Re: [PATCH] [RFC] Optimize select/poll by putting small data setson the stack

[Eric Dumazet]

Eric Dumazet a écrit :
> Andi Kleen a écrit :
> > This is a RFC for now. I would be interested in testing
> > feedback. Patch is for 2.6.15.
> >
> > Optimize select and poll by a using stack space for small fd sets
> >
> > This brings back an old optimization from Linux 2.0. Using
> > the stack is faster than kmalloc. On a Intel P4 system
> > it speeds up a select of a single pty fd by about 13%
> > (~4000 cycles -> ~3500)
>
> Was this result on UP or SMP kernel ? Preempt or not ?
>
> I think we might play in do_pollfd() and use fget_light()/fput_light() 
> instead of fget()/fput() that are somewhat expensive because of atomic 
> inc/dec on SMP.

Just for completeness I include this patch against 2.6.15


--- linux-2.6.15/fs/select.c	2006-01-03 04:21:10.000000000 +0100
+++ linux-2.6.15-ed/fs/select.c	2006-01-04 00:04:10.000000000 +0100
@@ -221,17 +221,18 @@
 			}
 
 			for (j = 0; j < __NFDBITS; ++j, ++i, bit <<= 1) {
+				int fput_needed;
 				if (i >= n)
 					break;
 				if (!(bit & all_bits))
 					continue;
-				file = fget(i);
+				file = fget_light(i, &fput_needed);
 				if (file) {
 					f_op = file->f_op;
 					mask = DEFAULT_POLLMASK;
 					if (f_op && f_op->poll)
 						mask = (*f_op->poll)(file, retval ? NULL : wait);
-					fput(file);
+					fput_light(file, fput_needed);
 					if ((mask & POLLIN_SET) && (in & bit)) {
 						res_in |= bit;
 						retval++;
@@ -417,14 +418,15 @@
 		fdp = fdpage+i;
 		fd = fdp->fd;
 		if (fd >= 0) {
-			struct file * file = fget(fd);
+			int fput_needed;
+			struct file * file = fget_light(fd, &fput_needed);
 			mask = POLLNVAL;
 			if (file != NULL) {
 				mask = DEFAULT_POLLMASK;
 				if (file->f_op && file->f_op->poll)
 					mask = file->f_op->poll(file, *pwait);
 				mask &= fdp->events | POLLERR | POLLHUP;
-				fput(file);
+				fput_light(file, fput_needed);
 			}
 			if (mask) {
 				*pwait = NULL;