aboutsummaryrefslogtreecommitdiffstats
path: root/target/linux/bcm4908/config-5.10
diff options
context:
space:
mode:
authorRafał Miłecki <rafal@milecki.pl>2022-07-28 17:46:33 +0200
committerRafał Miłecki <rafal@milecki.pl>2022-09-01 12:46:37 +0200
commitd3ccf0976ae9a7f93a72e0dfeee632620d9333ea (patch)
treedb45d4fa45cf1ce5978cb494e2b9ee83e7cbcb0b /target/linux/bcm4908/config-5.10
parent30bce90d84ad307e3c4f8426b580da9907e13e32 (diff)
downloadupstream-d3ccf0976ae9a7f93a72e0dfeee632620d9333ea.tar.gz
upstream-d3ccf0976ae9a7f93a72e0dfeee632620d9333ea.tar.bz2
upstream-d3ccf0976ae9a7f93a72e0dfeee632620d9333ea.zip
bcm4908: enable & setup packet steering
Without packet steering NAT masquarade speed on BCM4908 /jumps/ between two speeds: 1. 826 Mb/s (±3 Mb/s) 2. 909 Mb/s (±8 Mb/s) and it never reaches ~940 Mb/s. Proper packet steering can improve it. Below are testing results for running iperf TCP traffic from LAN to WAN. They were used to pick up golden values. ┌──────────┬──────────┬───────────┐ │ eth0 │ br-lan │ speed │ │ rps_cpus │ rps_cpus │ [Mbps] │ ├──────────┼──────────┼───────────┤ │ 0 │ 0 │ 743 / 804 │ │ 0 │ 1 │ 738 / 821 │ │ 0 │ 2 │ ✓ 940 │ │ 0 │ 4 │ ✓ 938 │ │ 0 │ 8 │ ✓ 941 │ ├──────────┼──────────┼───────────┤ │ 1 │ 0 │ 829 │ │ 1 │ 1 │ 829 │ │ 1 │ 2 │ ✓ 942 │ │ 1 │ 4 │ ✓ 941 │ │ 1 │ 8 │ ✓ 941 │ ├──────────┼──────────┼───────────┤ │ 2 │ 0 │ ✓ 942 │ │ 2 │ 1 │ 926 │ │ 2 │ 2 │ ✓ 942 │ │ 2 │ 4 │ ✓ 942 │ │ 2 │ 8 │ ✓ 941 │ ├──────────┼──────────┼───────────┤ │ 4 │ 0 │ ✓ 941 │ │ 4 │ 1 │ 925 │ │ 4 │ 2 │ ✓ 941 │ │ 4 │ 4 │ ✓ 941 │ │ 4 │ 8 │ ✓ 941 │ ├──────────┼──────────┼───────────┤ │ 8 │ 0 │ ✓ 942 │ │ 8 │ 1 │ 925 │ │ 8 │ 2 │ ✓ 941 │ │ 8 │ 4 │ ✓ 942 │ │ 8 │ 8 │ ✓ 942 │ └──────────┴──────────┴───────────┘ Ref: fcbd39689ebfe ("bcm53xx: enable & setup packet steering") Signed-off-by: Rafał Miłecki <rafal@milecki.pl> (cherry picked from commit 57cad53f4e52be987cdd61308ff7d2704baca539)
Diffstat (limited to 'target/linux/bcm4908/config-5.10')
0 files changed, 0 insertions, 0 deletions