Re: [PATCH net] rxrpc: Call state should be read with READ_ONCE() under some circumstances
From: Jakub Kicinski <kuba@kernel.org>
Date: 2021-01-13 02:26:32
Also in:
lkml
On Tue, 12 Jan 2021 15:59:15 +0000 David Howells wrote:
From: Baptiste Lepers <redacted> The call state may be changed at any time by the data-ready routine in response to received packets, so if the call state is to be read and acted upon several times in a function, READ_ONCE() must be used unless the call state lock is held. As it happens, we used READ_ONCE() to read the state a few lines above the unmarked read in rxrpc_input_data(), so use that value rather than re-reading it. Signed-off-by: Baptiste Lepers <redacted> Signed-off-by: David Howells <dhowells@redhat.com>
Fixes: a158bdd3247b ("rxrpc: Fix call timeouts")
maybe?
quoted hunk ↗ jump to hunk
diff --git a/net/rxrpc/input.c b/net/rxrpc/input.c index 667c44aa5a63..dc201363f2c4 100644 --- a/net/rxrpc/input.c +++ b/net/rxrpc/input.c@@ -430,7 +430,7 @@ static void rxrpc_input_data(struct rxrpc_call *call, struct sk_buff *skb) return; } - if (call->state == RXRPC_CALL_SERVER_RECV_REQUEST) { + if (state == RXRPC_CALL_SERVER_RECV_REQUEST) { unsigned long timo = READ_ONCE(call->next_req_timo); unsigned long now, expect_req_by;