diff options
| author |   Damjan Marion <damarion@cisco.com> | 2018-08-23 09:49:46 +0200 |
|---|---|---|
| committer |   Neale Ranns <nranns@cisco.com> | 2018-08-23 14:36:15 +0000 |
| commit | ca3ff1a25f62af33d7966726811f714a0c30a431 (patch) | |
| tree | 94a54c3ed462133bef818cdb772a926cbb5c6969 /src/vpp-api.am | |
| parent | 184870ac5a266c37987e4a4d97ab4d4efefacb1f (diff) | |
ip4: rework rewrite node to use vlib_buffer_enqueue_to_next()
- perf improvement is around 3 clock per packet
- it properly prefetches ip header and rewrite area
Credits to zhiyong.yang@intel.com for spotting that in some cases
rewrite area is not prefetched.
Change-Id: Ie02913de7bd7f42b7df2617fb5fa87c74ab53c23
Signed-off-by: Damjan Marion <damarion@cisco.com>
Diffstat (limited to 'src/vpp-api.am')
0 files changed, 0 insertions, 0 deletions