Messages in this thread | ![/](/images/icornerl.gif) | | Date | Thu, 16 Jun 2011 15:00:39 -0700 | From | Andi Kleen <> | Subject | Re: REGRESSION: Performance regressions from switching anon_vma->lock to mutex |
| |
> Ok, I bet it doesn't make them a non-issue, but if doing this in > anon_vma_clone() helped a lot, then doing the exact same pattern in > unlink_anon_vmas() hopefully helps some more. Hi Linus,
I did essentially the same thing (just in a less elegant, more paranoid way) in my old batching patches.
http://comments.gmane.org/gmane.linux.kernel.mm/63130
They made it a bit better, but overall it's still much worse than 2.6.35 (.36 introduced the root chain locking). I guess we'll see if it can recover mutex though, but I'm not optimistic.
-Andi
| ![\](/images/icornerr.gif) |