#
#asahi-gpu · 2026-03-08 08:59
chadmedim giving up on >2 planes again for now. im not sure what it is we could possibly be doing so differently to macos that causes this to not work
#asahi-gpu · 2026-03-08 08:15
jannauwe can just document a different stride for interchange
#asahi-gpu · 2026-03-08 08:14
jannauwe can't really break user-space with the interchange modifier
#asahi-gpu · 2026-03-08 08:02
jannaumacos sets stride to width (already a multiple of 16) * 16 * blocksize for &b38
#asahi-gpu · 2026-03-08 08:00
chadmedunless drm_info is lying
#asahi-gpu · 2026-03-08 08:00
chadmedit doesnt seem to care at all if theyre not multiples of 64 though. currently have a cursor displays just fine with stride = 132b
#asahi-gpu · 2026-03-08 07:58
chadmedlots of magenta and black
#asahi-gpu · 2026-03-08 07:58
chadmedi tried rounding them up just before and it exploded pretty badly
#asahi-gpu · 2026-03-08 07:56
jannaudoes dcp actually care about the stride for tiled formats?
#asahi-gpu · 2026-03-08 07:56
jannauI think we should change the comment but changing dcp isn't out of question
#asahi-gpu · 2026-03-08 07:55
jannauoliver: does https://paste.centos.org/view/4881ba18 help?
#asahi-gpu · 2026-03-08 07:38
chadmedwhy would swap_enabled ever be set to >0xf when there is a surface wtf
#asahi-gpu · 2026-03-08 07:31
jannauI doubt there is user space (ignoring mesa) which has code specific to apple modifiers and much less so code written to explicitly considering that comment
#asahi-gpu · 2026-03-08 07:29
chadmedwe cannot break userspace, the uapi has been upstreamed
#asahi-gpu · 2026-03-08 07:27
oliverWe can now either "break" userspace and change the comment (and adjust mesa too) or update the atomic_check and only verify stride for linear, because we know tiled will have proper stride. Which one should it be?
#asahi-gpu · 2026-03-08 07:03
oliverAnd due to that userspace sets pitch based on the width, not rounded up like jannau suggested.
#asahi-gpu · 2026-03-08 06:57
oliverThe pitch is actually 64 bytes aligned, it just does not match the comment.
#asahi-gpu · 2026-03-08 06:48
jannaudcp very much needs (or needed) 64-byte aligned pitch, see 0f27ba328875947a513
#asahi-gpu · 2026-03-08 06:47
jannausince the tiles are 16x16 with a 4-byte element size the pitch should be 64-byte aligned
#asahi-gpu · 2026-03-08 06:43
chadmedi commented out the check and cursors work again. if this is a limitation purely on the agx side then it should not spill into the dcp driver and mesa or the asahi driver needs to check for this
#asahi-gpu · 2026-03-08 06:40
jannaurounding the stride up to the next multiple of 64
#asahi-gpu · 2026-03-08 06:40
chadmedi actually dont think dcp cares too much. the stride check in plane_atomic_check is relatively new, and linear worked fine before we added it
#asahi-gpu · 2026-03-08 06:39
oliverWhat do you guys thing is the best way to go here? Only check the stride for the linear modifier?
#asahi-gpu · 2026-03-08 06:37
oliverThis then fails in atomic_check for textured that do not have a stride of a multiple of 64 bytes.
#asahi-gpu · 2026-03-08 06:36
oliver * stride is arbitrary here, we pick the simplest convention.
#asahi-gpu · 2026-03-08 06:36
oliver * therefore require that stride equals (width * bytes per pixel). Since the
#asahi-gpu · 2026-03-08 06:36
oliver * software assumes the stride is at least (width * bytes per pixel). We
#asahi-gpu · 2026-03-08 06:36
oliver * convention would be setting the image stride to 0. Unfortunately, some
#asahi-gpu · 2026-03-08 06:36
oliver * Userspace aware of the image layouts do not use strides. The most "correct"
#asahi-gpu · 2026-03-08 06:36
oliver * will be unable to use correctly the specified stride for any purpose.
#asahi-gpu · 2026-03-08 06:36
oliver * specify strides for these layouts, userspace unaware of Apple image layouts
#asahi-gpu · 2026-03-08 06:36
oliverAlyssa comments on the tiled modifiers: These layouts fundamentally do not have meaningful strides. No matter how we
#asahi-gpu · 2026-03-01 02:27
karolherbstand PIPE_SHADER_SUBGROUP_FEATURE_MASK (and GL is using PIPE_SHADER_SUBGROUP_NUM_FEATURES)
#asahi-gpu · 2026-03-01 02:26
karolherbstyeah.. I've added the PIPE_SHADER_SUBGROUP_FEATURE_ROTATE defines for CL specifically
#asahi-gpu · 2026-03-01 02:20
alyssathen probably you can merge
#asahi-gpu · 2026-03-01 02:20
alyssaoh
#asahi-gpu · 2026-03-01 02:13
karolherbstalyssa: on the gl side? Maybe.. Though rotate isn't used there and not sure if there is even a gl ext for it anyway
#asahi-gpu · 2026-03-01 02:12
alyssakarolherbst: probably should get cts'd..
#asahi-gpu · 2026-03-01 01:08
karolherbstanybody want to review that trivial subgroup rotate patch for asahi? or should I just merge it 🙃 https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37169/diffs?commit_id=11b15fa4cf88ab776e3c504ddb44d8784a3aee87
#asahi-gpu · 2026-02-28 14:44
jannauthat looks simple enough
#asahi-gpu · 2026-02-28 14:04
alyssa& no I'm not getting sniped into writing it :p
#asahi-gpu · 2026-02-28 14:04
alyssano uapi change needed, should be just a small amount of code
#asahi-gpu · 2026-02-28 14:00
alyssashould be straightforward for asahi
#asahi-gpu · 2026-02-28 14:00
alyssajannau: 06b41ca5899 ("iris: add support for fence signal capability") enables both exts
#asahi-gpu · 2026-02-28 08:14
ffxcant remove
#asahi-gpu · 2026-02-28 08:10
jannauI'm not sure if the bug is related and your comment isn't really helpful
#asahi-gpu · 2026-02-28 08:04
ffxjannau thanks, your comments pointed me in the right direction finally
#asahi-gpu · 2026-02-28 08:01
ffxhttps://bugzilla.mozilla.org/show_bug.cgi?id=1994071
#asahi-gpu · 2026-02-28 07:59
ffxjannau ok so this is a pure firefox / browser vendor problem?
#asahi-gpu · 2026-02-28 07:56
ffxthought that gfxInfo is part of the os

plus ›