Date: Fri, 12 Sep 1997 01:12:12 +0200 From: sthaug@nethelp.no To: fenner@FreeBSD.ORG Cc: freebsd-bugs@hub.freebsd.org Subject: Re: kern/3925 Message-ID: <402.874019532@verdi.nethelp.no> In-Reply-To: Your message of "Sat, 5 Jul 1997 11:55:59 -0700 (PDT)" References: <199707051855.LAA16321@hub.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
> Synopsis: SO_SNDLOWAT of 0 causes kernel to use 99% of CPU time on TCP send
>
> State-Changed-From-To: open-feedback
> State-Changed-By: fenner
> State-Changed-When: Sat Jul 5 11:55:17 PDT 1997
> State-Changed-Why:
> Could you verify that rev 1.27 of kern/uipc_socket.c fixes this problem?
Hi, sorry this took so long - I'm afraid your message arrived during
vacation time here, and got lost in my mailbox.
Anyway, I have tested the patch, and it looks good. My problem has been
resolved, and the test program now gets EINVAL as it should.
Please note that I haven't been able to test 1.27 directly. What I did was:
- Grab uipc_socket.c 1.30 from FreeBSD-current at ftp.cdrom.com.
- Locate the changes related to kern/3925.
- Backport these changes to my current system (2.2-970801-RELENG, with
uipc_socket.c 1.20.2.1).
- And test the new kernel based on these changes.
Below is a context diff against uipc_socket.c 1.20.2.1 - this could
probably go into the RELENG_2_2 branch.
Steinar Haug, Nethelp consulting, sthaug@nethelp.no
----------------------------------------------------------------------
*** uipc_socket.c.orig Tue Dec 3 11:48:58 1996
--- uipc_socket.c Fri Sep 12 00:31:57 1997
***************
*** 893,922 ****
case SO_RCVBUF:
case SO_SNDLOWAT:
case SO_RCVLOWAT:
if (m == NULL || m->m_len < sizeof (int)) {
error = EINVAL;
goto bad;
}
switch (optname) {
case SO_SNDBUF:
case SO_RCVBUF:
if (sbreserve(optname == SO_SNDBUF ?
&so->so_snd : &so->so_rcv,
! (u_long) *mtod(m, int *)) == 0) {
error = ENOBUFS;
goto bad;
}
break;
case SO_SNDLOWAT:
! so->so_snd.sb_lowat = *mtod(m, int *);
break;
case SO_RCVLOWAT:
! so->so_rcv.sb_lowat = *mtod(m, int *);
break;
}
break;
case SO_SNDTIMEO:
case SO_RCVTIMEO:
--- 893,945 ----
case SO_RCVBUF:
case SO_SNDLOWAT:
case SO_RCVLOWAT:
+ {
+ int optval;
+
if (m == NULL || m->m_len < sizeof (int)) {
error = EINVAL;
goto bad;
}
+
+ /*
+ * Values < 1 make no sense for any of these
+ * options, so disallow them.
+ */
+ optval = *mtod(m, int *);
+ if (optval < 1) {
+ error = EINVAL;
+ goto bad;
+ }
+
switch (optname) {
case SO_SNDBUF:
case SO_RCVBUF:
if (sbreserve(optname == SO_SNDBUF ?
&so->so_snd : &so->so_rcv,
! (u_long) optval) == 0) {
error = ENOBUFS;
goto bad;
}
break;
+ /*
+ * Make sure the low-water is never greater than
+ * the high-water.
+ */
case SO_SNDLOWAT:
! so->so_snd.sb_lowat =
! (optval > so->so_snd.sb_hiwat) ?
! so->so_snd.sb_hiwat : optval;
break;
case SO_RCVLOWAT:
! so->so_rcv.sb_lowat =
! (optval > so->so_rcv.sb_hiwat) ?
! so->so_rcv.sb_hiwat : optval;
break;
}
break;
+ }
case SO_SNDTIMEO:
case SO_RCVTIMEO:
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?402.874019532>