mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: "Ken Chen" <kenchen@google.com>
To: "Jeff Moyer" <jmoyer@redhat.com>
Cc: "Zach Brown" <zach.brown@oracle.com>,
	akpm@linux-foundation.org, linux-aio@kvack.org,
	linux-kernel@vger.kernel.org
Subject: Re: [patch] convert aio event reap to use atomic-op instead of spin_lock
Date: Thu, 12 Apr 2007 17:57:45 -0700	[thread overview]
Message-ID: <b040c32a0704121757p728654cdte652c2a900e4c2d4@mail.gmail.com> (raw)
In-Reply-To: <x49ejmp3igr.fsf@segfault.boston.devel.redhat.com>

On 4/12/07, Jeff Moyer <jmoyer@redhat.com> wrote:
> I didn't see any response to Zach's request for code that actually
> tests out the shared ring buffer.  Do you have such code?

Yes, I do.  I was stress testing the code since last night.  After 20+
hours of stress run with fio and aio-stress, now I'm posting it with
confidence.

I modified libaio's io_getevents to take advantage of new user level
reap function. The feature is exported out via ring->compat_features.
btw, is compat_feature suppose to be a version number or a bit mask?
I think bitmask make more sense and more flexible.

(warning: some lines are extremely long in the patch and my email
client will probably mangle it badly).


diff -Nurp libaio-0.3.104/src/io_getevents.c
libaio-0.3.104-new/src/io_getevents.c
--- libaio-0.3.104/src/io_getevents.c	2003-06-18 12:58:21.000000000 -0700
+++ libaio-0.3.104-new/src/io_getevents.c	2007-04-12 17:35:06.000000000 -0700
@@ -21,10 +21,13 @@
 #include <stdlib.h>
 #include <time.h>
 #include "syscall.h"
+#include <asm/system.h>

 io_syscall5(int, __io_getevents_0_4, io_getevents, io_context_t, ctx,
long, min_nr, long, nr, struct io_event *, events, struct timespec *,
timeout)

 #define AIO_RING_MAGIC                  0xa10a10a1
+#define AIO_RING_BASE			1
+#define AIO_RING_USER_REAP		2

 /* Ben will hate me for this */
 struct aio_ring {
@@ -41,7 +44,11 @@ struct aio_ring {

 int io_getevents_0_4(io_context_t ctx, long min_nr, long nr, struct
io_event * events, struct timespec * timeout)
 {
+	long i = 0, ret;
+	unsigned head;
+	struct io_event *evt_base;
 	struct aio_ring *ring;
+
 	ring = (struct aio_ring*)ctx;
 	if (ring==NULL || ring->magic != AIO_RING_MAGIC)
 		goto do_syscall;
@@ -49,9 +56,35 @@ int io_getevents_0_4(io_context_t ctx, l
 		if (ring->head == ring->tail)
 			return 0;
 	}
-	
+
+	if (!(ring->compat_features & AIO_RING_USER_REAP))
+		goto do_syscall;
+
+	if (min_nr > nr || min_nr < 0 || nr < 0)
+		return -EINVAL;
+
+	evt_base = (struct io_event *) (ring + 1);
+	while (i < nr) {
+		head = ring->head;
+		if (head == ring->tail)
+			break;
+
+		*events = evt_base[head & (ring->nr - 1)];
+		if (head == cmpxchg(&ring->head, head, head + 1)) {
+			events++;
+			i++;
+		}
+	}
+
+	if (i >= min_nr)
+		return i;
+
 do_syscall:	
-	return __io_getevents_0_4(ctx, min_nr, nr, events, timeout);
+	ret = __io_getevents_0_4(ctx, min_nr - i, nr - i, events, timeout);
+	if (ret >= 0)
+		return i + ret;
+	else
+		return i ? i : ret;
 }

 DEFSYMVER(io_getevents_0_4, io_getevents, 0.4)

  reply	other threads:[~2007-04-13  0:57 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-04-10 23:53 Ken Chen
2007-04-11  5:18 ` Andrew Morton
2007-04-11 16:54   ` Ken Chen
2007-04-11 18:15     ` Zach Brown
2007-04-11 18:29       ` Ken Chen
2007-04-11 18:00 ` Zach Brown
2007-04-11 19:11   ` Andrew Morton
2007-04-11 19:34     ` Zach Brown
2007-04-11 19:28   ` Ken Chen
2007-04-11 19:44     ` Zach Brown
2007-04-11 19:45     ` Benjamin LaHaise
2007-04-11 19:52       ` Zach Brown
2007-04-11 20:03         ` Benjamin LaHaise
2007-04-12  7:29     ` Ken Chen
2007-04-12 12:06       ` Jeff Moyer
2007-04-13  0:57         ` Ken Chen [this message]
2007-04-13  1:08           ` Ken Chen
2007-04-12  7:50   ` Ken Chen
2007-04-12 14:31     ` Benjamin LaHaise
2007-04-12 15:13       ` Benjamin LaHaise

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=b040c32a0704121757p728654cdte652c2a900e4c2d4@mail.gmail.com \
    --to=kenchen@google.com \
    --cc=akpm@linux-foundation.org \
    --cc=jmoyer@redhat.com \
    --cc=linux-aio@kvack.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=zach.brown@oracle.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®