numam-dpdk/lib/librte_kni
Hemant Agrawal 3e12a98fe3 kni: optimize Rx burst
The current implementation of rte_kni_rx_burst polls the fifo for buffers.
Irrespective of success or failure, it allocates the mbuf and try to put them into the alloc_q
if the buffers are not added to alloc_q, it frees them.
This waste lots of cpu cycles in allocating and freeing the buffers if alloc_q is full.

The logic has been changed to:
1. Initially allocand add buffer(burstsize) to alloc_q
2. Add buffers to alloc_q only when you are pulling out the buffers.

Signed-off-by: Hemant Agrawal <hemant@freescale.com>
Reviewed-by: Jay Rolette <rolette@infiniteio.com>
2015-02-24 02:26:24 +01:00
..
Makefile mk: add library version extension 2015-02-03 16:56:58 +01:00
rte_kni_fifo.h remove trailing whitespaces 2014-06-11 00:29:34 +02:00
rte_kni_version.map lib: provide initial versioning 2015-02-03 16:56:58 +01:00
rte_kni.c kni: optimize Rx burst 2015-02-24 02:26:24 +01:00
rte_kni.h kni: memzone pool for alloc and release 2014-10-21 17:24:53 +02:00