mirrors/haproxy
1
0
Fork 0
mirror of https://git.haproxy.org/git/haproxy.git/ synced 2025-09-24 23:31:40 +02:00
Code Issues Projects Releases Wiki Activity

BUG/MEDIUM: tcp-check: properly indicate polling state before performing I/O

Browse Source 
While porting the connection to use the mux layer, it appeared that
tcp-checks wouldn't receive anymore because the polling is not enabled
before attempting to call xprt->rcv_buf() nor xprt->snd_buf(), and it
is illegal to call these functions with polling disabled as they
directly manipulate the FD state, resulting in an inconsistency where
the FD is enabled and the connection's polling flags disabled.

Till now it happened to work only because when recv() fails on EAGAIN
it calls fd_cant_recv() which enables polling while signaling the
failure, so that next time the message is received. But the connection's
polling is never enabled, and any tiny change resulting in a call to
conn_data_update_polling() immediately disables reading again.

It's likely that this problem already happens on some corner cases
such as multi-packet responses. It definitely breaks as soon as the
response buffer is full but we don't support consuming more than one
response buffer.

This fix should be backported to 1.7 and 1.6. In order to check for the
proper behaviour, this tcp-check must work and clearly show an SSH
banner in recvfrom() as observed under strace, otherwise it's broken :

   tcp-check connect port 22
   tcp-check expect rstring SSH
   tcp-check send blah
This commit is contained in:
Willy Tarreau 2017-10-04 08:45:19 +02:00
parent 3cad394520
commit a3782e7594
1 changed files with 2 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
src/checks.c
View File

@ -2591,6 +2591,7 @@ static void tcpcheck_main(struct connection *conn)
check->current_step->action != TCPCHK_ACT_SEND || check->current_step->action != TCPCHK_ACT_SEND ||
check->current_step->string_len >= buffer_total_space(check->bo))) { check->current_step->string_len >= buffer_total_space(check->bo))) {
__conn_data_want_send(conn);
if (conn->xprt->snd_buf(conn, check->bo, 0) <= 0) { if (conn->xprt->snd_buf(conn, check->bo, 0) <= 0) {
if (conn->flags & CO_FL_ERROR) { if (conn->flags & CO_FL_ERROR) {
chk_report_conn_err(conn, errno, 0); chk_report_conn_err(conn, errno, 0);
@ -2792,6 +2793,7 @@ static void tcpcheck_main(struct connection *conn)
if (unlikely(check->result == CHK_RES_FAILED)) if (unlikely(check->result == CHK_RES_FAILED))
goto out_end_tcpcheck; goto out_end_tcpcheck;
__conn_data_want_recv(conn);
if (conn->xprt->rcv_buf(conn, check->bi, check->bi->size) <= 0) { if (conn->xprt->rcv_buf(conn, check->bi, check->bi->size) <= 0) {
if (conn->flags & (CO_FL_ERROR | CO_FL_SOCK_RD_SH)) { if (conn->flags & (CO_FL_ERROR | CO_FL_SOCK_RD_SH)) {
done = 1; done = 1;