[PATCH net] rxrpc: Call state should be read with READ_ONCE() under some circumstances

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>
---

 net/rxrpc/input.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

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;
 

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?

Jakub Kicinski [off-list ref] wrote:

On Tue, 12 Jan 2021 15:59:15 +0000 David Howells wrote:

quoted

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?

Ah, yes.  I missed there wasn't a Fixes line.  Can you add that one in, or do
I need to resubmit the patch?

David

On Wed, 13 Jan 2021 08:23:54 +0000 David Howells wrote:

Jakub Kicinski [off-list ref] wrote:

quoted

On Tue, 12 Jan 2021 15:59:15 +0000 David Howells wrote:  

quoted

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?  

Ah, yes.  I missed there wasn't a Fixes line.  Can you add that one in, or do
I need to resubmit the patch?

Sure, added, just checking if you didn't leave it out on purpose.

Applied, thanks!