MINOR: debug: add a new DEBUG_FD build option

When DEBUG_FD is set at build time, we'll keep a counter of per-FD events
in the fdtab. This counter is reported in "show fd" even for closed FDs if
not zero. The purpose is to help spot situations where an apparently closed
FD continues to be reported in loops, or where some events are dismissed.

Makefile

@@ -213,7 +213,8 @@ SMALL_OPTS =
 # You can enable debugging on specific code parts by setting DEBUG=-DDEBUG_xxx.
 # Currently defined DEBUG macros include DEBUG_FULL, DEBUG_MEMORY, DEBUG_FSM,
 # DEBUG_HASH, DEBUG_AUTH, DEBUG_SPOE, DEBUG_UAF and DEBUG_THREAD, DEBUG_STRICT,
-# DEBUG_DEV. Please check sources for exact meaning or do not use at all.
+# DEBUG_DEV, DEBUG_FD. Please check sources for exact meaning or do not use at
+# all.
 DEBUG =
 
 #### Trace options

include/haproxy/fd-t.h

@@ -134,6 +134,9 @@ struct fdtab {
 	unsigned char cloned:1;              /* 1 if a cloned socket, requires EPOLL_CTL_DEL on close */
 	unsigned char initialized:1;         /* 1 if init phase was done on this fd (e.g. set non-blocking) */
 	unsigned char et_possible:1;         /* 1 if edge-triggered is possible on this FD */
+#ifdef DEBUG_FD
+	unsigned int event_count;            /* number of events reported */
+#endif
 } THREAD_ALIGNED(64);
 
 /* polled mask, one bit per thread and per direction for each FD */

include/haproxy/fd.h

@@ -446,6 +446,9 @@ static inline void fd_insert(int fd, void *owner, void (*iocb)(int fd), unsigned
 	fdtab[fd].linger_risk = 0;
 	fdtab[fd].cloned = 0;
 	fdtab[fd].et_possible = 0;
+#ifdef DEBUG_FD
+	fdtab[fd].event_count = 0;
+#endif
 
 	/* conn_fd_handler should support edge-triggered FDs */
 	if ((global.tune.options & GTUNE_FD_ET) && fdtab[fd].iocb == conn_fd_handler)

src/cli.c

@@ -1002,7 +1002,14 @@ static int cli_io_handler_show_fd(struct appctx *appctx)
 
 		fdt = fdtab[fd];
 
-		if (!fdt.owner)
+		/* When DEBUG_FD is set, we also report closed FDs that have a
+		 * non-null event count to detect stuck ones.
+		 */
+		if (!fdt.owner
+#ifdef DEBUG_FD
+		    && !fdt.event_count
+#endif
+		    )
 			goto skip; // closed
 
 		if (fdt.iocb == conn_fd_handler) {
@@ -1038,7 +1045,10 @@ static int cli_io_handler_show_fd(struct appctx *appctx)
 			     fdt.iocb);
 		resolve_sym_name(&trash, NULL, fdt.iocb);
 
-		if (fdt.iocb == conn_fd_handler) {
+		if (!fdt.owner) {
+			chunk_appendf(&trash, ")");
+		}
+		else if (fdt.iocb == conn_fd_handler) {
 			chunk_appendf(&trash, ") back=%d cflg=0x%08x", is_back, conn_flags);
 			if (px)
 				chunk_appendf(&trash, " px=%s", px->id);
@@ -1061,6 +1071,9 @@ static int cli_io_handler_show_fd(struct appctx *appctx)
 			              li->bind_conf->frontend->id);
 		}
 
+#ifdef DEBUG_FD
+		chunk_appendf(&trash, " evcnt=%u", fdtab[fd].event_count);
+#endif
 		chunk_appendf(&trash, ")\n");
 
 		if (ci_putchk(si_ic(si), &trash) == -1) {

src/ev_epoll.c

@@ -222,6 +222,9 @@ static void _do_poll(struct poller *p, int exp, int wake)
 		e = epoll_events[count].events;
 		fd = epoll_events[count].data.fd;
 
+#ifdef DEBUG_FD
+		_HA_ATOMIC_ADD(&fdtab[fd].event_count, 1);
+#endif
 		if (!fdtab[fd].owner) {
 			activity[tid].poll_dead_fd++;
 			continue;

src/ev_evports.c

@@ -216,6 +216,9 @@ static void _do_poll(struct poller *p, int exp, int wake)
 		fd = evports_evlist[i].portev_object;
 		events = evports_evlist[i].portev_events;
 
+#ifdef DEBUG_FD
+		_HA_ATOMIC_ADD(&fdtab[fd].event_count, 1);
+#endif
 		if (fdtab[fd].owner == NULL) {
 			activity[tid].poll_dead_fd++;
 			continue;

src/ev_kqueue.c

@@ -183,6 +183,9 @@ static void _do_poll(struct poller *p, int exp, int wake)
 		unsigned int n = 0;
 		fd = kev[count].ident;
 
+#ifdef DEBUG_FD
+		_HA_ATOMIC_ADD(&fdtab[fd].event_count, 1);
+#endif
 		if (!fdtab[fd].owner) {
 			activity[tid].poll_dead_fd++;
 			continue;

src/ev_poll.c

@@ -219,6 +219,9 @@ static void _do_poll(struct poller *p, int exp, int wake)
 		int e = poll_events[count].revents;
 		fd = poll_events[count].fd;
 
+#ifdef DEBUG_FD
+		_HA_ATOMIC_ADD(&fdtab[fd].event_count, 1);
+#endif
 		if (!(e & ( POLLOUT | POLLIN | POLLERR | POLLHUP | POLLRDHUP )))
 			continue;
 

src/ev_select.c

@@ -197,6 +197,9 @@ static void _do_poll(struct poller *p, int exp, int wake)
 		for (count = BITS_PER_INT, fd = fds * BITS_PER_INT; count && fd < maxfd; count--, fd++) {
 			unsigned int n = 0;
 
+#ifdef DEBUG_FD
+			_HA_ATOMIC_ADD(&fdtab[fd].event_count, 1);
+#endif
 			if (!fdtab[fd].owner) {
 				activity[tid].poll_dead_fd++;
 				continue;

src/fd.c

@@ -322,6 +322,9 @@ static void fd_dodelete(int fd, int do_close)
 
 	fdtab[fd].state = 0;
 
+#ifdef DEBUG_FD
+	fdtab[fd].event_count = 0;
+#endif
 	port_range_release_port(fdinfo[fd].port_range, fdinfo[fd].local_port);
 	fdinfo[fd].port_range = NULL;
 	fdtab[fd].owner = NULL;