<< Previous Message Main Index Next Message >>
<< Previous Message in Thread This Month Next Message in Thread >>
Date   : Sun, 04 Jan 2015 16:53:02 +0000
From   : public@... (Daniel Beardsmore)
Subject: Serial console active at power-on

On 2015-01-04 16:23, J.G.Harston wrote:
> If you remove the keyboard a BBC will hang as it "sees" that both the 
> Ctrl and Shift keys are being pressed, and what happens if both the Ctrl 
> and Shift keys are being pressed?
> ...
> On RESET it detects if there is no keyboard, then intercepts KEYV so 
> that when the Shift and Ctrl keys are tested it returns "Not Pressed". 
> Of course, if there is no keyboard, then the BBC also "sees" that all 
> the keyboard links are made, so does a Shift-Break.

This seems odd.

So, to clarify the suggestions this implies:

a) The keyboard test is run towards the end of POST, *after* all ROMs have
been interrogated

b) Service ROMs are permitted to participate in POST
, and both detect failure states, and prevent the OS from halting on failure
c) The keyboard access code reports that every key is pressed and every link
is made if it can't actually communicate with the strobe control and multiplexer
chips on the keyboard itself

Holding ctrl+shift at power on causes a different behaviour to unplugging 
the keyboard. The former pauses the OS in MODE 7 before POST completes, 
while the latter hangs the machine in what seems to be MODE 0 or 3 (the 
cursor is clearly a lot narrower). I can see how intercepting KEYV will 
prevent ctrl+shift pausing the OS start-up process, but how does this affect
the situation where the keyboard isn't detected? Holding down as many keys
as I can doesn't cause POST to hang (as it can on a PC with a stuck key failure
error).
<< Previous Message Main Index Next Message >>
<< Previous Message in Thread This Month Next Message in Thread >>