On Sat Jan 11 2003 at 21:05:36 +0100, Ignatios Souvatzis wrote:
> a) to reliably reset it (there's the WDCTL_RST bit)

Yes. This I was planning to 'work around' by adding a
WDC_CAPABILITY_NOCONTROL or something to tell the MI wdc code there
isn't an aux port available so it simply doesn't try to reset the
controller (which is not possible anyway).

> b) maybe to disable the controller interupts (I don't remember anymore)

Can't see any such use.

> c) to read the alternate status register, which I think is readable without
> clearing the status (or the pending interupt).

Neither can I see this around.

omoikane:/scratch/nb/idedoubler/src/sys/dev/ic$ grep ctl_iot wdc.c
        bus_space_write_1(chp->ctl_iot, chp->ctl_ioh, wd_aux_ctlr,
        bus_space_write_1(chp->ctl_iot, chp->ctl_ioh, wd_aux_ctlr,
        bus_space_write_1(chp->ctl_iot, chp->ctl_ioh, wd_aux_ctlr, WDCTL_4BIT);
        bus_space_write_1(chp->ctl_iot, chp->ctl_ioh, wd_aux_ctlr,
        bus_space_write_1(chp->ctl_iot, chp->ctl_ioh, wd_aux_ctlr,
        bus_space_write_1(chp->ctl_iot, chp->ctl_ioh, wd_aux_ctlr,

Reading these uses of aux port it seems we only use it to reset the
controller in wdcreset() and wdcprobe(), which gives me hope that the
idedoubler support might be doable without too much pain.

-- 
Ilpo Ruotsalainen - <lonewolf@iki.fi> - http://www.iki.fi/lonewolf/