From mboxrd@z Thu Jan  1 00:00:00 1970
Message-Id: <53F6A637-CCBC-4E69-ACDE-EAB37F9B4BCF@ar.aichi-u.ac.jp>
From: Kenji Arisawa <arisawa@ar.aichi-u.ac.jp>
To: Fans of the OS Plan 9 from Bell Labs <9fans@9fans.net>
In-Reply-To: <8fd6e1a17bbdf42b64aca866573bc38c@quanstro.net>
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v926)
Date: Mon, 17 Nov 2008 21:04:14 +0900
References: <8fd6e1a17bbdf42b64aca866573bc38c@quanstro.net>
Subject: Re: [9fans] broken smtpd
Topicbox-Message-UUID: 47590938-ead4-11e9-9d60-3106f5b1d025

Hello,

Today, I observed smtpd that was in endless loop and then broken.



ar% ps
...
none          20392    0:01  28:27      264K Running  smtpd
...



ar% ACID 20392
/proc/20392/text:386 plan 9 executable
/sys/lib/acid/port
/sys/lib/acid/386
acid: lstk()
receiver(path=0x41f58)+0x1c5 /sys/src/cmd/upas/smtp/smtpd.c:625
	sender=0x379bf
	rcpt=0x41ad8
zzparse()+0x2c9 /sys/src/cmd/upas/smtp/smtpd.y:773
	save1=0x0
	save2=0x0
	save3=0x0
	save4=0x0
	zzstate=0xf
	zzchar=0xffffffff
	zzs=0x3f36b
	zzp=0xdfffe41c
	zzn=0xfffffc18
main(argv=0xdfffefa4,argc=0x0)+0x261 /sys/src/cmd/upas/smtp/smtpd.c:194
	netdir=0xdfffefbc
	_argc=0x6e
	_args=0x376c2
	buf=0x0
_main+0x31 /sys/src/libc/386/main9.s:16
acid: src(receiver+0x1c5)
/sys/src/cmd/upas/smtp/smtpd.c:625
  620
  621		logged = 0;
  622
  623		/* forwarding() can modify 'path' on loopback request */
  624		if(filterstate == ACCEPT && fflag && !authenticated &&
forwarding(path)) {
 >625			syslog(0, "smtpd", "Bad Forward %s (%s/%s) (%s)",
  626				s_to_c(senders.last->p), him, nci->rsys, s_to_c(path));
  627			rejectcount++;
  628			reply("550 5.7.1 we don't relay.  send to your-path@[] for "
  629				"loopback.\r\n");
  630			return;


I don't observe "Bad Forward" of this mail in /sys/log/smtpd.

After some amount of time,

ar% ps
...
none          20392    0:02  44:57      264K Broken   smtpd
...

acid: lstk()
errstr()+0x7 /sys/src/libc/9syscall/errstr.s:5
syslog(logname=0x3760a,cons=0x0,fmt=0x37610)+0x20 /sys/src/libc/9sys/
syslog.c:46
	err=0x616d2f27
	d=0x0
	ctim=0x0
	buf=0x0
	p=0x0
	arg=0x0
	n=0x0
catchalarm(msg=0xdfffe1ac)+0x5d /sys/src/cmd/upas/smtp/smtpd.c:68
	rv=0x1
notifier+0x30 /sys/src/libc/port/atnotify.c:15
acid: src(syslog+0x29)
/sys/src/libc/9sys/syslog.c:46
  41		Dir *d;
  42		char err[ERRMAX];
  43
  44		err[0] = '\0';
  45		errstr(err, sizeof err);
 >46		lock(&sl);
  47
  48		/*
  49		 *  paranoia makes us stat to make sure a fork+close
  50		 *  hasn't broken our fd's
  51		 */
acid:


I hope the above information is helpful to fix the problem.

Kenji Arisawa


On 2008/11/06, at 21:59, erik quanstrom wrote:

>>> Is this phenomenon specific to me?
>>
>> No, it isn't:
>>
>> term% ps|grep Broken
>> none            917    2:08  42:50      264K Broken   smtpd
>
> impressive!
>
> i don't see this problem, but i have implemented
> some fairly agressive spam control.  perhaps this
> shows up after i tend to hang up.
>
> (it is possible to use nupas smtpd with regular old
> upas.)
>
> would anyone mind sending me a full backtrace?
>
> - erik
>