[all-commits] [llvm/llvm-project] 83879f: [SimplifyCFG] Don't block sinking for allocas if n...

[Nikita Popov via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 83879f4f5311af334550c54c8279397a8aa33e7b
      https://github.com/llvm/llvm-project/commit/83879f4f5311af334550c54c8279397a8aa33e7b
  Author: Nikita Popov <npopov at redhat.com>
  Date:   2024-08-19 (Mon, 19 Aug 2024)

  Changed paths:
    M llvm/lib/Transforms/Utils/SimplifyCFG.cpp
    M llvm/test/CodeGen/Hexagon/block-addr.ll
    M llvm/test/DebugInfo/ARM/single-constant-use-preserves-dbgloc.ll
    M llvm/test/Transforms/SimplifyCFG/X86/sink-common-code.ll
    M llvm/test/Transforms/SimplifyCFG/sink-and-convert-switch.ll

  Log Message:
  -----------
  [SimplifyCFG] Don't block sinking for allocas if no phi created (#104579)

SimplifyCFG sinking currently does not sink loads/stores of allocas,
because historically SROA was unable to handle the resulting IR. Since
then, SROA both learned to speculate loads/stores over selects and phis,
*and* SimplifyCFG sinking has been deferred to the end of the function
simplification pipeline, which means that SROA happens before it.

As such, I believe that this workaround should no longer be necessary.
Given how sensitive SimplifyCFG sinking seems to be, this patch takes a
very conservative step towards removing this, by allowing sinking if we
don't actually need to form a phi over the pointer argument.

This fixes https://github.com/llvm/llvm-project/issues/104567, where
sinking a store to an escaped alloca allows converting a switch into
arithmetic.



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications