In message <200506300230.j5U2U1sS014978@guild.plethora.net>, Peter Seebach writ
es:
>Oh!  That's a very likely guess.  In particular, it fits with the very
>different addresses I was seeing in Windows.  Thanks, I'll try that.

The good news is that this fixes the problem, the bad news is that new I get
panics when it tries to attach uhub3 on an ehci controller, with some message
about usb_add_event and a register value somewhere over RBUS_MIN_START.



I still get some errors in dmesg:

cbb0 at pci2 dev 0 function 0: vendor 0x1180 product 0x0476 (rev. 0xa8)
cbb0: can't map socket base address 0x50000000
cbb0: can't map socket base address 0x1: io mode
cbb1 at pci2 dev 0 function 1: vendor 0x1180 product 0x0476 (rev. 0xa8)
cbb1: can't map socket base address 0x50100000
cbb1: can't map socket base address 0x1: io mode
cbb0: interrupting at irq 11
cardslot0 at cbb0 slot 0 flags 0
cbb1: interrupting at irq 11
cardslot1 at cbb1 slot 1 flags 0
cbb0: wait took too long 0.199987s
cbb0: wait took too long 0.199958s

But it now works enough to probe devices... Unfortunately, the one USB
controller I have handy doesn't work.  I'll test more when I get home and
have more options.  (I'm actually, ironically, only trying to get the USB
card "on" so I can use it as a power source.)

-s