24 Feb
2003
24 Feb
'03
3:41 a.m.
You are right: 000137. The funny thing is, I had it right the first time.
I'll go update the web page.
On the other header fields, I agree with your idea, but for some reason, I
have this vague recollection that there was a reason for what I did with
them. It's always a bad idea to second guess something you did a year ago
and "fix" it (see the 000137 thing). It's working, so I'd best leave
it
alone until it's proven to be broken. I think I remember setting it up that
way after reading through the source for some boot loader.
Hmm, yes, it was the "boot" program. VTserver squirts down a small loader
via ODT, which then loads boot.dd. Same as the boot program that the 512
byte rauboot program loads. This one:
73Boot from ra(0,0,0) at 0172150
: ra(0,0,0)unix
Boot.dd knows enough to chat with the console so you can tell it what file
you want off the virtual tape or hard drive, and it does some decoding of
the header, but it's limited. Maybe I should go read that code again.
--
Jonathan Engdahl
http://users.safeaccess.com/engdahl
"The things which are seen are temporary,
but the things which are not seen are eternal." II Cor. 4:18
----- Original Message -----
From: "Michael Davidson" <michael_davidson(a)pacbell.net>
To: "Jonathan Engdahl" <j.r.engdahl(a)adelphia.net>
Cc: <cctech(a)classiccmp.org>; <pups(a)minnie.tuhs.org>
Sent: Sunday, February 23, 2003 4:40 PM
Subject: Re: [pups] Re: PDP11 - RQDX1 standalone format utility? --> XXDP
Jonathan Engdahl wrote:
>I documented the procedure to extract a program from XXDP and put a
header
>on it so that you can boot it from VTserver or a
UNIX disk. It has been a
>long time since I did this to ZRQCH0, so I practiced on the program that
I
>think Christos needs (ZRQBC1) and sent it to him.
>
>There's a lot I don't know about the a.out header, so if someone can
improve
my method, let
me know.
http://users.safeaccess.com/engdahl/xxdp.htm#hackXXDP
There isn't very much to know about the a.out header,
and what you are doing looks OK although I would probably
have set a_text to 160000 and left a_data and a_bss as 0.
While I am not really familiar with the BSD boot code that
vtserver uses when loading a program I am almost certain that
it will use the entrypoint address in the a.out header and
not just jump to address 0. (remember it's an actual program
loader that understands a.out format, not just something that
only knows how to load a boot block and jump to it)
This is confirmed by the fact that the code you are patching
into the image at address 0 is wrong if you want it to get
you to the restart address.
In your example you have:
000020/ 000167
000022/ 145702
If this was loaded at address 0 in memory and the boot code
jumped to address 0, you would end up at 145706 *not* 145702
(of course, you might very well get lucky and still land on
an instruction boundary, but you would have skipped the first
2 words of the restart code).
If this code were really necessary (and I'm 99.99% sure that
it isn't) you would want either:
000137
145702
or:
000167
145676
_______________________________________________
PUPS mailing list
PUPS(a)minnie.tuhs.org
http://minnie.tuhs.org/mailman/listinfo/pups