Re: When a computer start


Bill Leary answered:

....
For example, on a 32 bit machine at 0xFFFFFFF0.
Right. One could built a mainboard with the BIOS ROM up there :)
When I've looked at some non-PC, but x86, based machines with an ICE
I've seen their (E(P))ROMs up there.

**"Assuming "PC" means "Personal Computer" in the current ..."** :)
Yes, it's reasonable for non-PC machines to have either a gated register
or a tiny ROM or even a CMOS-RAM(boot option) on top of the memory range.

I don't know IA64 nor Penryn, all others start in 16-bit RealMode
so the motherboard (chipset) remaps this 4GByte-16 to 000FFFF0h
which again is 16 bytes below, but now below the 1 MB border.

Again, on some of the non-PC but X86 machines I note above, yes, this
occurred. The code there did either a "big-real" mapping and jumped to
the
ROM up at the top, or went into protected mode and jumped there. On every
PC type machine, what you describe is what I've always observed. Again,
exceptin the 64 bit versions which I've not yet had an opportunity to go
at
with an ICE.

At least AMD64 work (almost) the same way as 32-bit CPUs on RESET.

And that's why all BIOS-ROMs have its RESET-routine entry point
at F000:FFF0 (or FFFF:0000, or any that means the same address),
and start with a far jump to initialise CS:IP.

I once checked the memory contents below the top 4GB assuming
a mirror of the BIOS ROM, but there was nothing at all (all FFh)
even with PCI mem-mapping disabled.

Did you do this with code on the machine, or with an ICE?

I checked it right after the BIOS finished.
My old in-circuit-emulator wont work on modern CPUS anymore,
hard to imagine a 1 meter, 2 GHz ICE cable ...
And when I look at various versions of PC-BIOS startup code,
then it's obvious that it can execute without memory and stack,
but then all references rely on a valid/initialised 16-bit CS.
So there isn't much that could be done before the first far-junp.

On the machines I note above, and on the one or two PCs I used an ICE on,
I *DID* see the ROM repeated at the top of memory DURING reset, but on
the PC machines, it disappeared when the main-board was released from
reset. I guess I'd assumed that was normal, for a PC, but now you've
got me wondering.

What you see 'during reset' will never reach the CPU ;)
The interesting point would be after the raising edge of the RESET-line,
IOW: what are the contents of AD0..32(..52) when the CPU becomes active ?
btw: x86 CPUs idle for a few hundred cycles after RESET is released.

Where could you connect your ICE ?
my old workstation got a CPU-socket-link-cable.
Or are you talking about a BUS-scanner connected to a PCI-slot ?
__
wolfgang



.

Relevant Pages

  • Hard core dec VT100 troubleshooting question (long)
    ... All power supply voltages are always good. ... and sure enough one bit was stuck low on one rom (and yes I reseated ... that could be programmed as high true or low true, ... reset (either at power on as generated by the RC circuit connected to ...
    (comp.terminals)
  • Re: PII vs PIII
    ... >> What CPUs were in those machines and what did you use them for Lane? ... >> breaking the pins off CPUs, a move always seen as being irreversible. ... I managed to break off the wrong pin. ...
    (comp.os.linux.hardware)
  • Re: Need help to reanimate an AT&T AT&T 3B1/7300 UNIX-PC
    ... There was only one Sun that I can recall. ... Ultra-60s (dual 450 MHz CPUs) as my desktop machine. ... which are 64-bit machines native -- with the ability to run ... That is part of my first unix box -- a Cosmos CMS/UNX running on an 8 ...
    (comp.sys.3b1)
  • Re: reset IE to default settings
    ... >>> machines with multiple OS's, not all of them attached to a network ... >>> though all machines will have at least IE6. ... >>> I would like to reset IE to all default values in all areas ... ... >>> without user interaction other than maybe a one button prompt at most. ...
    (microsoft.public.windows.inetexplorer.ie6.browser)
  • Re: Turing machines
    ... would be having programs in RAM instead of ROM. ... finite-state machines that are only allowed to work in one way). ... Depending on the design problem, one could easily try out some patterns ... a finite-state machine, only that the states here are more difficult to ...
    (comp.ai.philosophy)