Booting off RL01's

utzoo!decvax!duke!chico!teklabs!tekmdp!teklabs!terryl utzoo!decvax!duke!chico!teklabs!tekmdp!teklabs!terryl
Tue Sep 15 14:02:52 AEST 1981


     The real problem(at least the only one I found, actually there
were three) was the boot program that goes on block 0 of the RL. The
boot program was hacked from a 512 byte-format filesystem to a 1k
filesystem; there was a constant WC (stands for word count I assume)
that was changed from -256 to -512; however, in the device driver
part of the program the constant WC was not used; the value -256 was
hard coded in. Another bug that I found was a little more subtle;
the code for relocating the program into high core forgot to take
into account the 16-byte executable header; the code to do this was
there, but because the program is too big, it was commented out and
the code assumed one had already stripped off the 16-byte header.
The third problem was the interrupt vector for the RL in l40.s was
0320 (octal) instead of 0160 (octal) as others have noted. This problem
didn't occur in the boot standalone because it all runs non-interruptable.
     If anyone would like a copy of the working RL boot program I would
be happy to send it.



				Terry Laskodi
				chico!teklabs!terryl
				or
				ucbvax!teklabs!terryl



More information about the Comp.bugs.2bsd mailing list