------- Start of forwarded message -------


A few comments about the pcmcia interface discussions:

	1. I also agree that it is pretty important to get a
	   pcmcia interface into the kernel.  One has been floating
	   around for 9 plus months.  This is too long.  It makes
	   it hard on people who want to do further development on
	   laptops, support new pcmcia devices ...

	2. "The pcic-chip can have only 2 devices. 
	    And unless I goofed somewhere the limit is on the chip-driver and not
	    in the pcmciabus framework."
	   This statement is incorrect.  The Intel 82365 architecture
	   was setup to support 8 slots.  I believe that you are
	   correct, that an individual chip only supports 2 slots.
	   There is also some magic on how you address the device
	   registers for a slot.  The first 4 slots are accessed thru
	   ports 0x3e0 & 0x3e1.  The second 4 slots are accessed thru
	   ports 0x3e2 and 0x3e3.  (For the curious each slots has 0x40
	   devices registers; 4*0x40 = 256)  The bottom line is that
	   an IBM 701C (or 755) has two pcmcia slots (the first two
	   addressed at 0x3e0 & 0x3e1).  The IBM Dock II also has two
	   pcmcia slots (these are addressed as the first two slots at
	   0x3e0 & 0x3e1).  So I would say that slots 0, and 1 and 4
	   and 5 are available.


	3. ... The main reason for it's absence is, that you need to modify all drivers
	   used with PCMCIA, to not crash or hang the system if the card is removed.
	   This is going to be pretty near impossible.  You have to
	   guard every port reference that is used anywhere in the
	   driver to accept the possibility that the card was pulled
	   out and the value is bogus.  You then have to get out of
	   whatever nesting of routines the code is in.


	4. Last point.  If a user level program is necessary to
	   turn on pcmcia ethernet cards, (I got this impression from
	   previous comments.) please don't forget to put this program
	   on the install floppy.