Adam Agnew <[EMAIL PROTECTED]> writes:

> > Adam Agnew <[EMAIL PROTECTED]> writes:
> >
> > > In other news, ide access in RedBoot is working fine and work is well
> > > underway on the necissary Red Hat eCos Public License version of an
> > > elfloader.
> >
> > Cool.  What was the problem with ide?
> >
> 
> Argh, well this is addressed in a later post today too. My initial
> assumptions were that if I turned on the IDE controllers, left redboot
> running, then reset, the hard drives should already be spun up and it
> would Just Work. There were conversations on this list about a year ago
> (which should be preserved in our archive www.missl.cs.umd.edu/linuxbios
> but i'm going from memory (i'm young)) which stated that in the linux
> kernel case, a good reset was needed until we simply added code to wait
> for spin up. So, my initial attempts at turning the ide controller on
> didn't appear to work, and my later attempts which were much more
> elaborate were no better. It appeared that RedBoot had thought about it,
> and would wait up to 30 seconds provided the ide controllers were on, but
> that turned out to be false (i'm not sure why yet). For now I just have an
> ugly brute force delay in there, but i'm going to go back and just
> implement the *minimum* needed to get it working soon, clean up, submit
> patch.
> >From what I've seen, I think I can deduce that either
> 1) resetting seems to start the spin up process all over again, with no
> advantage in time
> or
> 2) For some reason, IDE controllers need a few seconds to think about it..
> 
> I would hope 2 is false, and 1 seems more likely.

Phoenix has a white paper on how they were doing this in 1997.

http://www.phoenix.com/PlatSS/products/wp.html

If needed I can also dig up a some of the draft ATA standards,
found from references by Andre Hendrick.

Eric



Reply via email to