From owner-freebsd-hardware  Tue Apr  2 09:27:32 1996
Return-Path: owner-hardware
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id JAA10409
          for hardware-outgoing; Tue, 2 Apr 1996 09:27:32 -0800 (PST)
Received: from sneezy (sneezy.sri.com [128.18.40.6])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id JAA10402
          for <hardware@FreeBSD.ORG>; Tue, 2 Apr 1996 09:27:29 -0800 (PST)
Received: by sneezy (SMI-8.6/SMI-SVR4)
	id JAA29670; Tue, 2 Apr 1996 09:22:23 -0800
Date: Tue, 2 Apr 1996 09:22:23 -0800
From: nate@sneezy.sri.com (Nate Williams )
Message-Id: <199604021722.JAA29670@sneezy>
To: "Brett Glass" <Brett_Glass@ccgate.infoworld.com>
Cc: Michael Smith <msmith@atrad.adelaide.edu.au>, hardware@FreeBSD.ORG
Subject: Re: Cannot boot after install
In-Reply-To: <9603018283.AA828378286@ccgate.infoworld.com>
References: <9603018283.AA828378286@ccgate.infoworld.com>
Sender: owner-hardware@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

[
Been off the air on a business trip.  Don't expect rapid replies as my
email machine is 3000 miles away, and the lag is no fun.
]

> > Note that there are some interesting comments in the ATA spec I was
> > referring to regarding power-saving modes.  Basically, the "lowest" a
> > drive is allowed to go without an explicit command should still respond,
> > as per normal, to commands, however response may be delayed by up to 30
> > seconds.  If your drive is responding with an error condition (indicating
> > that it's gone into 'sleep' mode), it is violating the spec.

...
> However, FreeBSD literally locks up until it sees a response from
> the drive. This is a problem with FreeBSD. It should not busy-wait in
> the kernel, forsaking all other tasks, until it gets a response. This is,
> after all, a multitasking OS! ;-)
> 
> The kernel hack I did (turning off the power-saving mode) is appropriate
> for desktop machines; in fact, for them, it's a good idea. But it's NOT
> appropriate for laptops and other low-power applications. The kernel needs
> to be fixed so that the drive *can* spin down without locking up the entire
> machine.

FWIW, the same problem occurs on my laptop occasionally when it goes
into APM mode.  Most of the time it'll come up fine, but occasionally
it'll lockup tight and I have to power-cycle the box.  This *is* a
problem, but unless someone dives in (not me) and figures it out it will
continue to to crop up in FreeBSD.


Nate