[llvm-commits] [llvm] r72507 - in /llvm/trunk: include/llvm/Target/TargetLowering.h lib/CodeGen/SelectionDAG/DAGCombiner.cpp lib/Target/X86/X86ISelLowering.cpp lib/Target/X86/X86ISelLowering.h test/CodeGen/X86/narrow_op-1.ll test/CodeGen/X86/narrow_op-2.ll

[Evan Cheng]

On May 28, 2009, at 11:09 AM, Duncan Sands wrote:

> Hi Evan,
>
>>> Hi Evan, nice transform!  But I think you shouldn't do it if the
>>> original store was volatile.  Also, mightn't this turn an aligned
>>
>> Why not? This would just reduce the store width.
>
> yes, that's unvalid if you are writing into i/o memory for example.
> The very fact of writing to an address can have an effect.  This
> is the case even if the value you are writing is the value that
> you just read from it.

Very well. The semantics of "volatile" is somewhat overloaded, but  
it's better to be prudent.

Thanks,

Evan

>
> Ciao,
>
> Duncan.
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits