[LMH]Cold Booting
ford@objs.com
ford@objs.com
Tue May 21 08:06:01 2002
--Boundary_(ID_DDxqs/R0gbjpe/fuhy3b+A)
Content-type: text/plain; charset=us-ascii
Content-transfer-encoding: 7BIT
Daniel Seagraves wrote:
> On Sun, 19 May 2002, Paul Fuqua wrote:
>
> > The last question is about the CPU cold-start - Did the Explorer (either
> > one) have a front-end CPU to load microcode and start the CPU, or was the
> > Explorer CPU the only one? If the Explorer CPU is the only CPU, how does
> > it reload I-Memory without confusing itself?
> >
> > No, no FEP. It probably ran the loading code in high memory and
> > arranged for a reset to happen when that was overwritten. I'll see if I
> > can find a hint of that in some ancient listings, if I can find that
> > box.
>
> I figured out this much - The E1 had a 2Kx56-bit PROM with the
> boot/diagnostic ROM. I have the images of these and I'm working out the
> format now. I think I have it (I can read text in the ROMs) but I don't
> know the entry addresses. It appears to overlay I-Memory at location 0
> until bit 11 (MCR-PROM-Disable) in the MCR is set. I tried to start it at
> 0, but that disassembles to a HALT. (I might still have the format
> wrong.)
>
> There are 7 PROMs, numbered 2236480-03 through 2236486-03.
>
> I started at 0 and ended at 6, dumping 8 bits from each to make a 56-bit
> word. This is a partial dump of the result:
>
> 00000000: 47 44 4f 53, 00 00 06 74, 69 6e 67 20, 00 06 0d 0a ;GDOS...ting ....
> 00000010: 0a 57 61 69, 06 53 6c 6f, 74 20 00 06, 20 00 00 00 ;.Wai.Slot .. ...
> 00000020: 00 00 06 3d, 20 53 6c 6f, 74 02 42 4f, 4f 54 00 00 ;...= Slot.BOOT..
> 00000030: 06 4f 52 3a, 20 00 00 06, 0d 0a 0a 45, 52 52 02 63 ;.OR: ......ERR.c
> 00000040: 65 20 3a 20, 00 02 64 20, 64 65 76 69, 02 63 74 20 ;e : ..d devi.ct
> 00000050: 6c 6f 61 02, 0d 0a 53 65, 6c 65 06 20, 3a 20 00 00 ;loa...Sele. : ..
> 00000060: 00 02 64 20, 6e 61 6d 65, 06 65 6d 20, 6c 6f 61 06 ;..d name.em loa.
> 00000070: 0d 0a 53 79, 73 74 06 20, 00 00 00 00, 00 06 6e 61 ;..Syst. ......na
> 00000080: 6d 65 20 3a, 02 6f 6c 6f, 61 64 20 06, 0d 0a 4d 69 ;me :.oload ...Mi
> 00000090: 63 72 06 00, 00 00 00 00, 00 02 73 74, 73 20 3a 20 ;cr........sts :
> 000000a0: 02 64 65 64, 20 74 65 06, 3d 45 78 74, 65 6e 06 65 ;.ded te.=Exten.e
> 000000b0: 73 74 2c 20, 45 02 20 52, 3d 52 65 74, 02 20 6c 6f ;st, E. R=Ret. lo
> 000000c0: 61 64 2c 02, 4d 3d 4d 65, 6e 75 06 6c, 6f 61 64 2c ;ad,.M=Menu.load,
> 000000d0: 20 02 66 61, 75 6c 74 20, 06 0d 0a 44, 3d 44 65 02 ; .fault ...D=De.
>
> There's ASCII in there (GDOS? ^_^) but it's still somewhat scrambled...
