[PATCH] D32980: Use the frame index side table for byval and inalloca arguments

[Reid Kleckner via Phabricator via llvm-commits]

rnk marked 2 inline comments as done.
rnk added a comment.

In https://reviews.llvm.org/D32980#749141, @aprantl wrote:

> Is there a testcase for the example that you mention in the description?


The dbg-declare-inalloca.ll is a more real-world example of the case in the description. You need a non-trivially copyable argument to trigger inalloca.

> Please keep a close eye on the green dragon bots, both the ones in the debuginfo-tests repository that execute as part of each check-clang, as well as the lldb testsuite, after landing this.

Sure!



================
Comment at: llvm/test/CodeGen/X86/dbg-baseptr.ll:62
+; DWARF:   DW_TAG_formal_parameter
+; DWARF-NEXT:     DW_AT_location [DW_FORM_exprloc]      (<0x2> 91 10 )
+; DWARF-NEXT:     DW_AT_name [DW_FORM_strp]     ( {{.*}}"input")
----------------
aprantl wrote:
> please add the disassembly as a comment: (such as DW_OP_fbreg +16 )
I understand how this is supposed to work now. I added more CHECKs to ensure that this DWARF is consistent with the asm.


================
Comment at: llvm/test/DebugInfo/X86/dbg-declare-inalloca.ll:105
+
+attributes #0 = { nounwind "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="pentium4" "target-features"="+fxsr,+mmx,+sse,+sse2,+x87" "unsafe-fp-math"="false" "use-soft-float"="false" }
+attributes #1 = { nounwind readnone speculatable }
----------------
aprantl wrote:
> We typically delete all unnecessary attributes from testcases.
My intention is that the .ll is verbatim output from clang on the commented C++ at the beginning of the file. Is there any reason to treat reducing away attributes differently from the DI metadata, which is mostly retained below?


https://reviews.llvm.org/D32980