s3:events: Call all ready fd event handlers on each iteration of the main loop

Previously, only one fd handler was being called per main message loop
in all smbd child processes.

In the case where multiple fds are available for reading the fd
corresponding to the event closest to the beginning of the event list
would be run.  Obviously this is arbitrary and could cause unfairness.

Usually, the first event fd is the network socket, meaning heavy load
of client requests can starve out other fd events such as oplock
or notify upcalls from the kernel.

In this patch, I have changed the behavior of run_events() to unset
any fd that it has already called a handler function, as well
as decrement the number of fds that were returned from select().
This allows the caller of run_events() to iterate it, until all
available fds have been handled.

I then changed the main loop in smbd child processes to iterate
run_events().  This way, all available fds are handled on each wake
of select, while still checking for timed or signalled events between
each handler function call.  I also added an explicit check for
EINTR from select(), which previously was masked by the fact that
run_events() would handle any signal event before the return code
was checked.

This required a signature change to run_events() but all other callers
should have no change in their behavior.  I also fixed a bug in
run_events() where it could be called with a selrtn value of -1,
doing unecessary looping through the fd_event list when no fds were
available.

Also, remove the temporary echo handler hack, as all fds should be
treated fairly now.

1 files changed, 15 insertions, 7 deletions

diff --git a/source3/lib/events.c b/source3/lib/events.c
index 70a0d7c302..0e127f02dc 100644
--- a/source3/lib/events.c
+++ b/source3/lib/events.c
@@ -67,7 +67,7 @@ bool event_add_to_select_args(struct tevent_context *ev,
 }
 
 bool run_events(struct tevent_context *ev,
-		int selrtn, fd_set *read_fds, fd_set *write_fds)
+		int *selrtn, fd_set *read_fds, fd_set *write_fds)
 {
 	struct tevent_fd *fde;
 	struct timeval now;
@@ -112,7 +112,7 @@ bool run_events(struct tevent_context *ev,
 		return true;
 	}
 
-	if (selrtn == 0) {
+	if (*selrtn <= 0) {
 		/*
 		 * No fd ready
 		 */
@@ -122,8 +122,16 @@ bool run_events(struct tevent_context *ev,
 	for (fde = ev->fd_events; fde; fde = fde->next) {
 		uint16 flags = 0;
 
-		if (FD_ISSET(fde->fd, read_fds)) flags |= EVENT_FD_READ;
-		if (FD_ISSET(fde->fd, write_fds)) flags |= EVENT_FD_WRITE;
+		if (FD_ISSET(fde->fd, read_fds)) {
+			flags |= EVENT_FD_READ;
+			FD_CLR(fde->fd, read_fds);
+			(*selrtn)--;
+		}
+		if (FD_ISSET(fde->fd, write_fds)) {
+			flags |= EVENT_FD_WRITE;
+			FD_CLR(fde->fd, write_fds);
+			(*selrtn)--;
+		}
 
 		if (flags & fde->flags) {
 			fde->handler(ev, fde, flags, fde->private_data);
@@ -162,7 +170,7 @@ static int s3_event_loop_once(struct tevent_context *ev, const char *location)
 	struct timeval now, to;
 	fd_set r_fds, w_fds;
 	int maxfd = 0;
-	int ret;
+	int ret = 0;
 
 	FD_ZERO(&r_fds);
 	FD_ZERO(&w_fds);
@@ -170,7 +178,7 @@ static int s3_event_loop_once(struct tevent_context *ev, const char *location)
 	to.tv_sec = 9999;	/* Max timeout */
 	to.tv_usec = 0;
 
-	if (run_events(ev, 0, NULL, NULL)) {
+	if (run_events(ev, &ret, NULL, NULL)) {
 		return 0;
 	}
 
@@ -189,7 +197,7 @@ static int s3_event_loop_once(struct tevent_context *ev, const char *location)
 		return -1;
 	}
 
-	run_events(ev, ret, &r_fds, &w_fds);
+	run_events(ev, &ret, &r_fds, &w_fds);
 	return 0;
 }