liquidio: optimize DMA in NUMA systems
authorVSR Burru <veerasenareddy.burru@cavium.com>
Fri, 10 Mar 2017 01:03:24 +0000 (17:03 -0800)
committerDavid S. Miller <davem@davemloft.net>
Mon, 13 Mar 2017 06:17:00 +0000 (23:17 -0700)
commitb3ca9af0fb65098dd2afecca6831c361a585f37f
treee737e5e34e64ef0d9f428aeb14c7acf615cdf36f
parentde12e4fb8f5f35bfeada567dabf597ef6004127d
liquidio: optimize DMA in NUMA systems

Optimize DMA in NUMA systems by allocating memory from NUMA node that NIC
is plugged in to; DMA will no longer cross NUMA nodes.  If NIC IRQs are
pinned to a local CPU, that CPU's access to the DMA'd data is also
optimized.

Signed-off-by: VSR Burru <veerasenareddy.burru@cavium.com>
Signed-off-by: Felix Manlunas <felix.manlunas@cavium.com>
Signed-off-by: Raghu Vatsavayi <raghu.vatsavayi@cavium.com>
Signed-off-by: Satanand Burla <satananda.burla@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/ethernet/cavium/liquidio/lio_main.c
drivers/net/ethernet/cavium/liquidio/octeon_device.c
drivers/net/ethernet/cavium/liquidio/octeon_droq.c
drivers/net/ethernet/cavium/liquidio/octeon_iq.h
drivers/net/ethernet/cavium/liquidio/request_manager.c