If a call gets aborted (e.g. because kafs saw a signal) between it being
queued for connection and the I/O thread picking up the call, the abort
will be prioritised over the connection and it will be removed from
local->new_client_calls by rxrpc_disconnect_client_call() without a lock
being held. This may cause other calls on the list to disappear if a race
occurs.
Fix this by taking the client_call_lock when removing a call from whatever
list its ->wait_link happens to be on.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: linux-afs@lists.infradead.org
Reported-by: Marc Dionne <marc.dionne@auristor.com>
Fixes:
9d35d880e0e4 ("rxrpc: Move client call connection to the I/O thread")
Link: https://patch.msgid.link/726660.1730898202@warthog.procyon.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
EM(rxrpc_call_see_input, "SEE input ") \
EM(rxrpc_call_see_release, "SEE release ") \
EM(rxrpc_call_see_userid_exists, "SEE u-exists") \
+ EM(rxrpc_call_see_waiting_call, "SEE q-conn ") \
E_(rxrpc_call_see_zap, "SEE zap ")
#define rxrpc_txqueue_traces \
spin_lock(&local->client_call_lock);
list_move_tail(&call->wait_link, &bundle->waiting_calls);
+ rxrpc_see_call(call, rxrpc_call_see_waiting_call);
spin_unlock(&local->client_call_lock);
if (rxrpc_bundle_has_space(bundle))
_debug("call is waiting");
ASSERTCMP(call->call_id, ==, 0);
ASSERT(!test_bit(RXRPC_CALL_EXPOSED, &call->flags));
+ /* May still be on ->new_client_calls. */
+ spin_lock(&local->client_call_lock);
list_del_init(&call->wait_link);
+ spin_unlock(&local->client_call_lock);
return;
}