[PATCH nf-next 0/2] netfilter: nft_set_pipapo: speed up insertions

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Always prefer the avx2 implementation if its available.
This greatly improves insertion performance (each insertion
checks if the new element would overlap with an existing one):

time nft -f - <<EOF
table ip pipapo {
	set s {
		typeof ip saddr . tcp dport
		flags interval
		size 800000
		elements = { 10.1.1.1 - 10.1.1.4 . 3996,
[.. 800k entries elided .. ]

before:
real    1m55.993s
user    0m2.505s
sys     1m53.296s

after:
real    0m42.586s
user    0m2.554s
sys     0m39.811s

First patch does some refactoring so the common part can be reused
for both packetpath and control plane.
Second patch alters control plane to use avx2.

Florian Westphal (2):
  netfilter: nft_set_pipapo_avx2: split lookup function in two parts
  netfilter: nft_set_pipapo: use avx2 algorithm for insertions too

 net/netfilter/nft_set_pipapo.c      |  47 ++++++++--
 net/netfilter/nft_set_pipapo_avx2.c | 127 +++++++++++++++++-----------
 net/netfilter/nft_set_pipapo_avx2.h |   4 +
 3 files changed, 122 insertions(+), 56 deletions(-)

-- 
2.49.1





[Index of Archives]     [Netfitler Users]     [Berkeley Packet Filter]     [LARTC]     [Bugtraq]     [Yosemite Forum]

  Powered by Linux