OSDN Git Service

net: thunderx: Adjust CQE_RX drop levels for better performance
authorSunil Goutham <sgoutham@marvell.com>
Mon, 2 Mar 2020 09:59:00 +0000 (15:29 +0530)
committerDavid S. Miller <davem@davemloft.net>
Mon, 2 Mar 2020 19:13:58 +0000 (11:13 -0800)
commitc0d2507abc2634ff6c6081d450924ffc68e740e0
tree0a9d916cde8174055512eba1acadc520279fcb29
parentaeaf0cc5b7b4ac37d782aa999d41b2cb68377e1b
net: thunderx: Adjust CQE_RX drop levels for better performance

With the current RX RED/DROP levels of 192/184 for CQE_RX, when
packet incoming rate is high, LLC is getting polluted resulting
in more cache misses and higher latency in packet processing. This
slows down the whole process and performance loss. Hence reduced
the levels to 224/216 (ie for a CQ size of 1024, Rx pkts will be
red dropped or dropped when unused CQE are less than 128/160 respectively)

Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/ethernet/cavium/thunder/nicvf_queues.h