Add clip distances validation for maxInterStageShaderVariables
#8762
Conversation
ErichDonGubler
force-pushed
the
max_inter_stage_shader_variables-clip_distances
branch
3 times, most recently
from
2004d9a to
f70488b
Compare
ErichDonGubler
force-pushed
the
max_inter_stage_shader_variables-clip_distances
branch
2 times, most recently
from
9a8ea87 to
152fedd
Compare
ErichDonGubler
force-pushed
the
max_inter_stage_shader_variables-clip_distances
branch
7 times, most recently
from
6454d73 to
4a816bf
Compare
| ViewIndex, | ||
| BaseInstance, | ||
| BaseVertex, | ||
| ClipDistance { array_size: u32 }, |
There was a problem hiding this comment.
Thoughts on putting array_size: Option<u32> in the naga type rather then duplicating the whole thing?
There was a problem hiding this comment.
The Naga type seems like its usage starts for classifying attributes, which only really makes sense as a flat set of unit variants. Naga might want to grow another representation for attributes and all relevant information, but it doesn't currently have that, and that would be strictly more general than what this is trying to accomplish.
CC @jimblandy and/or @teoxoy for thoughts on this, non-blockingly.
There was a problem hiding this comment.
It doesn't seem most of the variants are used. There is also quite a bit of overlap with InterStageBuiltIn, which despite the name is used for fragment inputs.
ErichDonGubler
force-pushed
the
max_inter_stage_shader_variables-clip_distances
branch
from
4a816bf to
ca04a30
Compare
Shortly, we're going to add our own functionality to it.
ErichDonGubler
force-pushed
the
max_inter_stage_shader_variables-clip_distances
branch
from
ca04a30 to
6a58c70
Compare
ErichDonGubler
force-pushed
the
max_inter_stage_shader_variables-clip_distances
branch
2 times, most recently
from
5b49daf to
672262f
Compare
ErichDonGubler
force-pushed
the
max_inter_stage_shader_variables-clip_distances
branch
from
672262f to
07c352f
Compare
ErichDonGubler
force-pushed
the
max_inter_stage_shader_variables-clip_distances
branch
from
07c352f to
e7966f5
Compare
|
The last blocker for merging I haven't figured out yet is GL failures on Linux. CTS appears to believe that limits are wrong, for some reason that I haven't investigated yet. |
|
@ErichDonGubler Can you update this to mention that it closes #6236? |
4 participants
Connections
maxInterStageShaderComponentstomaxInterStageShaderVariables(minus coverage ofclip-distances) #8652.Testing
New CTS tests have been enabled that cover this functionality, which were previously failing.
Squash or Rebase?
Rebase.