diff options
| author | Jakub Kicinski <kuba@kernel.org> | 2026-01-22 06:28:34 +0300 |
|---|---|---|
| committer | Jakub Kicinski <kuba@kernel.org> | 2026-01-22 06:28:35 +0300 |
| commit | 9de76f55b9f856b317773e106253051fb33a9e92 (patch) | |
| tree | c22b5d9ac23eab395709dfcce832ac9f679a7e6f /include/linux/raid | |
| parent | a4674aa58be53921f2aba62e143cc338d6ab142c (diff) | |
| parent | b8d9b7daf0af367f3fff017de0873ab825a6dbbe (diff) | |
| download | linux-9de76f55b9f856b317773e106253051fb33a9e92.tar.xz | |
Merge branch 'gro-inline-tcp6_gro_-receive-complete'
Eric Dumazet says:
====================
gro: inline tcp6_gro_{receive,complete}
On some platforms, GRO stack is too deep and causes cpu stalls.
Decreasing call depths by one shows a 1.5 % gain on Zen2 cpus.
(32 RX queues, 100Gbit NIC, RFS enabled, tcp_rr with 128 threads and 10,000 flows)
We can go further by inlining ipv6_gro_{receive,complete}
and take care of IPv4 if there is interest.
Note: two temporary __always_inline will be replaced with
inline_for_performance when/if available.
Cumulative size increase for this series (of 3):
$ scripts/bloat-o-meter -t vmlinux.0 vmlinux.3
add/remove: 2/2 grow/shrink: 5/1 up/down: 1572/-471 (1101)
Function old new delta
ipv6_gro_receive 1069 1846 +777
ipv6_gro_complete 433 733 +300
tcp6_check_fraglist_gro - 272 +272
tcp6_gro_complete 227 306 +79
tcp4_gro_complete 325 397 +72
ipv6_offload_init 218 274 +56
__pfx_tcp6_check_fraglist_gro - 16 +16
__pfx___skb_incr_checksum_unnecessary 32 - -32
__skb_incr_checksum_unnecessary 186 - -186
tcp6_gro_receive 959 706 -253
Total: Before=22592724, After=22593825, chg +0.00%
====================
Link: https://patch.msgid.link/20260120164903.1912995-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Diffstat (limited to 'include/linux/raid')
0 files changed, 0 insertions, 0 deletions