Re: [ovs-dev] [thread 12/15] ofproto-dpif: Lock rules to prevent eviction.

2013-08-09 Thread Ben Pfaff
On Fri, Aug 09, 2013 at 02:00:31PM -0700, Ethan Jackson wrote: > > The new failure case in ofproto_delete_flow() looks like a problem for > > fail-open, which never retries because previously it could never fail. > > Is there a reason we can't block waiting for the write lock here? The > > uses of

Re: [ovs-dev] [thread 12/15] ofproto-dpif: Lock rules to prevent eviction.

2013-08-09 Thread Ethan Jackson
> The new failure case in ofproto_delete_flow() looks like a problem for > fail-open, which never retries because previously it could never fail. > Is there a reason we can't block waiting for the write lock here? The > uses of ofproto_delete_flow() (in-band, fail-open) are not performance > sensi

Re: [ovs-dev] [thread 12/15] ofproto-dpif: Lock rules to prevent eviction.

2013-08-08 Thread Ben Pfaff
On Thu, Aug 08, 2013 at 06:33:02PM -0700, Ethan Jackson wrote: > > Well this is partially true. Previously if someone had a learn action > > the flow tables could change transiently. At any rate my thinking is > > that if we're in a situation where we've got inconsistent actions > > because of a

Re: [ovs-dev] [thread 12/15] ofproto-dpif: Lock rules to prevent eviction.

2013-08-08 Thread Ethan Jackson
> Well this is partially true. Previously if someone had a learn action > the flow tables could change transiently. At any rate my thinking is > that if we're in a situation where we've got inconsistent actions > because of a transiently broken flow table, a will be happening soon "a revalidatio

Re: [ovs-dev] [thread 12/15] ofproto-dpif: Lock rules to prevent eviction.

2013-08-08 Thread Ethan Jackson
> Previously, the processing of a given packet or flow always saw a > consistent set of flow tables. That is, the flow tables didn't change > from one lookup (e.g. resubmit) to another. I am not sure that this > is true any longer. Is it now possible for a flow_mod to take effect > in the middle

Re: [ovs-dev] [thread 12/15] ofproto-dpif: Lock rules to prevent eviction.

2013-08-08 Thread Ben Pfaff
On Thu, Aug 08, 2013 at 03:51:05PM -0700, Ben Pfaff wrote: > On Thu, Aug 08, 2013 at 12:58:31PM -0700, Ethan Jackson wrote: > > This patch uses a read-write lock to prevent rules from being evicted > > while they're used by child threads. > > > > Signed-off-by: Ethan Jackson > > I don't think an

Re: [ovs-dev] [thread 12/15] ofproto-dpif: Lock rules to prevent eviction.

2013-08-08 Thread Ben Pfaff
On Thu, Aug 08, 2013 at 12:58:31PM -0700, Ethan Jackson wrote: > This patch uses a read-write lock to prevent rules from being evicted > while they're used by child threads. > > Signed-off-by: Ethan Jackson I don't think anything ever destroys 'evict'. (But POSIX says that only an unlocked rwlo