Subject: Re: CD control commands
To: Andrew Gillham <gillham@vaultron.com>
From: Peter Fairbrother <peter.fairbrother@ntlworld.com>
List: tech-kern
Date: 10/07/2001 22:34:14
Yes. 

I use a high-priority process to create a 56kbit/s continuous rng stream,
which inputs among other things CD latency, so I'm reading the CD all the
time anyway (actually several processes, I'm not too good with threads) - it
takes up a big chunk of processor, but that's ok, the OS has only one user.

I'll just add a test to that, and not lock the drive when the CD is mounted
- should be an easy hack.

Thanks everyone.

-- Peter

> Andrew Gillham wrote:

> On Sun, Oct 07, 2001 at 01:35:39PM +0100, Peter Fairbrother wrote:
>>> You may be able to do what you want in hardware though; if you can
>>> disconnect
>>> the eject switch from the CD circuitery and connect it in place of the ATX
>>> power switch.
>>> Or maybe just remove the eject button ?
>> 
>> It's for a public program <www.m-o-o-t.org> to be used on many machines, so
>> regrettably that wouldn't be practical :)
> 
> You might be able to create a kernel thread that polls the CDROM frequently.
> (or polls it when it is not being actively read from)
> If you think about the "Automatic Insert Notification" that Windows does, it
> is basically just polling the media status on the drive at an interval.
> 
> If you get a "media not inserted" status, power off. :)
> 
> -Andrew