[llvm-commits] [llvm] r43353 - /llvm/trunk/lib/CodeGen/BranchFolding.cpp

[Evan Cheng]

Hi Bill,

Is the XXX marker used anywhere else in llvm? I think we use FIXME or  
TODO everywhere else. It's not that important, but it makes sense to  
be somewhere consistent so we can grep it later.

Please also add an entry to README.txt under CodeGen.

Evan

On Oct 25, 2007, at 11:23 AM, Bill Wendling wrote:

> Author: void
> Date: Thu Oct 25 13:23:45 2007
> New Revision: 43353
>
> URL: http://llvm.org/viewvc/llvm-project?rev=43353&view=rev
> Log:
> Added comment explaining why we are doing this check.
>
> Modified:
>    llvm/trunk/lib/CodeGen/BranchFolding.cpp
>
> Modified: llvm/trunk/lib/CodeGen/BranchFolding.cpp
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/CodeGen/BranchFolding.cpp?rev=43353&r1=43352&r2=43353&view=diff
>
> = 
> = 
> = 
> = 
> = 
> = 
> = 
> = 
> ======================================================================
> --- llvm/trunk/lib/CodeGen/BranchFolding.cpp (original)
> +++ llvm/trunk/lib/CodeGen/BranchFolding.cpp Thu Oct 25 13:23:45 2007
> @@ -272,6 +272,11 @@
>   while (I1 != MBB1->begin() && I2 != MBB2->begin()) {
>     --I1; --I2;
>     if (!I1->isIdenticalTo(I2) ||
> +        // XXX: This check is dubious. It's used to get around a  
> problem where
> +        // people incorrectly expect inline asm directives to  
> remain in the same
> +        // relative order. This is untenable because normal compiler
> +        // optimizations (like this one) may reorder and/or merge  
> these
> +        // directives.
>         I1->getOpcode() == TargetInstrInfo::INLINEASM) {
>       ++I1; ++I2;
>       break;
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits