diff mbox series

[v2] ipc: fix use-after-free on failed send

Message ID 69497bc41646c6dd5eebb097b3bdc8a2136aeb53.1542730524.git.anatoly.burakov@intel.com (mailing list archive)
State Accepted, archived
Delegated to: Thomas Monjalon
Headers show
Series [v2] ipc: fix use-after-free on failed send | expand

Checks

Context Check Description
ci/checkpatch success coding style OK
ci/mellanox-Performance-Testing success Performance Testing PASS
ci/intel-Performance-Testing success Performance Testing PASS
ci/Intel-compilation success Compilation OK

Commit Message

Burakov, Anatoly Nov. 20, 2018, 4:18 p.m. UTC
Previous fix for rte_panic has moved setting of alarm before
sending the message. This means that whether we send a message,
the alarm would still trigger. The comment noted that cleanup
would happen in the alarm handler, but that's not what actually
happened - instead, in the event of failed send we freed the
memory in-place, before putting the request on the queue.

This works OK when the message is sent, but when sending the
message fails, the alarm would still trigger with a pointer
argument that points to non-existent memory, and cause
memory corruption.

There probably is a "proper" fix for this issue, with correct
handling of sent vs. unsent requests, however it would be
simpler just to sacrifice the sent request in the (extremely
unlikely) event of alarm set failing. The other process would
still send a response, but it will be ignored by the sender.

Fixes: 45e5f49e87fb ("ipc: remove panic in async request")

Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
---

Notes:
    v2:
    - Move incrementing of nb_sent to before setting the alarm. This will
      let the API consumers know that some messages were sent, but failed
      to receive a response (due to failure of IPC to set the alarm).

 lib/librte_eal/common/eal_common_proc.c | 40 ++++++++-----------------
 1 file changed, 13 insertions(+), 27 deletions(-)

Comments

Thomas Monjalon Nov. 22, 2018, 10:09 p.m. UTC | #1
20/11/2018 17:18, Anatoly Burakov:
> Previous fix for rte_panic has moved setting of alarm before
> sending the message. This means that whether we send a message,
> the alarm would still trigger. The comment noted that cleanup
> would happen in the alarm handler, but that's not what actually
> happened - instead, in the event of failed send we freed the
> memory in-place, before putting the request on the queue.
> 
> This works OK when the message is sent, but when sending the
> message fails, the alarm would still trigger with a pointer
> argument that points to non-existent memory, and cause
> memory corruption.
> 
> There probably is a "proper" fix for this issue, with correct
> handling of sent vs. unsent requests, however it would be
> simpler just to sacrifice the sent request in the (extremely
> unlikely) event of alarm set failing. The other process would
> still send a response, but it will be ignored by the sender.
> 
> Fixes: 45e5f49e87fb ("ipc: remove panic in async request")
> 
> Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>

Applied, thanks
diff mbox series

Patch

diff --git a/lib/librte_eal/common/eal_common_proc.c b/lib/librte_eal/common/eal_common_proc.c
index f65ef56c0..1c3f09aad 100644
--- a/lib/librte_eal/common/eal_common_proc.c
+++ b/lib/librte_eal/common/eal_common_proc.c
@@ -827,20 +827,19 @@  mp_request_async(const char *dst, struct rte_mp_msg *req,
 		goto fail;
 	}
 
-	/*
-	 * set the alarm before sending message. there are two possible error
-	 * scenarios to consider here:
-	 *
-	 * - if the alarm set fails, we free the memory right there
-	 * - if the alarm set succeeds but sending message fails, then the alarm
-	 *   will trigger and clean up the memory
-	 *
-	 * Even if the alarm triggers too early (i.e. immediately), we're still
-	 * holding the lock to pending requests queue, so the interrupt thread
-	 * will just spin until we release the lock, and either release the
-	 * memory, or doesn't find any pending requests in the queue because we
-	 * never added any due to send message failure.
-	 */
+	ret = send_msg(dst, req, MP_REQ);
+	if (ret < 0) {
+		RTE_LOG(ERR, EAL, "Fail to send request %s:%s\n",
+			dst, req->name);
+		ret = -1;
+		goto fail;
+	} else if (ret == 0) {
+		ret = 0;
+		goto fail;
+	}
+	param->user_reply.nb_sent++;
+
+	/* if alarm set fails, we simply ignore the reply */
 	if (rte_eal_alarm_set(ts->tv_sec * 1000000 + ts->tv_nsec / 1000,
 			      async_reply_handle, pending_req) < 0) {
 		RTE_LOG(ERR, EAL, "Fail to set alarm for request %s:%s\n",
@@ -848,21 +847,8 @@  mp_request_async(const char *dst, struct rte_mp_msg *req,
 		ret = -1;
 		goto fail;
 	}
-
-	ret = send_msg(dst, req, MP_REQ);
-	if (ret < 0) {
-		RTE_LOG(ERR, EAL, "Fail to send request %s:%s\n",
-			dst, req->name);
-		ret = -1;
-		goto fail;
-	} else if (ret == 0) {
-		ret = 0;
-		goto fail;
-	}
 	TAILQ_INSERT_TAIL(&pending_requests.requests, pending_req, next);
 
-	param->user_reply.nb_sent++;
-
 	return 0;
 fail:
 	free(pending_req);