On Mon, Nov 21, 2011 at 9:49 AM, Ben Pfaff <b...@nicira.com> wrote: > On Sat, Nov 19, 2011 at 10:20:24AM -0800, Jesse Gross wrote: >> We currently use a seqcount to prevent reading partial 64-bit stats >> on 32-bit CPUs. u64_stats_sync uses the same logic but elides it on >> 64-bit and uniprocessor machines. This improves performance (primarily >> on non-x86 architectures) at the cost of not guaranteeing that packet >> and byte counts were necessarily read together. >> >> Suggested-by: Stephen Hemminger <shemmin...@vyatta.com> >> Signed-off-by: Jesse Gross <je...@nicira.com> > > Acked-by: Ben Pfaff <b...@nicira.com> > > (These acks are going to look silly going upstream, maybe we should > switch to "Looks good." for the datapath :-)
These patches will get squashed before I send another round upstream, so it won't show up. I suppose we can solve the problem when we get to it... > I don't see anything that documents how one initializes a > u64_stats_sync. I guess all-zero-bits initialization is sufficient, > and it looks like percpu data is initially zeroed. Both of those statements are true so it's definitely correct. I don't think that there's an "official" way to initialize it though. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
