Changes

Summary

  1. [AMDGPU] Propagate LDS align into to instructions (details)
  2. [mlir][linalg] Adapt the FillOp builder signature. (details)
  3. [AMDGPU] Simplify collectReachableCallees. NFCI. (details)
  4. [SCEV] Add tests with single-cond range check generated by InstComb. (details)
  5. [SCEV] Support signed predicates in applyLoopGuards. (details)
  6. [llvm] Update tests that got missed in adee485adf84ae8a. (details)
  7. [mlir][linalg] Change the FillOp library call signature. (details)
  8. [IR] Simplify createReplacementInstr (details)
  9. [AMDGPU] Stop using LegacyLegalizerInfo. NFCI. (details)
  10. [mlir][LLVMIR] Fold ExtractValueOp coming from InsertValueOp (details)
  11. [LLParser] Remove special handling for call address space (details)
  12. [TTI] Make assertion compatible with opaque pointers (details)
Commit 2b43209ee37ce27cede17cf5424e7655adfe3ac2 by Stanislav.Mekhanoshin
[AMDGPU] Propagate LDS align into to instructions

Differential Revision: https://reviews.llvm.org/D104316
The file was modifiedllvm/test/CodeGen/AMDGPU/lower-kernel-lds-global-uses.ll
The file was modifiedllvm/test/CodeGen/AMDGPU/ds_read2.ll
The file was modifiedllvm/test/CodeGen/AMDGPU/lower-module-lds.ll
The file was modifiedllvm/lib/Target/AMDGPU/AMDGPULowerModuleLDSPass.cpp
The file was modifiedllvm/test/CodeGen/AMDGPU/lower-module-lds-used-list.ll
The file was modifiedllvm/test/CodeGen/AMDGPU/replace-lds-by-ptr-lds-offsets.ll
The file was modifiedllvm/test/CodeGen/AMDGPU/lower-kernel-and-module-lds.ll
The file was modifiedllvm/test/CodeGen/AMDGPU/ds_write2.ll
The file was modifiedllvm/test/CodeGen/AMDGPU/lower-kernel-lds-constexpr.ll
The file was modifiedllvm/test/CodeGen/AMDGPU/lower-kernel-lds-super-align.ll
The file was modifiedllvm/test/CodeGen/AMDGPU/lower-kernel-lds.ll
Commit 7cef24ee83cd85b0d7917e7d574287615a717e44 by gysit
[mlir][linalg] Adapt the FillOp builder signature.

Change the build operand order from output, value to value, output. The patch makes the argument order consistent with the pretty printed order updated by https://reviews.llvm.org/D104356.

Differential Revision: https://reviews.llvm.org/D104359
The file was modifiedmlir/lib/Conversion/TosaToLinalg/TosaToLinalg.cpp
The file was modifiedmlir/lib/Dialect/Vector/VectorTransforms.cpp
The file was modifiedmlir/lib/Dialect/Linalg/IR/LinalgOps.cpp
The file was modifiedmlir/test/lib/Dialect/Linalg/TestLinalgTransforms.cpp
The file was modifiedmlir/lib/Dialect/Linalg/Transforms/Transforms.cpp
The file was modifiedmlir/lib/Dialect/Linalg/Transforms/Bufferize.cpp
The file was modifiedmlir/include/mlir/Dialect/Linalg/IR/LinalgStructuredOps.td
The file was modifiedmlir/lib/Dialect/Linalg/Transforms/Promotion.cpp
The file was modifiedmlir/lib/Dialect/Linalg/Transforms/Vectorization.cpp
Commit c65f3f562b15588a444322afe3f2cd3693681730 by jay.foad
[AMDGPU] Simplify collectReachableCallees. NFCI.

Don't use SCC iterators when we're only interested in reachability.
Use df_begin/df_end inline to find reachable nodes.

Differential Revision: https://reviews.llvm.org/D104704
The file was modifiedllvm/lib/Target/AMDGPU/Utils/AMDGPULDSUtils.cpp
Commit 5ab96fa16b31c0f019da63ae2bce3820e672d897 by flo
[SCEV] Add tests with single-cond range check generated by InstComb.
The file was modifiedllvm/test/Analysis/ScalarEvolution/max-backedge-taken-count-guard-info.ll
Commit adee485adf84ae8af4105ab4e0b62753245f420d by flo
[SCEV] Support signed predicates in applyLoopGuards.