Looks close, though. FYI, Menuboot prints the line:
D=Default load, M=Menu load, R=Retest, E=Extended tests :
Steve
--Boundary_(ID_DDxqs/R0gbjpe/fuhy3b+A)
Content-type: text/html; charset=us-ascii
Content-transfer-encoding: 7BIT
<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
Daniel Seagraves wrote:
<blockquote TYPE=CITE>On Sun, 19 May 2002, Paul Fuqua wrote:
<p>> The last question is about the CPU cold-start
- Did the Explorer (either
<br>> one) have a front-end CPU to load microcode
and start the CPU, or was the
<br>> Explorer CPU the only one? If the Explorer
CPU is the only CPU, how does
<br>> it reload I-Memory without confusing itself?
<br>>
<br>> No, no FEP. It probably ran the loading code in high memory
and
<br>> arranged for a reset to happen when that was overwritten. I'll
see if I
<br>> can find a hint of that in some ancient listings, if I can find that
<br>> box.
<p>I figured out this much - The E1 had a 2Kx56-bit PROM with the
<br>boot/diagnostic ROM. I have the images of these and I'm working
out the
<br>format now. I think I have it (I can read text in the ROMs) but
I don't
<br>know the entry addresses. It appears to overlay I-Memory at location
0
<br>until bit 11 (MCR-PROM-Disable) in the MCR is set. I tried to start
it at
<br>0, but that disassembles to a HALT. (I might still have the format
<br>wrong.)
<p>There are 7 PROMs, numbered 2236480-03 through 2236486-03.
<p>I started at 0 and ended at 6, dumping 8 bits from each to make a 56-bit
<br>word. This is a partial dump of the result:
<p><tt>00000000: 47 44 4f 53, 00 00 06 74, 69 6e 67 20, 00 06 0d 0a ;GDOS...ting
....</tt>
<br><tt>00000010: 0a 57 61 69, 06 53 6c 6f, 74 20 00 06, 20 00 00 00 ;.Wai.Slot
.. ...</tt>
<br><tt>00000020: 00 00 06 3d, 20 53 6c 6f, 74 02 42 4f, 4f 54 00 00 ;...=
Slot.BOOT..</tt>
<br><tt>00000030: 06 4f 52 3a, 20 00 00 06, 0d 0a 0a 45, 52 52 02 63 ;.OR:
......ERR.c</tt>
<br><tt>00000040: 65 20 3a 20, 00 02 64 20, 64 65 76 69, 02 63 74 20 ;e
: ..d devi.ct</tt>
<br><tt>00000050: 6c 6f 61 02, 0d 0a 53 65, 6c 65 06 20, 3a 20 00 00 ;loa...Sele.
: ..</tt>
<br><tt>00000060: 00 02 64 20, 6e 61 6d 65, 06 65 6d 20, 6c 6f 61 06 ;..d
name.em loa.</tt>
<br><tt>00000070: 0d 0a 53 79, 73 74 06 20, 00 00 00 00, 00 06 6e 61 ;..Syst.
......na</tt>
<br><tt>00000080: 6d 65 20 3a, 02 6f 6c 6f, 61 64 20 06, 0d 0a 4d 69 ;me
:.oload ...Mi</tt>
<br><tt>00000090: 63 72 06 00, 00 00 00 00, 00 02 73 74, 73 20 3a 20 ;cr........sts
:</tt>
<br><tt>000000a0: 02 64 65 64, 20 74 65 06, 3d 45 78 74, 65 6e 06 65 ;.ded
te.=Exten.e</tt>
<br><tt>000000b0: 73 74 2c 20, 45 02 20 52, 3d 52 65 74, 02 20 6c 6f ;st,
E. R=Ret. lo</tt>
<br><tt>000000c0: 61 64 2c 02, 4d 3d 4d 65, 6e 75 06 6c, 6f 61 64 2c ;ad,.M=Menu.load,</tt>
<br><tt>000000d0: 20 02 66 61, 75 6c 74 20, 06 0d 0a 44, 3d 44 65 02 ;
.fault ...D=De.</tt>
<p>There's ASCII in there (GDOS? ^_^) but it's still somewhat scrambled...</blockquote>
Looks close, though. FYI, Menuboot prints the line:
<p>D=Default load, M=Menu load, R=Retest, E=Extended tests :
<br><br>
Steve</html>
--Boundary_(ID_DDxqs/R0gbjpe/fuhy3b+A)--