svn commit: r280759 - head/sys/netinet
Adrian Chadd
adrian at freebsd.org
Sun Mar 29 21:15:44 UTC 2015
> To illustrate the probability of 1 instruction race, I would suggest to
> look at PCPU_ADD() implementation for all arches except amd64/i386 and
> find out that it is prone to this race. Hitting the race in PCPU_ADD()
> will usually lead to mangling vm_meter statistics. Did we ever observe
> this in practice? No.
I dunno. Has anyone actually looked at that? :)
On a busy system you may find those small race windows exist enough to
be annoying. Case in point - the recent callout code race. The race
between "read" "modify" "write" was tiny, but enough to break things.
-adrian
More information about the svn-src-head
mailing list