hi,

> You could simply add your device ID to the list of recognised boards
> in cy_match_pci() and see what happens.
ah, ok! *that's* the place to add the device-id. so here we go:

        case PCI_PRODUCT_CYCLADES_CYCLOMY_1:
                break;
        case PCI_PRODUCT_CYCLADES_CYCLOMY_2:
                break;
        case 0x0105:
                break;

compiling a new kernel... booting....and voila!
the board is recognized and seems to do something (I added option
CY_DEBUG, which gives the ugly reading below...and a lot of type cast
errors during compilation. I'll try to fix these tonight and send a
PR...):

Cyclades product 0x0105 (miscellaneous communications, revision 0x01) at ? dev 20 function 0 (tag 0x8000a000, intrtag 0x8000a000, intrswiz 0, intrpin 0x1, i/o on, mem on, no quirks)cy: card reset done
cy0 probe chip 0 offset 0x0 ... firmware version 0x48
cy0 probe chip 1 offset 0x800 ... firmware version 0x48
cy0 probe chip 2 offset 0x1000 ... not ready for command
found 2 CD1400s
: interrupting at irq 10
cy0attach CD1400 #0 offset 0x0
attach CD1400 #1 offset 0x800
: 8 ports

hm, the major number for cy devices is 38 (from /sys/arch/i386/i386/conf.c)
but there's no support in /dev/MAKEDEV... so hack MAKEDEV by adding:

tcy*)
        unit=${i#???}
        rm -f tcy$unit dcy$unit
        mknod tcy$unit c 38 $(($unit + $dialin ))
        mknod dcy$unit c 38 $(($unit + $dialout ))
        chown uucp.wheel tcy$unit dcy$unit
        ;;

(better names welcome ;-)

adding a line to /etc/remote using /dev/dcy0 and tip connects.
There's verbose debugging output on the console, so I think it's
basically working. I'll do some testing toninght (after building or
finding a usable cable... ;-)

thanks for the help!

       -kurt