[PATCH] Move the segmented stack switch to a function attribute

[Alex Crichton]

ping?

On Wed, Apr 2, 2014 at 4:48 PM, Alex Crichton <alex at crichton.co> wrote:
> Oh dear, it's also probably useful if I attach the patch, sorry about that!
>
> On Wed, Apr 2, 2014 at 4:33 PM, Alex Crichton <alex at crichton.co> wrote:
>> Awhile back, Luqman Aden made a request [1] to merge a
>> "no-split-stack" function attribute for disabling the split-stack
>> prologue in an otherwise segmented-stack-enabled program. I believe it
>> was later recommended to instead remove the command line switch and
>> use an opt-in system instead (with a "split-stack" attribute for
>> functions) [2].
>>
>> I asked Luqman if it was ok for me to rebase his patch and try to
>> re-submit it, and he said that I had the green light. I've rebased his
>> work from [2] on today's LLVM, and addressed a few comments that were
>> raised on the review.
>>
>> I've only recently found the phabricator instance for LLVM, so if it
>> would be more helpful to move the patch onto there, please just let me
>> know!
>>
>> [1] - http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20131007/190913.html
>> [2] - http://llvm-reviews.chandlerc.com/D1905