FailedChanges

Summary

  1. Temporarily Revert "[clangd] Add Random Forest runtime for code completion." (details)
  2. [test][InstrProf] Fix always_inline.ll under NPM (details)
  3. [instcombine][x86] Converted pdep/pext with shifted mask to simple arithmetic (details)
  4. [clang][module] Improve incomplete-umbrella warning (details)
Commit 549e55b3d5634870aa9d42135f51ad46a6a0e347 by echristo
Temporarily Revert "[clangd] Add Random Forest runtime for code completion."
as a header doesn't appear to have made it into the commit.

This reverts commit 9b6765e784b39c88cb8cdb85ab083e6c95a997ed and followup
The file was removedclang-tools-extra/clangd/unittests/DecisionForestTests.cpp
The file was modifiedclang-tools-extra/clangd/unittests/CodeCompleteTests.cpp (diff)
The file was removedclang-tools-extra/clangd/unittests/decision_forest_model/features.json
The file was removedclang-tools-extra/clangd/unittests/decision_forest_model/forest.json
The file was removedclang-tools-extra/clangd/quality/CompletionModel.cmake
The file was removedclang-tools-extra/clangd/quality/README.md
The file was removedclang-tools-extra/clangd/quality/CompletionModelCodegen.py
The file was removedclang-tools-extra/clangd/quality/model/features.json
The file was removedclang-tools-extra/clangd/unittests/decision_forest_model/CategoricalFeature.h
The file was modifiedclang-tools-extra/clangd/CMakeLists.txt (diff)
The file was removedclang-tools-extra/clangd/quality/model/forest.json
The file was modifiedclang-tools-extra/clangd/unittests/CMakeLists.txt (diff)
Commit 7c10129f5a2145cf8f6dbe259269fd2a781a8dbe by aeubanks
[test][InstrProf] Fix always_inline.ll under NPM

NPM's inliner does not clean up dead functions.

Differential Revision: https://reviews.llvm.org/D87922
The file was modifiedllvm/test/Instrumentation/InstrProfiling/always_inline.ll (diff)
Commit 06f136f61e6d23fde5c91f7fa0813d0291c17c97 by listmail
[instcombine][x86] Converted pdep/pext with shifted mask to simple arithmetic

If the mask of a pdep or pext instruction is a shift masked (i.e. one contiguous block of ones) we need at most one and and one shift to represent the operation without the intrinsic. One all platforms I know of, this is faster than the pdep/pext.

The cost modelling for multiple contiguous blocks might be worth exploring in a follow up, but it's not relevant for my current use case. It would almost certainly be a win on AMDs where these are really really slow though.

Differential Revision: https://reviews.llvm.org/D87861
The file was modifiedllvm/lib/Target/X86/X86InstCombineIntrinsic.cpp (diff)
The file was modifiedllvm/test/Transforms/InstCombine/X86/x86-bmi-tbm.ll (diff)
Commit ed79827aea444e6995fb3d36abc2bfd36331773c by zixu_wang
[clang][module] Improve incomplete-umbrella warning

Change the warning message for -Wincomplete-umbrella to report the location of the umbrella header;

Differential Revision: https://reviews.llvm.org/D82118
The file was modifiedclang/test/Modules/incomplete-umbrella.m (diff)
The file was modifiedclang/lib/Lex/PPLexerChange.cpp (diff)