[patch 2/2] fix compiler warning and optimize the non-blocking path

[patch 2/2] fix compiler warning and optimize the non-blocking path

[Davide Libenzi]

This patch adds a comment to ep_poll(), renames labels a bit clearly, 
fixes a warning of unused variable from gcc, and optimizes the 
non-blocking path a little.



From: Shawn Bohrer <shawn.bohrer@gmail.com>
Hinted-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Davide Libenzi <davidel@xmailserver.org>


- Davide


---
 fs/eventpoll.c |   32 ++++++++++++++++++++++++++++----
 1 file changed, 28 insertions(+), 4 deletions(-)

Index: linux-2.6.mod/fs/eventpoll.c
===================================================================
--- linux-2.6.mod.orig/fs/eventpoll.c	2011-01-15 10:50:16.000000000 -0800
+++ linux-2.6.mod/fs/eventpoll.c	2011-01-15 11:01:33.000000000 -0800
@@ -1127,27 +1127,50 @@
 	return ep_scan_ready_list(ep, ep_send_events_proc, &esed);
 }
 
+/**
+ * ep_poll - Retrieves ready events, and delivers them to the caller supplied
+ *           event buffer.
+ *
+ * @ep: Pointer to the eventpoll context.
+ * @events: Pointer to the userspace buffer where the ready events should be
+ *          stored.
+ * @maxevents: Size (in terms of number of events) of the caller event buffer.
+ * @timeout: Maximum timeout for the ready events fetch operation, in
+ *           milliseconds. If the @timeout is zero, the function will not block,
+ *           while if the @timeout is less than zero, the function will block
+ *           until at least one event has been retrieved (or an error
+ *           occurred).
+ *
+ * Returns: Returns the number of ready events which have been fetched, or an
+ *          error code, in case of error.
+ */
 static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events,
 		   int maxevents, long timeout)
 {
 	int res, eavail, timed_out = 0;
 	unsigned long flags;
-	long slack;
+	long slack = 0;
 	wait_queue_t wait;
 	struct timespec end_time;
 	ktime_t expires, *to = NULL;
 
 	if (timeout > 0) {
 		ktime_get_ts(&end_time);
-		timespec_add_ns(&end_time, (u64)timeout * NSEC_PER_MSEC);
+		timespec_add_ns(&end_time, (u64) timeout * NSEC_PER_MSEC);
 		slack = select_estimate_accuracy(&end_time);
 		to = &expires;
 		*to = timespec_to_ktime(end_time);
 	} else if (timeout == 0) {
+		/*
+		 * Avoid the unnecessary trip to the wait queue loop, if the
+		 * caller specified a non blocking operation.
+		 */
 		timed_out = 1;
+		spin_lock_irqsave(&ep->lock, flags);
+		goto check_events;
 	}
 
-retry:
+fetch_events:
 	spin_lock_irqsave(&ep->lock, flags);
 
 	res = 0;
@@ -1184,6 +1207,7 @@
 
 		set_current_state(TASK_RUNNING);
 	}
+check_events:
 	/* Is it worth to try to dig for events ? */
 	eavail = ep_events_available(ep);
 
@@ -1196,7 +1220,7 @@
 	 */
 	if (!res && eavail &&
 	    !(res = ep_send_events(ep, events, maxevents)) && !timed_out)
-		goto retry;
+		goto fetch_events;
 
 	return res;
 }

Re: [patch 2/2] fix compiler warning and optimize the non-blocking path

[Andrew Morton]

On Sat, 15 Jan 2011 11:43:45 -0800 (PST)
Davide Libenzi <davidel@xmailserver.org> wrote:

> +/**
> + * ep_poll - Retrieves ready events, and delivers them to the caller supplied
> + *           event buffer.
> + *
> + * @ep: Pointer to the eventpoll context.
> + * @events: Pointer to the userspace buffer where the ready events should be
> + *          stored.
> + * @maxevents: Size (in terms of number of events) of the caller event buffer.
> + * @timeout: Maximum timeout for the ready events fetch operation, in
> + *           milliseconds. If the @timeout is zero, the function will not block,
> + *           while if the @timeout is less than zero, the function will block
> + *           until at least one event has been retrieved (or an error
> + *           occurred).

Ah, that's what timeout<0 means.

My `man epoll_wait' says "a timeout of -1 makes epoll_wait() wait
indefinitely", whereas the implementation accepts any -ve value.

So we lost the ability to later make -[2..MAX_INT] meaningful in some
new way.  Oh well.

> + * Returns: Returns the number of ready events which have been fetched, or an
> + *          error code, in case of error.
> + */

Re: [patch 2/2] fix compiler warning and optimize the non-blocking path

[Davide Libenzi]

On Wed, 19 Jan 2011, Andrew Morton wrote:

> On Sat, 15 Jan 2011 11:43:45 -0800 (PST)
> Davide Libenzi <davidel@xmailserver.org> wrote:
> 
> > +/**
> > + * ep_poll - Retrieves ready events, and delivers them to the caller supplied
> > + *           event buffer.
> > + *
> > + * @ep: Pointer to the eventpoll context.
> > + * @events: Pointer to the userspace buffer where the ready events should be
> > + *          stored.
> > + * @maxevents: Size (in terms of number of events) of the caller event buffer.
> > + * @timeout: Maximum timeout for the ready events fetch operation, in
> > + *           milliseconds. If the @timeout is zero, the function will not block,
> > + *           while if the @timeout is less than zero, the function will block
> > + *           until at least one event has been retrieved (or an error
> > + *           occurred).
> 
> Ah, that's what timeout<0 means.
> 
> My `man epoll_wait' says "a timeout of -1 makes epoll_wait() wait
> indefinitely", whereas the implementation accepts any -ve value.
> 
> So we lost the ability to later make -[2..MAX_INT] meaningful in some
> new way.  Oh well.

I guess we can throw a -EINVAL if timeout <-1, if you prefer ...



- Davide

Re: [patch 2/2] fix compiler warning and optimize the non-blocking path

[Andrew Morton]

On Wed, 19 Jan 2011 15:49:09 -0800 (PST)
Davide Libenzi <davidel@xmailserver.org> wrote:

> On Wed, 19 Jan 2011, Andrew Morton wrote:
> 
> > On Sat, 15 Jan 2011 11:43:45 -0800 (PST)
> > Davide Libenzi <davidel@xmailserver.org> wrote:
> > 
> > > +/**
> > > + * ep_poll - Retrieves ready events, and delivers them to the caller supplied
> > > + *           event buffer.
> > > + *
> > > + * @ep: Pointer to the eventpoll context.
> > > + * @events: Pointer to the userspace buffer where the ready events should be
> > > + *          stored.
> > > + * @maxevents: Size (in terms of number of events) of the caller event buffer.
> > > + * @timeout: Maximum timeout for the ready events fetch operation, in
> > > + *           milliseconds. If the @timeout is zero, the function will not block,
> > > + *           while if the @timeout is less than zero, the function will block
> > > + *           until at least one event has been retrieved (or an error
> > > + *           occurred).
> > 
> > Ah, that's what timeout<0 means.
> > 
> > My `man epoll_wait' says "a timeout of -1 makes epoll_wait() wait
> > indefinitely", whereas the implementation accepts any -ve value.
> > 
> > So we lost the ability to later make -[2..MAX_INT] meaningful in some
> > new way.  Oh well.
> 
> I guess we can throw a -EINVAL if timeout <-1, if you prefer ...
> 

Too late!