Messages in this thread | | | Date | Thu, 28 Feb 2013 16:15:45 -0800 | From | "H. Peter Anvin" <> | Subject | Re: [PATCH] mm/x86: Patch out arch_flush_lazy_mmu_mode() when running on bare metal |
| |
On 02/28/2013 04:05 PM, Josh Boyer wrote: > On Fri, Mar 01, 2013 at 12:37:20AM +0100, Borislav Petkov wrote: >> On Thu, Feb 28, 2013 at 05:55:49PM -0500, Boris Ostrovsky wrote: >>> Invoking arch_flush_lazy_mmu_mode() results in calls to >>> preempt_enable()/disable() which may have performance impact. >>> >>> Since lazy MMU is not used on bare metal we can patch away >>> arch_flush_lazy_mmu_mode() so that it is never called in such >>> environment. >>> >>> Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> >> >> Looks straight-forward enough to me. >> >> Acked-by: Borislav Petkov <bp@suse.de> > > I'll try to get someone to test this tomorrow. >
Sounds good.
I'd like look at this after the merge window close. I'm kind of bothered about having a choice with an oops on PV, a performance regression on native, or putting a pretty complex patch in as a fix, but since we'll be right after -rc1 it is probably reasonable to pick the last option.
-hpa
| |