[llvm] r216119 - X86: Align the stack on word boundaries in LowerFormalArguments()

[Duncan P. N. Exon Smith]

+tjablin

> On 2014-Aug-21, at 15:42, Reid Kleckner <rnk at google.com> wrote:
> 
> How was this not handled by r214980?

It was.  Reverted in r216246.

I looked carefully at this on August 4th (before r214980), and when it
popped back up I didn't check that the testcase still failed.

Thomas: I left your testcase in tree (and it still passes).  Can you
confirm that r214980 is sufficient?