I tested again with some different parameters, and this actually made the performance little worse, so I’m holding off on this patch for now.
Jarno On Feb 11, 2014, at 5:43 PM, Ben Pfaff <b...@nicira.com> wrote: > On Tue, Feb 11, 2014 at 04:30:49PM -0800, Jarno Rajahalme wrote: >> The code setting wc masks shows up in 'perf' reports. This patch streamlines >> the code and avoids setting the mask bits for registers or metadata, as >> those fields do not exist on datapaths. >> >> Signed-off-by: Jarno Rajahalme <jrajaha...@nicira.com> > > In mf_mask_subfield_wc(), for MFF_IN_PORT and MFF_IN_PORT_OXM, I would > set the whole field unconditionally to all-1-bits. (We don't allow > bitwise matching on input ports and I hope we never do.) Same for > MFF_ETH_TYPE, although bitwise matching there is a little less crazy. > > There's an extra space on this line: > bitwise_one (&value, field->n_bytes, sf->ofs, sf->n_bits); > > Acked-by: Ben Pfaff <b...@nicira.com> _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
