On Thu, 8 Oct 1998, Greg A. Woods wrote:

> I don't know which specification you might be thinking of, but the PCIMG
> PCI/ISA specification only allows for 4 PCI slots.

Interesting. I must be behind the times, then, because my MindShare
book, which covers only up to PCI 2.1, claims that one can add
several additional slots or extend the bus length reliably if one
is willing to reduce the bus speed appropriately.

What is the version of the spec that reduces the limit to four
slots, and what section is this limit in?

> The specifications for the 2115[0234] PCI-to-PCI bridge chips also show
> that 4 devices per PCI slot is the maximum.

Last I checked, we were talking about the PCI bus, not a particular
chipset's implementation of it. If your scheme does not support
anything but DEC's 2115[0234] chipsets, please make that clear.

> Of course there's nothing
> preventing buses from being chained together with such bridges, so each
> card might have as many as 9 devices....

If you count bridges on a card, you can have far more than nine
devices. Each slot has a single package on it. However, a package
may have up to eight devices in it. None of this so far involves
bridges.

If you wanted to go to the extreme, you could have a package with
eight bridges in it, one at each device position. (This is still
one PCI package, in one slot.) Each of these could have up to four
packages attached, for a total of 32 packages behind those eight
bridges, and each of those 32 packages could have up to eight
devices each, giving a total of 256 devices on the one card.

cjs
--
Curt Sampson  <cjs@portal.ca>  604-257-9400    De gustibus, aut bene aut nihil.
Any opinions expressed are mine and mine alone.
The most widely ported operating system in the world: http://www.netbsd.org