[llvm] r242173 - MachineRegisterInfo: Remove UsedPhysReg infrastructure

[Matthias Braun via llvm-commits]

I just committed r245329 to llvm ToT which should fix the issue. However I can't test this myself.

- Matthias

> On Aug 13, 2015, at 5:32 PM, Quentin Colombet via llvm-commits <llvm-commits at lists.llvm.org> wrote:
> 
> 
>> On Aug 13, 2015, at 4:52 PM, Tom Stellard <tom at stellard.net <mailto:tom at stellard.net>> wrote:
>> 
>> On Thu, Aug 13, 2015 at 03:53:06PM -0700, Quentin Colombet via llvm-commits wrote:
>>> 
>>>> On Aug 13, 2015, at 3:20 PM, Marek Olšák <maraeo at gmail.com <mailto:maraeo at gmail.com>> wrote:
>>>> 
>>>> On Fri, Aug 14, 2015 at 12:09 AM, Quentin Colombet <qcolombet at apple.com <mailto:qcolombet at apple.com>> wrote:
>>>>> Hi Marek,
>>>>> 
>>>>> I will take care of it, Matthias is on vacation.
>>>>> 
>>>>> Sorry for not noticing earlier.
>>>>> 
>>>>>> On Aug 13, 2015, at 3:06 PM, Marek Olšák via llvm-commits <llvm-commits at lists.llvm.org <mailto:llvm-commits at lists.llvm.org>> wrote:
>>>>>> 
>>>>>> Hi,
>>>>>> 
>>>>>> Just a reminder: We really need to revert this in the 3.7 branch ASAP.
>>>>> 
>>>>> Should I revert on trunk as well?
>>>> 
>>>> Hi Quentin,
>>>> 
>>>> That depends on whether or not the regression will be fixed there in
>>>> the near future.
>>> 
>>> Ok, I’ll wait for Matthias to look into it before deciding.
>>> 
>>> My guess is that the fix will likely be in the target specific part, not here.
>>> 
>>> Anyway, I have reverted this on release_37: r244976.
>>> 
>> 
>> I opened an audit for this commit, and I think I've spotted the problem:
> 
> I think you’re right Tom, the iterator used now does not consider super/sub registers.
> 
> Not sure what was the motivation for this commit, so I leave the final word to Matthias on how he would like to fix that.
> Iterating on the regunits for every target sounds like a lot of code duplication and even factorizing this functionality somewhere does not sound super efficient.
> 
> Anyway, Matthias will follow-up when he gets back on Monday IIRC.
> 
> Cheers,
> -Quentin
> 
>> 
>> http://reviews.llvm.org/rL242173 <http://reviews.llvm.org/rL242173>
>> 
>> -Tom
>> 
>>> Cheers,
>>> -Quentin
>>> 
>>>> 
>>>> Marek
>>> 
>>> _______________________________________________
>>> llvm-commits mailing list
>>> llvm-commits at lists.llvm.org <mailto:llvm-commits at lists.llvm.org>
>>> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits <http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150818/5ed98e73/attachment.html>