Commit graph

76582 commits

Author SHA1 Message Date
Jason Ekstrand
bbebd2de13 nir: Add a helper for getting the bitmask for a variable's location 2015-12-24 00:47:20 -08:00
Jason Ekstrand
4ff4310a78 nir/types: Expose glsl_type::count_attribute_slots() 2015-12-24 00:47:19 -08:00
Jason Ekstrand
0bc1b0fd23 nir/lower_return: Do it for real this time 2015-12-24 00:47:19 -08:00
Jason Ekstrand
e1b1d58bec nir/cf: Make extracting or re-inserting nothing a no-op 2015-12-23 23:46:04 -08:00
Jason Ekstrand
eae352e75c nir: Add a function for comparing cursors 2015-12-23 18:09:42 -08:00
Jason Ekstrand
54c870ff61 nir/spirv: Add support for undefs in vtn_ssa_value() 2015-12-23 14:14:39 -08:00
Jason Ekstrand
2e823d5754 nir/spirv: Properly handle vector times matrix 2015-12-23 13:49:56 -08:00
Jason Ekstrand
452ba4db2b nir/spirv: Create the correct type if a matrix-vector multiply produces a vector 2015-12-23 13:49:56 -08:00
Jason Ekstrand
5b30132388 nir/spirv: Fix some mem_ctx issues with create_vec 2015-12-23 13:49:56 -08:00
Jason Ekstrand
66168a798b nir/spirv: Better document vtn_ssa_value.transposed 2015-12-23 13:49:56 -08:00
Jason Ekstrand
3b391892aa anv/descriptor_set: Use anv_foreach_stage 2015-12-23 13:49:56 -08:00
Jason Ekstrand
72ceb99bab anv: Mask out invalid stages in foreach_stage 2015-12-23 13:49:56 -08:00
Jason Ekstrand
5644b1cece nir/spirv: Handle LogicalNot 2015-12-23 13:49:56 -08:00
Jason Ekstrand
6219a69589 nir/spirv: Handle derefs in vtn_ssa_value
This is kind of a hack, but it makes vtn_ssa_value insert a load if the
value requested is actually a deref.  This shouldn't happen normally but,
thanks to the impedence mismatch of the NIR function parameter model vs.
the SPIR-V model, this can happen for function arguments.
2015-12-23 13:49:56 -08:00
Jason Ekstrand
3ab1b7afa8 nir/spirv: Do boolean fixup on block loads
We used to do it for variable loads on things of type "uniform" but that
never got ported to block loads.
2015-12-23 13:49:56 -08:00
Jason Ekstrand
af74ce5a19 spirv/nir: Handle non-vector extractions in vtn_composite_extract 2015-12-23 13:49:56 -08:00
Jason Ekstrand
79b8b42081 nir/spirv: Handle function calls 2015-12-23 13:49:56 -08:00
Jason Ekstrand
95990c96cc nir: Create the params array in function_impl_create 2015-12-23 13:49:56 -08:00
Jason Ekstrand
a7f3e113ad i965/nir: Remove return handling
This was added because we were getting spurrious returns coming out of
SPIR-V.  Now that we're calling lower_returns, we don't need this.
2015-12-23 13:49:56 -08:00
Jason Ekstrand
ac975b73cf anv/pipeline: Run lower_returns and inline_functions after spirv_to_nir 2015-12-23 13:49:56 -08:00
Jason Ekstrand
8fba4bf79f nir: Add a function inlining pass 2015-12-23 13:49:56 -08:00
Jason Ekstrand
b21db9cea5 nir/builder: Add a copy_deref_var helper 2015-12-23 13:49:56 -08:00
Jason Ekstrand
23cfa683d5 nir: move nir_copy_var from anv_nir_builder to nir_builder 2015-12-23 13:49:56 -08:00
Jason Ekstrand
4aac03fe61 nir/clone: Add support for cloning a single function_impl
This will be useful for things such as function inlining.
2015-12-23 13:49:56 -08:00
Jason Ekstrand
98291b8f2c nir: Add a helper for creating a "bare" nir_function_impl
This is useful if you want to clone a single function_impl if, for
instance, you wanted to do function inlining.
2015-12-23 13:49:56 -08:00
Jason Ekstrand
86772c2488 nir/control_flow: Handle relinking top-level blocks
This can happen if a function ends in a return instruction and you remove
the return.
2015-12-23 13:49:56 -08:00
Jason Ekstrand
1749e667ea nir: Add a stub function inlining pass
All it does is remove the return at the end, but it's good enough for
simple functions.
2015-12-23 13:49:56 -08:00
Jason Ekstrand
413a9d3517 nir/print: Factor variable name lookup into a helper
Otherwise, we have a problem when we go to print functions with arguments
because their names get added to the hash table during declaration which
