[PATCH] Fix PR 23525 - Separate header mass propagation in irregular loops.

[Duncan P. N. Exon Smith]

> On 2015-Jun-10, at 10:31, Xinliang David Li <xinliangli at gmail.com> wrote:
> 
> 
> 
> On Wed, Jun 10, 2015 at 10:19 AM, Diego Novillo <dnovillo at google.com> wrote:
> 
> 
> On Wed, Jun 10, 2015 at 1:06 PM, Xinliang David Li <xinliangli at gmail.com> wrote:
> This illustrates the problem mentioned by Duncan : the frequency of SeventyFive should be close to half of the entry (which is 0.5), but the computed frequency is too low (on the other hand, without the fix, it is too high). Again this only applies to short trip counted irreducible loops.
> 
> 
> This is due to the patch not taking incoming masses into account.  I could address this in the same patch, but I think I'd rather do it separately.
> 
> sgtm.

Okay, I finally caught up with the rest of the thread.  Looks like
David noticed the same thing I did.  (Sorry for my delay coming back to
this.)

I agree that doing it separately is fine, as long as there aren't any
regressions in LNT.  I doubt there will be, since I don't think we have
much irreducible control flow there.