aboutsummaryrefslogtreecommitdiffstats
path: root/net/tipc/msg.c
diff options
context:
space:
mode:
authorJon Paul Maloy <jon.maloy@ericsson.com>2015-02-05 08:36:41 -0500
committerDavid S. Miller <davem@davemloft.net>2015-02-05 16:00:02 -0800
commitc637c1035534867b85b78b453c38c495b58e2c5a (patch)
tree77cd2a48a5b04e43b014da64168a6c1e209a1d40 /net/tipc/msg.c
parenttipc: use existing sk_write_queue for outgoing packet chain (diff)
downloadlinux-dev-c637c1035534867b85b78b453c38c495b58e2c5a.tar.xz
linux-dev-c637c1035534867b85b78b453c38c495b58e2c5a.zip
tipc: resolve race problem at unicast message reception
TIPC handles message cardinality and sequencing at the link layer, before passing messages upwards to the destination sockets. During the upcall from link to socket no locks are held. It is therefore possible, and we see it happen occasionally, that messages arriving in different threads and delivered in sequence still bypass each other before they reach the destination socket. This must not happen, since it violates the sequentiality guarantee. We solve this by adding a new input buffer queue to the link structure. Arriving messages are added safely to the tail of that queue by the link, while the head of the queue is consumed, also safely, by the receiving socket. Sequentiality is secured per socket by only allowing buffers to be dequeued inside the socket lock. Since there may be multiple simultaneous readers of the queue, we use a 'filter' parameter to reduce the risk that they peek the same buffer from the queue, hence also reducing the risk of contention on the receiving socket locks. This solves the sequentiality problem, and seems to cause no measurable performance degradation. A nice side effect of this change is that lock handling in the functions tipc_rcv() and tipc_bcast_rcv() now becomes uniform, something that will enable future simplifications of those functions. Reviewed-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net/tipc/msg.c')
-rw-r--r--net/tipc/msg.c34
1 files changed, 34 insertions, 0 deletions
diff --git a/net/tipc/msg.c b/net/tipc/msg.c
index 697223a21240..b6eb90cd3ef7 100644
--- a/net/tipc/msg.c
+++ b/net/tipc/msg.c
@@ -327,6 +327,40 @@ bool tipc_msg_bundle(struct sk_buff_head *list, struct sk_buff *skb, u32 mtu)
}
/**
+ * tipc_msg_extract(): extract bundled inner packet from buffer
+ * @skb: linear outer buffer, to be extracted from.
+ * @iskb: extracted inner buffer, to be returned
+ * @pos: position of msg to be extracted. Returns with pointer of next msg
+ * Consumes outer buffer when last packet extracted
+ * Returns true when when there is an extracted buffer, otherwise false
+ */
+bool tipc_msg_extract(struct sk_buff *skb, struct sk_buff **iskb, int *pos)
+{
+ struct tipc_msg *msg = buf_msg(skb);
+ int imsz;
+ struct tipc_msg *imsg = (struct tipc_msg *)(msg_data(msg) + *pos);
+
+ /* Is there space left for shortest possible message? */
+ if (*pos > (msg_data_sz(msg) - SHORT_H_SIZE))
+ goto none;
+ imsz = msg_size(imsg);
+
+ /* Is there space left for current message ? */
+ if ((*pos + imsz) > msg_data_sz(msg))
+ goto none;
+ *iskb = tipc_buf_acquire(imsz);
+ if (!*iskb)
+ goto none;
+ skb_copy_to_linear_data(*iskb, imsg, imsz);
+ *pos += align(imsz);
+ return true;
+none:
+ kfree_skb(skb);
+ *iskb = NULL;
+ return false;
+}
+
+/**
* tipc_msg_make_bundle(): Create bundle buf and append message to its tail
* @list: the buffer chain
* @skb: buffer to be appended and replaced