Re: SMSC LAN9118 Family driver

[Izumi Tsutsui <tsutsui%ceres.dti.ne.jp@localhost>]

> > I'm still afraid some people might complain that the name 'slan'
> > is too generic and ambiguous. (it could imply something like vlan(4))
> > 
> > How about smh(4)?
> > - we already have sm(4) for SMC 91c100/91c96 (FEAST)
> > - SMSC calls 9115/9116/9117/9118  "High Performance" or "Highly Efficient"
> >   http://www.smsc.com/index.php?tid=145
> >   (I guess that's because many people claim the FEAST is
> >    the "worst" controller, as src/sys/dev/ic/rtl81x9.c says ;-p)
> 
> hmm.  I propose smsh(4) though it is trifling.
> 
>   sm(4) -> SMC's ethernet controller
>            ^^
>   smsh(4) -> SMSC's "High *" ethernet controller

IIRC SMSC has spinned out from SMC and we can assume they are
the same network product vendor, but smsh(4) is also fine for me.

> > I wonder if LAN9118_FLAGS_SWAP should be handled by MD bus_space(9) or not.
> 
> I think that I am significant to the use of the function of LSI as an
> approach different from bus_space(9).

IMO, setting WORD_SWAP before checking BYTE_TEST will
cause less confusion, because using bus_space(9) before
setting up LAN9118_WOAD_SWAP to check BYTE_TEST looked a bit weird
(even if it use bswap32() by LAN9118_FLAGS_SWAP flag).

> > BTW, do you have a backend driver for ISA variants?
> > If not, it should not be mentioned in man page yet.
> 
> Should I delete the description concerning ISA?

At least ISA example in SYNOPSIS should be commented out
if there is no actual ISA attachment yet.

BTW, do you have any attachment sources for evbarm? (none is your patch)
If so, they should be in SYNOPSIS instead.

> I should describe 'It operates on the evbarm board.  Maybe works on ISA-
> bus.  However not exists the ISA back end yet'.

The description in the DESCRIPTION section may be okay, but
it's also better to mention "no ISA attachment yet" in BUGS section
otherwise users will be confused.
---
Izumi Tsutsui