Re: linux-kernel-digest V1 #104

alex (alex@bach.cis.temple.edu)
Wed, 5 Jul 1995 20:16:48 -0400 (EDT)


Hi,

On Wed, 5 Jul 1995, Mark H. Wood wrote:

> That's approximately what happens in OpenVMS. The bootstrap code has its
> own drivers that are really (comparatively) stupid, but very rugged and
> able to talk, after a fashion, with any bootable device. These drivers
> are kept around (kernel-read-only, of course!) and used by the dump
> code. (There is usually a separate, preallocated dump file which the
> boot code locates and opens at startup, so that the dump code doesn't
> have to depend on the state of the full-blown file system code. Because
> it is open, you can't delete the file out from under the OS, which used
> to cause interesting disk corruption problems before that feature was
> implemented.)

No, no, no - wrong!!!!!

[Alpha Support Group Hat on]

That all was overly simplified and I'd say even incorrect (Sorry). When
Alpha boots, powered on, or halts the control is being passed to a
special program that controls system console. That programs is always
present and it eats about 2Mb of system memory. The "console", in
addition to being able to perform some basic system tests and play with
some devices and firmware, knows how to start a boot sequence from any of
"bootable" devices using any of "bootable" protocols. From what I
understand, the PAL code is written in the way that it cannot screw up
those 2 Mb of RAM.

[Alpha Support Group HAT off]

Best wishes,
Alex

============================================================================
Alexander O. Yuriev Email: alex@bach.cis.temple.edu
CIS Labs, TEMPLE UNIVERSITY WWW: http://bach.cis.temple.edu/personal/alex
Philadelphia, PA, USA

PGP Key: 1024/ADF3EE95 Fingerprint: AB4FE7382C3627BC 6934EC2A2C05AB62

Unless otherwise stated, everything above is my personal opinion and not an
opinion of any organisation affiliated with me.
=============================================================================