On Fri, Sep 7, 2018 at 12:03 AM Eric Dumazet [off-list ref] wrote:

Problem is : we have platforms with more than 100 cpus, and
sk_memory_allocated() cost will be too expensive,
especially if the host is under memory pressure, since all cpus will
touch their private counter.

per cpu variables do not really scale, they were ok 10 years ago when
no more than 16 cpus were the norm.

I would prefer change TCP to not aggressively call
__sk_mem_reduce_allocated() from tcp_write_timer()

Ideally only tcp_retransmit_timer() should attempt to reduce forward
allocations, after recurring timeout.

Note that after 20c64d5cd5a2bdcdc8982a06cb05e5e1bd851a3d ("net: avoid
sk_forward_alloc overflows")
we have better control over sockets having huge forward allocations.

Something like :

Or something less risky :

diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index 7fdf222a0bdfe9775970082f6b5dcdcc82b2ae1a..0aee80b6966cb2898e46350c761f9eb431ff1206

100644

--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c

@@ -604,7 +604,8 @@ void tcp_write_timer_handler(struct sock *sk)
        }

 out:
-       sk_mem_reclaim(sk);
+       if (tcp_under_memory_pressure(sk))
+               sk_mem_reclaim(sk);
 }

 static void tcp_write_timer(struct timer_list *t)