case, but that's what most of the WebGL tests exercise. :( > miplevel 0 - 4x4 > miplevel 1 - 2x2 > miplevel 2 - 1x1 > > So, yes, I could see this tripping an assert...but such a crazy texture > will never be mipmap complete. If they're expecting mipmapping, then > it seems like they should get a fallback black texture (which normally > happens for incomplete textures). If not, maybe they should get a > single miplevel? Either way, seems like we should detect insanity and > bail, rather than change size calculations for the normal sane case. > > _______________________________________________ > mesa-dev mailing list > mesa-dev@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/mesa-dev
signature.asc
Description: OpenPGP digital signature
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
