*BSD News Article 40867


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!munnari.oz.au!spool.mu.edu!howland.reston.ans.net!pipex!uunet!newsflash.concordia.ca!news.mcgill.ca!homer.cs.mcgill.ca!fox
From: fox@cs.mcgill.ca (Colin BRADLEY)
Newsgroups: comp.os.386bsd.questions
Subject: Odd IDE paging under NetBSD-1.0/XF3.1 - old wd.c bug reborn?
Date: 15 Jan 1995 20:31:09 GMT
Organization: SOCS, McGill University, Montreal, Canada
Lines: 39
Distribution: world
Message-ID: <3fc0me$jv@homer.cs.mcgill.ca>
NNTP-Posting-Host: marge.cs.mcgill.ca


Hello, everyone.

	I'm running NetBSD-1.0 on a 486/66 Micronics VLB with
	16 megs, a Western Digital Caviar 340 meg IDE plugged into the
	builtin IDE controller, and I have a Seagate 3600N 525 meg
	plugged into an Ultrastor34f VLB controller.

	The machine is from Gateway 2k.

	When I run intensive apps (like a large xv histogram 
	equalization (or whatever it is)), the thing starts doing
	a little paging. The swap space is on the IDE. In the old 
	days, before the illustrious wd.c patch (under 0.9), the 
	drive would sometimes hang, forever timing out (or whatnot) 
	and the	disk active light would remain on until reboot. 

	What is happening now is that the light sorta flashes on,
	waits a sec, then flashes on again, as in: *blink* - *blink* -....

	And this continues for a while, and then it finds itself and
	continues as normal. It is reminiscent of the old timeout
	problem.

	Is anyone familiar with this problem? 

	It is quite annoying, since during the blink phase, the machine
	is effectively locked up (this being an IDE drive)....Is there
	a patch or some other workaround (short of getting a SCSI 
	replacement)?

	I'd be grateful for any tips or advice.

Regards, Colin.

-- 
fox@cs.mcgill.ca					
Colin Bradley 
panic: can't find /