happens after we print the prototype.
2015-12-23 13:49:56 -08:00
Kristian Høgsberg Kristensen
220ac9337b vk: Only require wc bo mmap for !llc GPUs 2015-12-19 22:25:57 -08:00
Kristian Høgsberg Kristensen
b49aaf5de0 vk: Remove stale 48 bit addresses FIXMEs
This has worked fine for a long time.
2015-12-19 22:20:45 -08:00
Kristian Høgsberg Kristensen
c4802bc44c vk/gen8: Implement VkEvent for gen8
We use PIPE_CONTROL for setting and resetting the event from cmd buffers
and MI_SEMAPHORE_WAIT in polling mode for waiting on an event.
2015-12-19 22:17:19 -08:00
Kristian Høgsberg Kristensen
8ac46d84ff vk: Fix check for I915_PARAM_MMAP_VERSION
Comparing the wrong thing for < 1.
2015-12-18 17:24:19 -08:00
Jordan Justen
5e82a91324 anv/gen8: Add support for gl_NumWorkGroups
Co-authored-by: Kristian Høgsberg <krh@bitplanet.net>
Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
2015-12-18 01:45:11 -08:00
Jason Ekstrand
d7f66f9f6f nir/spirv: Array lengths are constants not literals 2015-12-17 16:36:29 -08:00
Jason Ekstrand
1473a8dc6f anv/formats: Add more 64-bit formats 2015-12-17 13:51:09 -08:00
Jason Ekstrand
167809365b anv/formats: Add more PACK32 formats 2015-12-17 13:44:50 -08:00
Jason Ekstrand
952bf05897 anv/image: Properly report buffer features 2015-12-17 11:52:31 -08:00
Jason Ekstrand
3395ca17d1 isl: Add a is_storage_image_format helper 2015-12-17 11:45:04 -08:00
Jason Ekstrand
b1325404c5 anv/device: Handle zero-sized memory allocations 2015-12-17 11:00:38 -08:00
Jason Ekstrand
c643e9cea8 anv/state: Allow levelCount to be 0
This can happen if the client is creating an image view of a textureable
surface and they only ever intend to render to that view.
2015-12-16 17:34:57 -08:00
Jason Ekstrand
b2fe8b4673 nir/spirv: Add a missing break statement 2015-12-15 17:24:18 -08:00
Jason Ekstrand
1c51d91bfe anv/pipeline: Allow the user to pass a null MultisampleCreateInfo
According to section 5.2 of the Vulkan spec, this is allowed for color-only
rendering pipelines.
2015-12-15 16:26:10 -08:00
Jason Ekstrand
d61ff1ed08 anv/descriptor_set: Initialize immutable_samplers to NULL
Previously this wasn't a problem.  However, with the new API update,
descriptor sets can now be sparse so the client doesn't have to provide an
entry for every binding.  This means that it's possible for a binding to be
uninitialized other than the memset.  In that case, we want to have a null
array of immutable samplers.
2015-12-15 16:24:22 -08:00
Jason Ekstrand
28c4ef9d6c anv/device: Bump the size of the instruction block pool
Some CTS test shaders were failing to compile.  At some point soon, we
really need to make a real pipeline cache and stop using a block pool for
this.
2015-12-15 11:49:28 -08:00
Jason Ekstrand
306abbead3 anv/pipeline: Properly set IncludeVertexHandles in 3DSTATE_GS 2015-12-15 11:37:18 -08:00
Jason Ekstrand
2d4b7eda23 nir/spirv: Add support for more CS intrinsics 2015-12-15 10:20:23 -08:00
Jason Ekstrand
1035108a7f nir/lower_system_values: Add support for computed builtins.
In particular, this commit adds support for computing gl_GlobalInvocationID
and gl_LocalInvocationIndex from other intrinsics.
2015-12-15 10:20:23 -08:00
Jason Ekstrand
630b9528b3 shader_enums: Add enums for gl_GlobalInvocationID and gl_LocalInvocationIndex 2015-12-15 10:20:23 -08:00
Jason Ekstrand
7ebd84fa4b nir/lower_system_values: Refactor and use the builder.
Now that we have a helper in the builder for system values and a helper in
core NIR to get the intrinsic opcode, there's really no point in having
things split out into a helper function.  This commit "modernizes" this
pass to use helpers better and look more like newer passes.
2015-12-15 10:20:23 -08:00
Jason Ekstrand
c26e889a44 nir/builder: Add a load_system_value helper
While we're at it, go ahead and make nir_lower_clip use it.

Cc: Rob Clark <robclark@gmail.com>
2015-12-15 10:20:23 -08:00