[llvm] r279417 - [LTO] Handles commons in monolithic LTO

Teresa Johnson via llvm-commits llvm-commits at lists.llvm.org
Wed Aug 31 07:10:17 PDT 2016


Thanks, Rafael. Right, this is essentially what gold-plugin was doing
before the new API. I will take a look at whether we can put something
back to avoid the RAUW (might not be until next week though). Is the
deficiency in the gold-plugin API or in gold's resolution? Do you know
why gold isn't already selecting the largest as prevailing?

Teresa

On Wed, Aug 31, 2016 at 7:00 AM, Rafael EspĂ­ndola via llvm-commits
<llvm-commits at lists.llvm.org> wrote:
> On 29 August 2016 at 17:07, Rafael EspĂ­ndola <rafael.espindola at gmail.com> wrote:
>>> +    // Common resolution: collect the maximum size/alignment.
>>> +    // FIXME: right now we ignore the prevailing information, it is not clear
>>> +    // what is the "right" behavior here.
>>
>> I think the correct behaviour is to expect the linker to keep as
>> prevailing the GV with the largest size and for the linker to update
>> the alignment.
>>
>> The rationale is that is that keeping the largest avoids the need for
>> RAUW since the alignment can be set in place.
>
> I just implemented this in lld (r280242).
>
> Cheers,
> Rafael
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits



-- 
Teresa Johnson | Software Engineer | tejohnson at google.com | 408-460-2413


More information about the llvm-commits mailing list