1 << 31 cleanup. Eitan Adler pointed out that there has been a
resurrection of the bad idiom in the tree.
sufficient review by miod, kettenis, tedu

Add port multiplier support, has been in snaps for a while with no reported
issues. No actual OKs, but general acknowledgement and 'get it in' from several.

Add some more device status related register definitions.

Correct offset for global interrupt status register.

this diff covers two changes, but theyre very tightly tied together.

implement long sgls by allocating 512 bytes per command, which gives us 7
scatter gather tables to fill in per command. we can now do proper sized
io. commands are still polled though, so its very very slow ;)

since you cant submit more than 128 bytes using direct command submission,
we needed to implement indirect submission too.

this almost worked, but i gave the controller to pascoe before i got it
working. he did the fixes to my dumb dumb mistakes. thanks pascoe. thascoe.

finish resetting the device by sending a soft reset via the post_direct
mechanism, and then read the device signature.

control bits in the prb. shuffle things around a bit...

start defining the scatter/gather bits and the port request blocks (PRBs).

interrupt enable set/clear defns

port control clear bits

Correct some sili port offsets.
ok dlg@

more bits and the %b thing for port set/status

Port Control Set register defs

initial atascsi glue.

define the size and offsets for each ports set of registers

bits for the global control register

oops, 64bit memory bars.

port registers

global register definitions

read and write funcs for the global register space

map the global and port register spaces.

import sili(4) so it can be worked on in the tree. this will support the
silicon image 3124/3132/3531 sata chipsets eventually.

so far this is the autoconf glue, an actual match routine for the 3124, and
interrupt establishment code. it is split up between pci and ic cos there
are cardbus variants of these controllers that we can support in the
future.

thanks to jolan@ for the name, it was better liked than my initial
suggestion of siisl(4).