On Sat, 27 Mar 1999, Michael Kukat wrote:

> Seems that every 4000 has this, add the 4000 VLC (KA48) to your list.

Done.

> I found something in the main driver, which detects the derivates of
> the chip,

The ESP100/100A/200 detection?  It's nice, but (mostly) irrelevent until
vax gets DMA (the differences beteen the three are FIFO depth and DMA
bugs).

> don't we have a nice solutions do see the difference between the 53c80
> and the 53c94?

*laugh* Yes.  Try to generate an interrupt for the '94 -- if it doesn't
work, it will probably lock up ;)

Apart from that, I can't see a good way to identify a '94 100% of the
time.  The other ports assume presence by platform ID.

> Don't just assume, these boxes *MUST* have one, it's nicer if we
> really look, if it's there.

Huh?  Assume that it is there, then try to generate an interrupt.  If the
interrupt does not occur, then the device is not present.  Ragge's new
device detection scheme is good, I think -- yes, it could be refined a
bit, but it's much better than the original scheme.

I don't trust device signature scans.  I've dealt with video drivers that
misidentify because the signature that they are looking for is fooled by
another chip.

> But to give you the info which *MUST* have it: KA46, KA48 and KA49 (VS 4000
> /60, VLC, /90)

Thanks.  That's exactly what I was looking for 8-)

--	Chris (wileyc@sekiya.twics.co.jp)