status on salto page

Juergen Buchmueller pullmoll at
Sat Jun 30 02:21:40 PDT 2007

On Fri, 29 Jun 2007 22:20:34 -0700
Al Kossow <aek at> wrote:

> the failure it's trying to print out is "BFSInit failed"
> basic file system initialization
> neat to see the cursor starting to appear..

Yep, it all plays together, and it all depends on the perfect correctness of
the emulation of the various carries.

I e.g. realized from the 74181 description, which suggests that it handles
subtract different from add operations, that indeed it looks like the carry
is inverted on subtracting arithmetic operations. Implementing this took me
one step forward: booting without the system trying to write to disk,
BITBLTs to the extended memory...

One more important bug squashed was that the bus source R<- (load R) does
_not_ load R in the late phase, if the F2 is DNS! This is because the F2
DNS entirely disables the write to the register file if the no-load flag
in the IR is set (this was for you, Eric).

Reading the schematics over and over again seems to be the only way to
track down the omissions or discrepancies in the Hardware Manual's
description of how things should work.


More information about the Altogether-devel mailing list