Acked-by: Jarno Rajahalme
Thanks for the fix and happy new year!
Jarno
On Jan 3, 2015, at 11:53 AM, Ethan Jackson wrote:
> Before this patch, dp_netdev_flow_add() inserted newly minted flows in
> the "flow_table" cmap before inserting them into the per core "dpcls"
> classifier. Since dpcl
I think this was the crash that sunk us last time. We're in a really
good position at this point. The automated test runner is pretty
stable and I'm getting some really nice consistent numbers. I should
have some graphs shortly.
Also btw with the perf stuff. Since the NSDI deadline the datapat
Before this patch, dp_netdev_flow_add() inserted newly minted flows in
the "flow_table" cmap before inserting them into the per core "dpcls"
classifier. Since dpcls_insert() initializes 'flow->cr.mask', there's
a brief window where the flow is accessible from the cmap, but has a
bogus mask value.