This adds handling for signed predicates, similar to how unsigned
predicates are already handled.

Reviewed By: nikic

Differential Revision: https://reviews.llvm.org/D104732
The file was modifiedllvm/test/Analysis/Delinearization/a.ll
The file was modifiedllvm/test/Analysis/ScalarEvolution/sext-mul.ll
The file was modifiedllvm/lib/Analysis/ScalarEvolution.cpp
The file was modifiedllvm/test/Analysis/ScalarEvolution/max-trip-count-address-space.ll
The file was modifiedllvm/test/Analysis/ScalarEvolution/max-backedge-taken-count-guard-info.ll
The file was modifiedllvm/test/Analysis/ScalarEvolution/max-trip-count.ll
Commit aa58fdb3960bf1ea25c2c088fba96700cb7e7071 by flo
[llvm] Update tests that got missed in adee485adf84ae8a.
The file was modifiedllvm/test/Transforms/LoopVersioningLICM/loopversioningLICM2.ll
The file was modifiedllvm/test/Transforms/LoopStrengthReduce/X86/nested-loop.ll
Commit f1844f15c1ad54b78f2d84087df4b51fe5f703f6 by gysit
[mlir][linalg] Change the FillOp library call signature.

Adapt the FillOp library call signature to the updated operand order introduced in https://reviews.llvm.org/D10412. The patch reverts the special treatment of FillOp in LinalgToStandard.

Differential Revision: https://reviews.llvm.org/D104360
The file was modifiedmlir/include/mlir/Dialect/Linalg/IR/LinalgOps.h
The file was modifiedmlir/lib/Conversion/LinalgToStandard/LinalgToStandard.cpp
Commit 157473a58f02b8f2ad12ecbaaa1af32d0342257b by jay.foad
[IR] Simplify createReplacementInstr

NFCI, although the test change shows that ConstantExpr::getAsInstruction
is better than the old implementation of createReplacementInstr because
it propagates things like the sdiv "exact" flag.

Differential Revision: https://reviews.llvm.org/D104124
The file was modifiedclang/test/CodeGenCUDA/managed-var.cu
The file was modifiedllvm/lib/IR/ReplaceConstant.cpp
Commit dfb8c0873984e24ac98a41a372dbe31d8feb629e by jay.foad
[AMDGPU] Stop using LegacyLegalizerInfo. NFCI.

Differential Revision: https://reviews.llvm.org/D103684
The file was modifiedllvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp
Commit f0d43a29e30da8f2758e17e1c242917288ab9a0b by nicolas.vasilache
[mlir][LLVMIR] Fold ExtractValueOp coming from InsertValueOp

Differential Revision: https://reviews.llvm.org/D104769
The file was modifiedmlir/include/mlir/Dialect/LLVMIR/LLVMOps.td
The file was addedmlir/test/Dialect/LLVMIR/canonicalize.mlir
The file was modifiedmlir/lib/Dialect/LLVMIR/IR/LLVMDialect.cpp
Commit 3ee6f1a4fa83fa1e737324e77c27d3cc967d36bc by nikita.ppv
[LLParser] Remove special handling for call address space

Spin-off from D104740: I don't think this special handling is needed
anymore. Calls in textual IR are annotated with addrspace(N) (which
defaults to the program address space from data layout) and specifies
the expected pointer address space of the callee. There is no need
to special-case the program address space on top of that, as it
already is the default expected address space, and we shouldn't
allow use of the program address space if the call was explicitly
annotated with some other address space.

The IsCall parameter is retained because it will be used again soon.

Differential Revision: https://reviews.llvm.org/D104752
The file was modifiedllvm/lib/AsmParser/LLParser.cpp
Commit cfb1cb4491d7aac218d9fd903a69b46bbc21118b by nikita.ppv
[TTI] Make assertion compatible with opaque pointers

Dropping the TODO here because it applies to all uses of this method.
The file was modifiedllvm/include/llvm/Analysis/TargetTransformInfoImpl.h