BUG/MEDIUM: mux-h1: Handle h1_process() failures on a pipelined request

On frontend side, when a conn-stream is detached from a H1 connection, the
H1 stream is destroyed and if we already have some data to parse (a
pipelined request), we process these data immedialtely calling
h1_process(). Then we adjust the H1 connection timeout. But h1_process() may
fail and release the H1 connection. For instance, a parsing error may be
reported. Thus, when that happens, we must not use anymore the H1 connection
and exit.

This patch must be backported as far as the 2.2. This bug can impact the 2.3
and the 2.2, in theory, if h1 stream creation fails. But, concretly, it only
fails on the 2.4 because the requests are now parsed at this step.
This commit is contained in:
Christopher Faulet 2020-12-18 15:13:47 +01:00
parent fac0f8f029
commit 0c366a8761

View File

@ -2782,8 +2782,10 @@ static void h1_detach(struct conn_stream *cs)
/* If we have a new request, process it immediately or /* If we have a new request, process it immediately or
* subscribe for reads waiting for new data * subscribe for reads waiting for new data
*/ */
if (unlikely(b_data(&h1c->ibuf))) if (unlikely(b_data(&h1c->ibuf))) {
h1_process(h1c); if (h1_process(h1c) == -1)
goto end;
}
else else
h1c->conn->xprt->subscribe(h1c->conn, h1c->conn->xprt_ctx, SUB_RETRY_RECV, &h1c->wait_event); h1c->conn->xprt->subscribe(h1c->conn, h1c->conn->xprt_ctx, SUB_RETRY_RECV, &h1c->wait_event);
} }