>Don't worry, Tim. I'd probably dismember anyone attempting to do so.

I just tried the following with -current on a machine with an
SMC card:
	we1 at isa0 port 0x300-0x31f iomem 0xcc000-0xcffff irq 10
	we1: SMC8216/SMC8216C Ethernet (16-bit)
	we1: Ethernet address 00:00:c0:6d:bd:7e

(as well as other stuff, including 2 3c905s).

   1.  Build and boot kernel with we[01] enabled and mcd disabled
       (normal GENERIC, sans EISA, sans el0 for paranioa). 
       Works fine.

   2.  Build and boot kernel with both we[01] and mcd enabnled,
       with the suspect mcd "promiscuous" match disabled, but
       otherwise same as above. Also works fine.
       Not suprising: if this broke, or configured an mcd device,
       there's a bug in ISA port accounting.

   3.  Build and boot kernel with we[01] DISabled and mcd ENabled,
       otherwise same as #2.  Works.  Hm. OK so far.

   4.  Build and boot kernel with both we[01] and mcd ENabled.
       we1 still works. Bingo.

   5.  Add el0 to config file. Repeat #3. Works OK.

   6.  Repeat #4 with  el0 still present.   Still OK.

When the mcd driver didnt have port accounting and had an overly-keen
match routine, sure, it was busted. But after the above, I see no
cause for especial concern, certainly not for SMC cards.
Am I missing something?