Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 25 May 2001 12:56:08 +0200
From:      Andre Oppermann <oppermann@monzoon.net>
To:        stable@freebsd.org
Cc:        forrestc@imach.com, jlemon@flugsvamp.com
Subject:   Re: fxp0 "sleeping" Problem (NOT NEW DRIVER RELATED)
Message-ID:  <3B0E3A48.B05316E4@monzoon.net>

next in thread | raw e-mail | index | archive | help
> In article <local.mail.freebsd-stable/Pine.BSF.4.21.0105241018350.25599-100000@workhorse.iMach.com> you write:
> >The symptoms are as follows:
> >
> >Several times per day the card seems to go to sleep.   We cannot ping
> >through the interface at all (in or out), but pings to the network card ip
> >address from the box comes back (networking is still alive).
> >
> >I have been watching mbufs.  The current output is:
> >
> >mail#  netstat -m
> >492/3536/10240 mbufs in use (current/peak/max):
> >        130 mbufs allocated to data
> >        362 mbufs allocated to packet headers
> >110/2560/2560 mbuf clusters in use (current/peak/max)
> >6004 Kbytes allocated to network (78% of mb_map in use)
> >49385 requests for memory denied
> >3 requests for memory delayed
> >0 calls to protocol drain routines
> >
> >Generally I never see the clusters above 150 or so UNLESS the card is
> >"hung" which I suspect is a symptom and not the cause...
> >
> >We don't need to do anything to have it come back - just wait, and then
> >everything seems just fine like as if nothing had happened at all.
> >It seems to sleep about 5 minutes. I can't tell if it sleeps for a
> >specific length of time or not, but from the missed  ping responses, it
> >seems like it varys from time to time.
> >
> >I am getting NO syslog or dmesg errors when this occurs.
> 
> Hmm.  Can you check your /etc/syslog.conf?  There should 
> be a line similar to:
> 
>         kern.debug                      /var/log/messages
> 
> to capture kernel output.  I'm a little hesitant to believe
> that there's no error output; maybe it isn't configured or 
> syslogd isn't running.

I'm getting the same sort of error on my eval embedded boards. First
both worked fine (4.2R) and somewhen when I was following -stable it
broke the same way Forrest sees it: The 100Mbit LED lights up when I
connect the cable (stays on when I remove it, but that's the same
with standard fxp cards). Although there is some activity on the LAN
(but definatly no SYN attack as someone mentioned) the activity LED
on this board does not light up (it does on standard fxp cards).

Now both my embedded boards are network wise dead because the fxp
chip is appearently in some sort of sleep mode. I tried running
Winblows on the board to see if the Intel Windows driver could get
the chip back alive but nope. The same symtoms, 100 LED on, no
activity.

But always I'm able to ping myself with the interface IP. Ifconfig
always says "no carrier". Setting it to whatever manually doesnt
help at all. Stays the same. The same behaviour with Winblows, can
ping itself but nothing outside.

My embed boards were working perfectly fine with -stable until approx.
two month ago when it happend to both after a cvsup and make world.
Might it be related to the power safe changes that went into fxp?

After giving it a last try with a fresh install of 4.2R which again
showed the same symtoms I was on the way so send the embeds back to
my supplier and tell him he's selling junk... luckily I stumbled
across this thread.

> The other thing is that you're running 100mb/s; the previous
> lockups I've seen with this board were for 10mb/s.  Can you
> send me a dmesg + ifconfig output?

I'll do as soon as I'm in the office. Will hook up a serial console.

-- 
Andre

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3B0E3A48.B05316E4>