[prev in list] [next in list] [prev in thread] [next in thread] 

List:       acpi4linux
Subject:    Re: [Acpi] Re: ACPI "thin driver" patch (against 2.3.17)
From:       Simon Richter <geier () phobos ! fs ! tum ! de>
Date:       1999-09-10 16:40:55
[Download RAW message or body]

On Fri, 10 Sep 1999, Linus Torvalds wrote:

> I don't buy for a moment that we don't have a functional system: it will
> have to be fairly functional just to load the kernel off the harddisk.

Yes, but there is a reason why you need the IDE driver linked statically
into the kernel if you wnat to boot off an IDE harddisk, and you also need
the PCI enumerator if the controller is a PCI one. The same goes for ACPI.
You will need ACPI enum if you want to access an ACPI IDE controller. This
is why we need this stuff in kernelspace later.

> Ergo, the harddisk controller has already been initialized, and the
> argument that you need ACPI to do initialization has to be a obvious red
> herring.

Nope. It has been initialized to read data. It has not been initialized to
do power saving. IDE has been a standard since years, I don't believe the
BIOS programmers changed their code to init those registers, because this
is not required in the spec and they can use their code on more different
chipsets if they do not use these registers. So we need to reinitialize
the devices, so that we can save power. If we use the AML methods without
initializing these registers, we risk data loss and weird crashes.

   Simon

PGP public key available from ftp://phobos.fs.tum.de/pub/pgp/geier.asc
 Fingerprint: 10 62 F6 F5 C0 5D 9E D8  47 05 1B 8A 22 E5 4E C1

[prev in list] [next in list] [prev in thread] [next in thread]