GitLab

This was trying to emulate the effect of mapping GL -> TGSI -> NIR,
but failed to handle VARYING_SLOT_PNTC which led to a kludgy workaround
in freedreno.

Part-of: <mesa/mesa!4732>

If we don't use the pre-draw flag here, we won't end up transitioning
these until some time after the draw, which isn't what we want.

These flags are never set, so let's simplify the logic a bit here.

This gets rid of some repeated code.

This is necessary to avoid having derefs with uniform mode, which
lower_explicit_io would turn into address computations.

The OpenCL image_width/height/depth functions have variants which can
take an LOD parameter.  More importantly, LLVM-SPIRV-Translator always
generates OpImageQuerySizeLod even if the LOD is guaranteed to be zero.
Given that over half the hardware out there has an LOD field for image
size queries (based on a rudimentary scan through their NIR -> whatever
code), we may as well just add the source to the NIR intrinsic.  If this
is ever a problem for anyone, the lowering is pretty trivial.

I've also added asserts to everyone's drivers that should alert them if
they ever see an LOD other than zero.  This will never happen with GL or
Vulkan so there's no need for panic.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

This reverts commit d9d4a1ae.

Delete unintentionally left debugging code

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <mesa/mesa!6472>

nir_shader::num_inputs isn't supposed to be a count of how many input
variables we have.  It's a size of the lowered input space.
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <mesa/mesa!6472>

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <mesa/mesa!6472>

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <mesa/mesa!6472>

This commit propagates the alignment information provided either through
the Alignment decoration on pointers or via the alignment mem operands
to OpLoad, OpStore, and OpCopyMemory to the NIR deref chain.  It does so
by wrapping the deref in a cast.  NIR should be able to clean up most
unnecessary casts only leaving us with the useful alignment information.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <mesa/mesa!6472>

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <mesa/mesa!6472>

If we have a cast deref with alignment information and we can get equal
or better alignment information from something further up the deref
chain, we can strip the alignment information from the cast and allow
other optimizations to potentially eliminate the cast.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <mesa/mesa!6472>

Generally, if a cast has alignment information, that information is
useful and we don't want to loose it.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <mesa/mesa!6472>