From owner-freebsd-isp Thu Sep 19 08:54:24 1996 Return-Path: owner-isp Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id IAA05705 for isp-outgoing; Thu, 19 Sep 1996 08:54:24 -0700 (PDT) Received: from brasil.moneng.mei.com (brasil.moneng.mei.com [151.186.109.160]) by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id IAA05680 for ; Thu, 19 Sep 1996 08:54:19 -0700 (PDT) Received: (from jgreco@localhost) by brasil.moneng.mei.com (8.7.Beta.1/8.7.Beta.1) id KAA11109; Thu, 19 Sep 1996 10:53:02 -0500 From: Joe Greco Message-Id: <199609191553.KAA11109@brasil.moneng.mei.com> Subject: Re: [Fwd: Silo Overflows?] To: brandon@glacier.cold.org (Brandon Gillespie) Date: Thu, 19 Sep 1996 10:53:02 -0500 (CDT) Cc: freebsd-isp@FreeBSD.ORG In-Reply-To: from "Brandon Gillespie" at Sep 19, 96 09:40:34 am X-Mailer: ELM [version 2.4 PL24] Content-Type: text Sender: owner-isp@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk > > Ok, I had a similar problem with that card, just a check: Does the irq > > setting is ok in the kernel config file, in the last sio line?. If you > > left out that setting, the probes will check the board with all the > > ports, but as soon as something goes trough the line it will give silo > > overflows. :-). (Yes, it happened when I was just copying the previous > > line, instead of type them, so I forgot to place it there). Hope that > > helps. > > We run a in-house dialup server off a BOCA 16 port board (BB2016). When > we started there were only 8 modems on the board, but as we have expanded > and as they are all used the board continues to freeze up all of the > serial ports off the board. The only way to un-freeze is to shutoff all > getty's on the serial ports, give it a few and turn them back on (via > /etc/ttys). The kernel has been patched in sio.c to set ftl_init to 8 > instead of 14, and we seem to have fewer overflows, but they still occur, > and when they do the entire board freezes. With our increased usage it > is to the point now that there are one to two freezes a day. We need to > find a solution, any suggestions? BOCA did tell us the UART on the card > is fast enough to handle what we are doing, so I dont suspect that as a > problem. I have seen that problem. The interrupt gets stuck "on" - you can see it with a logic probe. We called it "red light syndrome". I never found a solution for it, however you can detect the condition in software: When you connect to a port, and type "AT" to the modem, you do not get a response... but if you hit "" again, you will get the "OK". I hacked up a monitor that detected when the damn thing wedged. ... JG