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

List:       acpi4linux
Subject:    [Acpi] Good news: New strategy should make testing easier
From:       Simon Richter <geier () phobos ! fs ! tum ! de>
Date:       1999-09-05 11:17:20
[Download RAW message or body]

Hi,

we have determined that we would run into a huge problem with the current
approach because we need the VM to initialize the hardware, however we
cannot test the VM from a running system without initializing the
hardware, because the results would not be reliable. Also, having the VM
statically in the kernel would increase the kernel size too much.

Alan Cox has pointed out an interesting idea: We boot with a kernel that
preserves all ACPI tables. Then an userspace program is run, compiling
native "drivers" out of the AML code and appending them to the kernel
image. On the next boot, we have the methods needed to use the hardware
available from the start, and boot into ACPI mode. The kernel size
increases only a few kbytes by this, and the code generation is an
userspace program which can easily be tested.

   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]