From: Eric Wong <normalperson@yhbt.net>
To: kgio@librelist.org
Subject: [PATCH] accept: do not set blocking if non-blocking is set
Date: Sun, 30 Dec 2012 12:03:41 +0000 [thread overview]
Message-ID: <20121230120341.GA15638@dcvr.yhbt.net> (raw)
In-Reply-To: <20121230120341.GA15638@dcvr.yhbt.net>
This is prone to race conditions in multiprocess situations
where one process is relying on non-blocking operation while
another (likely newer process) relies on blocking operation.
Since the blocking process can always fall back to calling
rb_io_wait_readable(), use that instead and give up some
scalability for higher reliability.
Those interested in avoiding thundering herds will have to
stop/start their processes using blocking sockets (and tolerate
some downtime).
---
pushed to master of bogomips.org/kgio.git
commit c751f42f5f6a5e54a399df472015ab6d2ffc3f7a
ext/kgio/accept.c | 36 +++---------------------------------
1 file changed, 3 insertions(+), 33 deletions(-)
diff --git a/ext/kgio/accept.c b/ext/kgio/accept.c
index f8a073f..5f07532 100644
--- a/ext/kgio/accept.c
+++ b/ext/kgio/accept.c
@@ -79,15 +79,6 @@ static VALUE xaccept(void *ptr)
#ifdef HAVE_RB_THREAD_BLOCKING_REGION
# include <time.h>
# include "blocking_io_region.h"
-/*
- * Try to use a (real) blocking accept() since that can prevent
- * thundering herds under Linux:
- * http://www.citi.umich.edu/projects/linux-scalability/reports/accept.html
- *
- * So we periodically disable non-blocking, but not too frequently
- * because other processes may set non-blocking (especially during
- * a process upgrade) with Rainbows! concurrency model changes.
- */
static int thread_accept(struct accept_args *a, int force_nonblock)
{
if (force_nonblock)
@@ -95,28 +86,6 @@ static int thread_accept(struct accept_args *a, int force_nonblock)
return (int)rb_thread_io_blocking_region(xaccept, a, a->fd);
}
-static void set_blocking_or_block(int fd)
-{
- static time_t last_set_blocking;
- time_t now = time(NULL);
-
- if (last_set_blocking == 0) {
- last_set_blocking = now;
- (void)rb_io_wait_readable(fd);
- } else if ((now - last_set_blocking) <= 5) {
- (void)rb_io_wait_readable(fd);
- } else {
- int flags = fcntl(fd, F_GETFL);
- if (flags == -1)
- rb_sys_fail("fcntl(F_GETFL)");
- if (flags & O_NONBLOCK) {
- flags = fcntl(fd, F_SETFL, flags & ~O_NONBLOCK);
- if (flags == -1)
- rb_sys_fail("fcntl(F_SETFL)");
- }
- last_set_blocking = now;
- }
-}
#else /* ! HAVE_RB_THREAD_BLOCKING_REGION */
# include <rubysig.h>
static int thread_accept(struct accept_args *a, int force_nonblock)
@@ -134,7 +103,6 @@ static int thread_accept(struct accept_args *a, int force_nonblock)
TRAP_END;
return rv;
}
-#define set_blocking_or_block(fd) (void)rb_io_wait_readable(fd)
#endif /* ! HAVE_RB_THREAD_BLOCKING_REGION */
static void
@@ -209,7 +177,8 @@ retry:
return Qnil;
a->fd = my_fileno(a->accept_io);
errno = EAGAIN;
- set_blocking_or_block(a->fd);
+ (void)rb_io_wait_readable(a->fd);
+ /* fall-through to EINTR case */
#ifdef ECONNABORTED
case ECONNABORTED:
#endif /* ECONNABORTED */
@@ -217,6 +186,7 @@ retry:
case EPROTO:
#endif /* EPROTO */
case EINTR:
+ /* raise IOError if closed during sleep */
a->fd = my_fileno(a->accept_io);
goto retry;
case ENOMEM:
--
Eric Wong
parent reply other threads:[~2012-12-30 12:04 UTC|newest]
Thread overview: expand[flat|nested] mbox.gz Atom feed
[parent not found: <20121230120341.GA15638@dcvr.yhbt.net>]
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
List information: https://yhbt.net/kgio/
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20121230120341.GA15638@dcvr.yhbt.net \
--to=normalperson@yhbt.net \
--cc=kgio@librelist.org \
/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.
Code repositories for project(s) associated with this public inbox
https://yhbt.net/kgio.git/
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).