Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 8 Jul 1999 10:04:49 -0400 (EDT)
From:      Larry Lile <lile@stdio.com>
To:        Phil Regnauld <regnauld@ftf.net>
Cc:        freebsd-tokenring@FreeBSD.ORG
Subject:   Re: Status on Olicom driver ?
Message-ID:  <Pine.BSF.4.05.9907080956350.19285-100000@heathers.stdio.com>
In-Reply-To: <19990708160347.13356@ns.int.ftf.net>

next in thread | previous in thread | raw e-mail | index | archive | help


On Thu, 8 Jul 1999, Phil Regnauld wrote:

> Larry Lile writes:
> > 
> > On Wed, 7 Jul 1999, Phil Regnauld wrote:
> > 
> > > Hi,
> > > 
> > > I've been away for a while -- what's the status on the Olicom drivers --
> > > are ther still problems or are they working now ?
> > > 
> > > (I remind my problem: 4.0-SNAP-19990421 worked, but not 4.0-SNAP-19990527 for example).
> > 
> 	[...]
> 
> > The second and worst is that the PMW kit ceases to call
> > TransmitFrameCompleted.  This causes the driver/card to run out of
> > transmit buffers and there is no way to recover except reboot.
> 
> 	In my case, ifconfig down flushes the packets, and ifconfig up gets
> 	the interface up and running again (until next burst) -- mild
> 	traffic doesn't kill it...

You are suffering under the too many mbufs problem.  All of the available
transmit fragments get used up because you are trying to transmit a
packet with more mbufs than there are total (not available) adapter 
fragments.

> 
> > I can't help but think that since newbus went in on 4/16 that there is
> > some correlation with the second problem.
> 
> 	But 4/21 works... (except for repeated "receive error. (ReceiveStatus=3)"
> 	on the console, and it also says "adapter not ready to come up", or similar,
> 	when I ifconfig the oltr0 for the first time -- but it works anyway).

The only way I can figure out your ReceiveStatus=3 problem is with a
network trace.  

The adapter not ready problem is a race condition that I have to fix in
the probe/attach code.

If you go a little further forward in time you will find that the adapter
will hang on outgoing packets, because the PMW ceases to call
TransmitFrameCompleted.  Although without my new code to get around the
mbuf fragment problem "seeing" it would be sorta pot luck. :(

> 
> > Needless to say I am very frustrated that I haven't been able to fix
> > these problems, but I am still working on them.  I still have a lot
> > of cleaning up to do in the driver but these 3 things have me tied
> > up.
> 
> 	Is this purely FreeBSD related -- i.e.: wouldn't the Olicom people
> 	be able to help a little ?

The author of PMW is looking at some dumps for me already.  It is likely
an interrupted DMA transfer, caused by me failing to protect a PMW call
properly.

Larry Lile
lile@stdio.com



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




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.05.9907080956350.19285-100000>