[LLVMdev] Spill Interval Generation Question

[Evan Cheng]

Fixed:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of- 
Mon-20071008/054323.html

Evan
On Oct 7, 2007, at 10:50 PM, Evan Cheng wrote:

>
> On Oct 5, 2007, at 5:52 PM, David Greene wrote:
>
>>
>> In LiveIntervalAnalysis::addIntervalsForSpills, we have this:
>>
>>           if (HasUse) {
>>             LiveRange LR(getLoadIndex(index), getUseIndex(index),
>>                          nI.getNextValue(~0U, 0, VNInfoAllocator));
>>             DOUT << " +" << LR;
>>             nI.addRange(LR);
>>           }
>>
>> Should this actually extend from the load index to the def index  
>> since
>> value intervals are exclusive at end?  Otherwise, the interval for
>> the spill
>> doesn't include the use of the reloaded register, which doesn't
>> make sense to
>> me.
>
> You are probably right. Let me do some testing and fix it if that's
> the case.
>
>>
>> I think this probably went unnoticed because nothing happens with
>> intervals
>> after spill code generation in the current codebase.
>
> Yep. this will only trip over an iterative coalescer.
>
> Evan
>
>>
>>                                               -Dave
>> _______________________________________________
>> LLVM Developers mailing list
>> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
>> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
>
> _______________________________________________
> LLVM Developers mailing list
> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev