From: Brian Vazquez <brianvv@xxxxxxxxxx>
[ Upstream commit 5c1ebbfabcd61142a4551bfc0e51840f9bdae7af ]
Florian reported a regression and sent a patch with the following
changelog:
<quote>
There is a noticeable tcp performance regression (loopback or cross-netns),
seen with iperf3 -Z (sendfile mode) when generic retpolines are needed.
With SK_RECLAIM_THRESHOLD checks gone number of calls to enter/leave
memory pressure happen much more often. For TCP indirect calls are
used.
We can't remove the if-set-return short-circuit check in
tcp_enter_memory_pressure because there are callers other than
sk_enter_memory_pressure. Doing a check in the sk wrapper too
reduces the indirect calls enough to recover some performance.
Before,
0.00-60.00 sec 322 GBytes 46.1 Gbits/sec receiver
After:
0.00-60.04 sec 359 GBytes 51.4 Gbits/sec receiver
"iperf3 -c $peer -t 60 -Z -f g", connected via veth in another netns.
</quote>
It seems we forgot to upstream this indirect call mitigation we
had for years, lets do this instead.
[edumazet] - It seems we forgot to upstream this indirect call
mitigation we had for years, let's do this instead.
- Changed to INDIRECT_CALL_INET_1() to avoid bots reports.
Fixes: 4890b686f408 ("net: keep sk->sk_forward_alloc as small as possible")
Reported-by: Florian Westphal <fw@xxxxxxxxx>
Link: https://lore.kernel.org/netdev/20230227152741.4a53634b@xxxxxxxxxx/T/
Signed-off-by: Brian Vazquez <brianvv@xxxxxxxxxx>
Signed-off-by: Eric Dumazet <edumazet@xxxxxxxxxx>
Link: https://lore.kernel.org/r/20230301133247.2346111-1-edumazet@xxxxxxxxxx
Signed-off-by: Paolo Abeni <pabeni@xxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
net/core/sock.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/net/core/sock.c b/net/core/sock.c
index 4dfdcdfd00114..eb0b76acd9df1 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -2805,7 +2805,8 @@ static void sk_enter_memory_pressure(struct sock *sk)
static void sk_leave_memory_pressure(struct sock *sk)
{
if (sk->sk_prot->leave_memory_pressure) {
- sk->sk_prot->leave_memory_pressure(sk);
+ INDIRECT_CALL_INET_1(sk->sk_prot->leave_memory_pressure,
+ tcp_leave_memory_pressure, sk);
} else {
unsigned long *memory_pressure = sk->sk_prot->memory_pressure;
--
2.39.2
