From owner-freebsd-hackers Tue Jul 18 12:15:54 1995 Return-Path: hackers-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id MAA07460 for hackers-outgoing; Tue, 18 Jul 1995 12:15:54 -0700 Received: from kitten.mcs.com (Kitten.mcs.com [192.160.127.90]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id MAA07452 for ; Tue, 18 Jul 1995 12:15:52 -0700 Received: from Jupiter.mcs.net (Jupiter.mcs.net [192.160.127.89]) by kitten.mcs.com (8.6.10/8.6.9) with ESMTP id OAA12857; Tue, 18 Jul 1995 14:15:50 -0500 Received: (from karl@localhost) by Jupiter.mcs.net (8.6.11/8.6.9) id OAA00718; Tue, 18 Jul 1995 14:15:49 -0500 From: Karl Denninger Message-Id: <199507181915.OAA00718@Jupiter.mcs.net> Subject: Re: SCSI disk wedge (fwd) To: bugs@ns1.win.net (Mark Hittinger) Date: Tue, 18 Jul 1995 14:15:48 -0500 (CDT) Cc: hackers@freebsd.org In-Reply-To: <199507122036.QAA16323@ns1.win.net> from "Mark Hittinger" at Jul 12, 95 04:36:35 pm X-Mailer: ELM [version 2.4 PL24] Content-Type: text Content-Length: 1930 Sender: hackers-owner@freebsd.org Precedence: bulk > > > > Can you be more specific about which controller and/or drives you are > > > seing this with? Trying to recall the mail for the last month with > > > ``disk channel hangs'' as the keyword doesn't work for me, my poor old > > > mind is not what it use to be :-) > > Aha1742 and AHA 27xx -- both EISA. > > On the 27xx driver we get an error, on the 1742 the failure is silent. > > Other than that, the behavior is identical. > > I'm running two p60 32mb eisa with AHA27xx. I have some HP drives and > some seagate barracudas. Going to ditch the HP soon. I beat up on them > pretty heavily and do not see a problem. > > My motherboards are those weirdo broken DMA boards so even though I have > 32mb I have to use bounce buffers. Just for grins you might try a kernel > with bounce buffers and see if you get the hangs. > > Another dumb idea is that something else is causing the problem and the > symptom is a scsi timeout. I've made this mistake and chased a wild > goose. > > I don't wanna open the box and find out what my adaptec firmware rev > level is but...... > > The only other i/o in the boxes are the EISA style 3c509. > > I am running the -current build from June 30'th. > > Regards, > > Mark Hittinger > bugs@win.net Uh, I'm not betting on the disk wedge any longer. My bet is on a VM problem of some kind. Peter Wemm has provided some code which has *changed* the symptoms on the 1742 systems from a hang to a crash with a VM problem (freed objects that are already free). I'm going to continue to work this. -- -- Karl Denninger (karl@MCS.Net)| MCSNet - The Finest Internet Connectivity Modem: [+1 312 248-0900] | (shell, PPP, SLIP, leased) in Chicagoland Voice: [+1 312 248-8649] | 7 Chicagoland POPs, ISDN, 28.8, much more Fax: [+1 312 248-9865] | Email to "info@mcs.net" WWW: http://www.mcs.net ISDN - Get it here TODAY! | Home of Chicago's only FULL AP Clarinet feed!