projects / linux-block.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 768818d) | patch
net: lantiq_xrx200: convert to build_skb
authorAleksander Jan Bajkowski <olek2@wp.pl>
Tue, 4 Jan 2022 15:11:44 +0000 (16:11 +0100)
committerJakub Kicinski <kuba@kernel.org>
Thu, 6 Jan 2022 01:18:04 +0000 (17:18 -0800)
commite015593573b3e3f74bd8a63c05fa92902194a354
tree0610d06bbc028ddd06da2a12e08bd2c54f095c23tree | snapshot
parent768818d772d5d4ddc0c7eb2e62848929270ab7a3commit | diff
net: lantiq_xrx200: convert to build_skb

We can increase the efficiency of rx path by using buffers to receive
packets then build SKBs around them just before passing into the network
stack. In contrast, preallocating SKBs too early reduces CPU cache
efficiency.

NAT Performance results on BT Home Hub 5A (kernel 5.10.89, mtu 1500):

Down Up
Before 577 Mbps 648 Mbps
After 624 Mbps 695 Mbps

Signed-off-by: Aleksander Jan Bajkowski <olek2@wp.pl>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
drivers/net/ethernet/lantiq_xrx200.c diff | blob | blame | history
Linux 6.x block layer and io_uring tree(s)