[PATCH] D110558: [DebugInfo][LoopStrengthReduction] Prevent attempts to translate long SCEV expressions

[Orlando Cazalet-Hyams via Phabricator via llvm-commits]

Orlando added a comment.

Just a few nits inline. I otherwise defer to @StephenTozer's reviewing.



================
Comment at: llvm/test/Transforms/LoopStrengthReduce/pr51656.ll:78
+  %sub = sub nsw i32 %0, %mul.47, !dbg !40
+  ;; CHECK: call void @llvm.dbg.value(metadata !DIArgList(i32 undef, i32 %mul.47), metadata !20, metadata !DIExpression(DW_OP_LLVM_arg, 0, DW_OP_LLVM_arg, 1, DW_OP_minus, DW_OP_LLVM_convert, 32, DW_ATE_signed, DW_OP_LLVM_convert, 64, DW_ATE_signed, DW_OP_stack_value))
+  call void @llvm.dbg.value(metadata i32 %sub, metadata !20, metadata !DIExpression(DW_OP_LLVM_convert, 32, DW_ATE_signed, DW_OP_LLVM_convert, 64, DW_ATE_signed, DW_OP_stack_value)), !dbg !41
----------------
nit: I think the convention is to use a single `;` for FileCheck directive comments.


================
Comment at: llvm/test/Transforms/LoopStrengthReduce/pr51656.ll:2
+; RUN: opt -loop-reduce -S %s | FileCheck %s
+; REQUIRES: x86-registered-target
+
----------------
Is there a reason for this test to be X86 specific?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D110558/new/

https://reviews.llvm.org/D110558