Date: Tue, 14 Sep 1999 06:21:58 +0800 (CST) From: freedom@csie.nctu.edu.tw To: FreeBSD-gnats-submit@freebsd.org Subject: kern/13740: wrong IP statistics Message-ID: <199909132221.GAA12704@liberty.dorm11.nctu.edu.tw>
next in thread | raw e-mail | index | archive | help
>Number: 13740
>Category: kern
>Synopsis: wrong IP statistics
>Confidential: no
>Severity: serious
>Priority: high
>Responsible: freebsd-bugs
>State: open
>Quarter:
>Keywords:
>Date-Required:
>Class: sw-bug
>Submitter-Id: current-users
>Arrival-Date: Mon Sep 13 15:30:00 PDT 1999
>Closed-Date:
>Last-Modified:
>Originator: Tan Koan-Sin
>Release: FreeBSD 4.0-CURRENT i386
>Organization:
NCTU, Taiwan
>Environment:
While doing some network experiment, we found that we
always see
0 output packets dropped due to no bufs, etc.
from the output of "netstat -s". It's unbelievable, because
we generated heavy traffic.
>Description:
The following code fragment is buggy.
(line 362-370 of $Id: ip_output.c,v 1.90 1999/05/04 16:20:33 luigi Exp$)
/*
* Verify that we have any chance at all of being able to queue
* the packet or packet fragments
*/
if ((ifp->if_snd.ifq_len + ip->ip_len / ifp->if_mtu + 1) >=
ifp->if_snd.ifq_maxlen) {
error = ENOBUFS;
goto bad;
}
1. I don't think we should check queue length here. Because even
the condition is true. The output queue could have enough space
when the packet is in if_output().
2. Even 1. is wrong, there should be
IF_DROP(&ifp->if_snd);
before
error = ENOBUFS;
>How-To-Repeat:
Just let some heavy UDP traffic compete for the same output queue.
>Fix:
Please remove the code fragment, or add IF_DROP.
>Release-Note:
>Audit-Trail:
>Unformatted:
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-bugs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199909132221.GAA12704>