From: Andrew Lunn <andrew@lunn.ch>
To: Hans H?bner <hans@huebner.org>
Cc: Will Lentz <will_lentz@trimble.com>, ecos-discuss@ecos.sourceware.org
Subject: Re: [ECOS] uipc_socket.c (and cyg_tcp_maxidle)
Date: Fri, 17 Jun 2005 20:12:00 -0000 [thread overview]
Message-ID: <20050617200916.GB17597@lunn.ch> (raw)
In-Reply-To: <20050616083626.U69813@web.m68k.de>
[-- Attachment #1: Type: text/plain, Size: 1297 bytes --]
On Thu, Jun 16, 2005 at 09:01:23AM +0200, Hans H?bner wrote:
> On Wed, 15 Jun 2005, Will Lentz wrote:
>
> >I may have found a potential bug in
> >packages/net/bsd_tcpip/current/src/sys/kern/uipc_socket.c (or I may be
> >completely wrong :-).
> >
> >At the end of sodealloc(), the following code exists:
> > zfreei(so->so_zone, so);
> > wakeup(so->so_zone);
> >The problem is that zfreei() changes so->so_zone. Shouldn't wakeup() be
> >done on the original so->so_zone? I only noticed this problem by:
> >1- while(1) {
> > sock = socket( AF_INET, SOCK_STREAM, IPPROTO_TCP );
> > connect( sock, ... );
> > close( sock );
> > }
> > Eventually this pauses in socket() (in cyg_tsleep()) when you run out
> >of eCos sockets.
> >
> >2- After 2*MSL or so, cyg_wakeup() gets called with chan == 0x0. Why?
> >The zfreei() call in sodealloc() changes so->so_zone to 0 before the
> >wakeup() call.
This is not quite correct. zfreei() does not change so->so_zone. What
it does is return the memory for the so structure to the pool. The
wakeup then uses the memory which has just been returned to the
pool. There is a race condition. Once back into the pool the memory
could be allocated to another thread before the call to wakeup is
made.
Attached is a patch to fix this.
Andrew
[-- Attachment #2: sodealloc.diff --]
[-- Type: text/plain, Size: 1495 bytes --]
Index: net/bsd_tcpip/current/ChangeLog
===================================================================
RCS file: /cvs/ecos/ecos-opt/net/net/bsd_tcpip/current/ChangeLog,v
retrieving revision 1.52
diff -u -r1.52 ChangeLog
--- net/bsd_tcpip/current/ChangeLog 27 Mar 2005 18:18:13 -0000 1.52
+++ net/bsd_tcpip/current/ChangeLog 17 Jun 2005 20:08:29 -0000
@@ -1,3 +1,8 @@
+2005-06-17 Andrew Lunn <andrew.lunn@ascom.ch>
+
+ * src/sys/kern/uipc_socket.c (sodealloc): Fixed a race condition
+ when freeing the socket memory. Problem reported by Will Lent.
+
2005-03-27 Andrew Lunn <andrew.lunn@ascom.ch>
* src/sys/net/if.c (ifioctl): Fixed a compiler warning about
Index: net/bsd_tcpip/current/src/sys/kern/uipc_socket.c
===================================================================
RCS file: /cvs/ecos/ecos-opt/net/net/bsd_tcpip/current/src/sys/kern/uipc_socket.c,v
retrieving revision 1.3
diff -u -r1.3 uipc_socket.c
--- net/bsd_tcpip/current/src/sys/kern/uipc_socket.c 24 Jul 2003 18:04:25 -0000 1.3
+++ net/bsd_tcpip/current/src/sys/kern/uipc_socket.c 17 Jun 2005 20:08:31 -0000
@@ -188,8 +188,10 @@
void
sodealloc(so)
struct socket *so;
+
{
-
+ vm_zone_t zone;
+
so->so_gencnt = ++so_gencnt;
#ifdef INET
if (so->so_accf != NULL) {
@@ -202,8 +204,9 @@
FREE(so->so_accf, M_ACCF);
}
#endif /* INET */
- zfreei(so->so_zone, so);
- wakeup(so->so_zone);
+ zone = so->so_zone;
+ zfreei(zone, so);
+ wakeup(zone);
}
int
[-- Attachment #3: Type: text/plain, Size: 148 bytes --]
--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss
next prev parent reply other threads:[~2005-06-17 20:12 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-06-15 22:39 [ECOS] uipc_socket.c Will Lentz
2005-06-16 7:06 ` [ECOS] uipc_socket.c (and cyg_tcp_maxidle) Hans Hübner
2005-06-16 15:36 ` Will Lentz
2005-06-16 15:51 ` Hans Hübner
2005-06-16 16:06 ` Will Lentz
[not found] ` <20050616183534.N69813@web.m68k.de>
2005-06-16 23:45 ` Will Lentz
2005-06-17 7:15 ` Andrew Lunn
2005-06-17 20:12 ` Andrew Lunn [this message]
2005-06-17 21:00 ` Will Lentz
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=20050617200916.GB17597@lunn.ch \
--to=andrew@lunn.ch \
--cc=ecos-discuss@ecos.sourceware.org \
--cc=hans@huebner.org \
--cc=will_lentz@trimble.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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).