Re: [PATCH] [REBASED for 4.14] once: add DO_ONCE_SLOW() for sleepable contexts

2022-12-14 Thread Greg KH
On Tue, Dec 13, 2022 at 01:22:40PM +0100, Christophe Leroy wrote: > From: Eric Dumazet > > [ Upstream commit 62c07983bef9d3e78e71189441e1a470f0d1e653 ] > > Christophe Leroy reported a ~80ms latency spike > happening at first TCP connect() time. > > This is because __inet_hash_connect() uses

[PATCH] [REBASED for 4.14] once: add DO_ONCE_SLOW() for sleepable contexts

2022-12-13 Thread Christophe Leroy
From: Eric Dumazet [ Upstream commit 62c07983bef9d3e78e71189441e1a470f0d1e653 ] Christophe Leroy reported a ~80ms latency spike happening at first TCP connect() time. This is because __inet_hash_connect() uses get_random_once() to populate a perturbation table which became quite big after