[PATCH] D21231: [StackProtector] Fix computation of GSCookieOffset and EHCookieOffset with SEH4

[Reid Kleckner via llvm-commits]

rnk added a comment.

The direction is consistent with what we've done before, but in retrospect, I kinda wish we did this in MI rather than IR. Changing that is out of scope though.

I'm fine with this if David majnemer is fine with this.


http://reviews.llvm.org/D21231