diff options
| author | Paolo Abeni <pabeni@redhat.com> | 2025-11-25 18:10:34 +0300 |
|---|---|---|
| committer | Paolo Abeni <pabeni@redhat.com> | 2025-11-25 18:10:35 +0300 |
| commit | 61e628023d79386e93d2d64f8b7af439d27617a6 (patch) | |
| tree | 3d1994b0c5f9165963ea3270de1eb96282d76341 /tools/lib/python | |
| parent | e3daf0e7fe9758613bec324fd606ed9caa187f74 (diff) | |
| parent | a6efc273ab8245722eee2150fa12cf75781dc410 (diff) | |
| download | linux-61e628023d79386e93d2d64f8b7af439d27617a6.tar.xz | |
Merge branch 'net_sched-speedup-qdisc-dequeue'
Eric Dumazet says:
====================
net_sched: speedup qdisc dequeue
Avoid up to two cache line misses in qdisc dequeue() to fetch
skb_shinfo(skb)->gso_segs/gso_size while qdisc spinlock is held.
Idea is to cache gso_segs at enqueue time before spinlock is
acquired, in the first skb cache line, where we already
have qdisc_skb_cb(skb)->pkt_len.
This series gives a 8 % improvement in a TX intensive workload.
(120 Mpps -> 130 Mpps on a Turin host, IDPF with 32 TX queues)
v2: https://lore.kernel.org/netdev/20251111093204.1432437-1-edumazet@google.com/
v1: https://lore.kernel.org/netdev/20251110094505.3335073-1-edumazet@google.com/T/#m8f562ed148f807c02fd02c6cd243604d449615b9
====================
Link: https://patch.msgid.link/20251121083256.674562-1-edumazet@google.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Diffstat (limited to 'tools/lib/python')
0 files changed, 0 insertions, 0 deletions