[llvm-commits] [llvm] r67139 - in /llvm/trunk: lib/Transforms/IPO/DeadArgumentElimination.cpp test/Transforms/DeadArgElim/2009-03-17-MRE-Invoke.ll

[Bill Wendling]

Chris,

These were causing XPASSes (on Darwin) in these tests:

Sending        test/FrontendC/2009-03-13-dbg.c
Sending        test/FrontendC++/2009-02-16-CtorNames-dbg.cpp
Sending        test/FrontendC++/2009-03-17-dbg.cpp

Is this expected? I modified them to make them passing. Please check
to see if that's correct.

Thanks!
-bw

On Tue, Mar 17, 2009 at 5:31 PM, Chris Lattner <sabre at nondot.org> wrote:
> Author: lattner
> Date: Tue Mar 17 19:31:45 2009
> New Revision: 67139
>
> URL: http://llvm.org/viewvc/llvm-project?rev=67139&view=rev
> Log:
> Fix PR3807 by inserting 'insertelement' instructions in the normal dest of
> an invoke instead of after the invoke (in its block), which is invalid.
>
> Added:
>    llvm/trunk/test/Transforms/DeadArgElim/2009-03-17-MRE-Invoke.ll
> Modified:
>    llvm/trunk/lib/Transforms/IPO/DeadArgumentElimination.cpp
>
> Modified: llvm/trunk/lib/Transforms/IPO/DeadArgumentElimination.cpp
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/IPO/DeadArgumentElimination.cpp?rev=67139&r1=67138&r2=67139&view=diff
>
> ==============================================================================
> --- llvm/trunk/lib/Transforms/IPO/DeadArgumentElimination.cpp (original)
> +++ llvm/trunk/lib/Transforms/IPO/DeadArgumentElimination.cpp Tue Mar 17 19:31:45 2009
> @@ -798,9 +798,13 @@
>         // Replace by null for now.
>         Call->replaceAllUsesWith(Constant::getNullValue(Call->getType()));
>       } else {
> -        assert(isa<StructType>(RetTy) && "Return type changed, but not into a"
> -                                         "void. The old return type must have"
> -                                         "been a struct!");
> +        assert(isa<StructType>(RetTy) &&
> +               "Return type changed, but not into a void. The old return type"
> +               " must have been a struct!");
> +        Instruction *InsertPt = Call;
> +        if (InvokeInst *II = dyn_cast<InvokeInst>(Call))
> +          InsertPt = II->getNormalDest()->begin();
> +
>         // We used to return a struct. Instead of doing smart stuff with all the
>         // uses of this struct, we will just rebuild it using
>         // extract/insertvalue chaining and let instcombine clean that up.
> @@ -813,12 +817,13 @@
>             if (RetTypes.size() > 1)
>               // We are still returning a struct, so extract the value from our
>               // return value
> -              V = ExtractValueInst::Create(New, NewRetIdxs[i], "newret", Call);
> +              V = ExtractValueInst::Create(New, NewRetIdxs[i], "newret",
> +                                           InsertPt);
>             else
>               // We are now returning a single element, so just insert that
>               V = New;
>             // Insert the value at the old position
> -            RetVal = InsertValueInst::Create(RetVal, V, i, "oldret", Call);
> +            RetVal = InsertValueInst::Create(RetVal, V, i, "oldret", InsertPt);
>           }
>         // Now, replace all uses of the old call instruction with the return
>         // struct we built
>
> Added: llvm/trunk/test/Transforms/DeadArgElim/2009-03-17-MRE-Invoke.ll
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/DeadArgElim/2009-03-17-MRE-Invoke.ll?rev=67139&view=auto
>
> ==============================================================================
> --- llvm/trunk/test/Transforms/DeadArgElim/2009-03-17-MRE-Invoke.ll (added)
> +++ llvm/trunk/test/Transforms/DeadArgElim/2009-03-17-MRE-Invoke.ll Tue Mar 17 19:31:45 2009
> @@ -0,0 +1,15 @@
> +; RUN: llvm-as < %s | opt -deadargelim | llvm-dis
> +; PR3807
> +
> +define internal { i32, i32 } @foo() {
> +  ret {i32,i32} {i32 42, i32 4}
> +}
> +
> +define i32 @bar() {
> +  %x = invoke {i32,i32} @foo() to label %T unwind label %T2
> +T:
> +  %y = extractvalue {i32,i32} %x, 1
> +  ret i32 %y
> +T2:
> +  unreachable
> +}
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>