tcp: annotate sk->sk_wmem_queued lockless reads

For the sake of tcp_poll(), there are few places where we fetch
sk->sk_wmem_queued while this field can change from IRQ or other cpu.

We need to add READ_ONCE() annotations, and also make sure write
sides use corresponding WRITE_ONCE() to avoid store-tearing.

sk_wmem_queued_add() helper is added so that we can in
the future convert to ADD_ONCE() or equivalent if/when
available.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit is contained in:
Eric Dumazet
2019-10-10 20:17:46 -07:00
committed by David S. Miller
parent e292f05e0d
commit ab4e846a82
8 changed files with 24 additions and 19 deletions

View File

@@ -659,7 +659,7 @@ static void skb_entail(struct sock *sk, struct sk_buff *skb)
tcb->sacked = 0;
__skb_header_release(skb);
tcp_add_write_queue_tail(sk, skb);
sk->sk_wmem_queued += skb->truesize;
sk_wmem_queued_add(sk, skb->truesize);
sk_mem_charge(sk, skb->truesize);
if (tp->nonagle & TCP_NAGLE_PUSH)
tp->nonagle &= ~TCP_NAGLE_PUSH;
@@ -1034,7 +1034,7 @@ new_segment:
skb->len += copy;
skb->data_len += copy;
skb->truesize += copy;
sk->sk_wmem_queued += copy;
sk_wmem_queued_add(sk, copy);
sk_mem_charge(sk, copy);
skb->ip_summed = CHECKSUM_PARTIAL;
WRITE_ONCE(tp->write_seq, tp->write_seq + copy);