Messages in this thread | | | Date | Thu, 14 Aug 2008 14:43:10 -0700 | From | "H. Peter Anvin" <> | Subject | Re: [PATCH] x86 alternatives : fix LOCK_PREFIX race with preemptible kernel and CPU hotplug |
| |
Andi Kleen wrote: >> So, unless these calculus are completely bogus, the difference between the nop >> and the DS case seems not to be statistically significant. > > My understanding was that the nops only make a measurable difference in the > -pg/mcount/ftrace patched out case where every function call starts with a 5 > byte nop. > > Presumably function calls are much more common than lock prefix operations.
Yeah, although the LOCK prefix was measurable.
-hpa
| |