[PATCH] D86319: [Attributor] Introduce module slice.

[Johannes Doerfert via Phabricator via llvm-commits]

jdoerfert added a comment.

I went to some tests and left some comments. tomorrow we go through the rest.



================
Comment at: llvm/test/Transforms/Attributor/ArgumentPromotion/X86/attributes.ll:1
-; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --function-signature --scrub-attributes
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --function-signature --scrub-attributes --check-attributes
 ; RUN: opt -attributor -enable-new-pm=0 -attributor-manifest-internal  -attributor-max-iterations-verify -attributor-annotate-decl-cs -attributor-max-iterations=4 -S < %s | FileCheck %s --check-prefixes=CHECK,NOT_CGSCC_NPM,NOT_CGSCC_OPM,NOT_TUNIT_NPM,IS__TUNIT____,IS________OPM,IS__TUNIT_OPM
----------------
Please pre-commit the additional flags and updates to the tests as NFC (no review)


================
Comment at: llvm/test/Transforms/Attributor/ArgumentPromotion/alignment.ll:19
+; IS________NPM-NEXT:    call void @g(i32 [[TMP0]])
+; IS________NPM-NEXT:    ret void
 ;
----------------
This is the expected result, the CGSCC and module pass are producing the same in NPM and OPM respectively. 


================
Comment at: llvm/test/Transforms/Attributor/ArgumentPromotion/alignment.ll:122
+; IS__CGSCC_NPM:       Return2:
+; IS__CGSCC_NPM-NEXT:    ret i32 [[A]]
 ;
----------------
It's really important that argument privatization works now in the CGSCC pass! yay!


================
Comment at: llvm/test/Transforms/Attributor/IPConstantProp/musttail-call.ll:100
 ; IS__CGSCC____-NEXT:    [[I1:%.*]] = call i32 @external()
-; IS__CGSCC____-NEXT:    ret i8* null
+; IS__CGSCC____-NEXT:    unreachable
 ;
----------------
This looks wrong though. Need to verify.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D86319/new/

https://reviews.llvm.org/D86319