Re: EFI/console general question
From: Robert Deininger (rdeininger_at_mindspringdot.com)
Date: 09/09/05
- Next message: Alan Greig: "Re: [OT] New Orleans disaster"
- Previous message: Paul Sture: "Re: Hurd to detail strategic plan Dec 13"
- In reply to: JF Mezei: "EFI/console general question"
- Next in thread: Paul A. Jacobi: "Re: EFI/console general question"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Fri, 09 Sep 2005 10:56:17 GMT
In article <432118C6.D0F44C87@teksavvy.com>, JF Mezei
<jfmezei.spamnot@teksavvy.com> wrote:
>Just trying to understand relationship between CPU and the rest.
>
>When a chip built/designed, is it console/boot medium agnostic ?
Yes, at least for mainstream CPUs.
>Does the CPU simply branch to a specific memory location when it is
>powered on/halted ? If so, it is *theoretically* possible to create a
>VAX board using already fabricated Microvax II chips, but put in an EFI
>console ROM that would have been compiled for VAX and which would
>support Q-BUS access to a hard drive (so it can find the rest of the
>EFI program/os loader) ?
Yes, branch to a specific memory location is a pretty common way of starting.
Alphas are a little fancier; they first load up the instruction cache from
a serial ROM using a 2-pin serial interface. Then start executing
instructions in the cache. This means the CPU HW doesn't need to know how
to talk to memory at first. The SROM code teaches the CPU how to access
memory.
Most Alpha systems have multiple versions of the SROM code, selectable via
a jumper. The "extra" versions enable manufacturing and debugging
functions.
>Could one go one more step and just replace the ROM chip on an existing
>Microvax II board (example) with a new ROM containing an EFI console
>written to support the MVII chip and the specific support chips on that
>board ?
Sure, if the ROM is big enough. (It probably isn't.)
>Or are CPUs FABBed with specific support for a specific console program
>?
>
>
>When a CPU powers up or is halted (or whatever other signal), is it as
>simple as simply branching to a predetermined physical memory address ?
>Or is it far more involved than this ?
It can be more involved, since the CPU may not be able to talk to memory
yet. These days, a secondary processor often does things like control the
power-up sequencing and do the first-pass initialization memory controller
chips.
>OK, different twist to the question:
>
>If intel makes a certain model Pentium chip, could HP build an EFI based
>motherboard/system while Dell would build a BIOS based console, but both
>using the same CPU ?
I don't see why not.
- Next message: Alan Greig: "Re: [OT] New Orleans disaster"
- Previous message: Paul Sture: "Re: Hurd to detail strategic plan Dec 13"
- In reply to: JF Mezei: "EFI/console general question"
- Next in thread: Paul A. Jacobi: "Re: EFI/console general question"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
