>Number:         32362
>Category:       port-evbmips
>Synopsis:       Alchemy aucom serial ports do not have /dev/ node
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    port-evbmips-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Dec 22 19:05:00 +0000 2005
>Originator:     Garrett D'Amore
>Release:        3.99.14
>Organization:
Tadpole Computer, Inc.
>Environment:
NetBSD cabernet 3.99.14 NetBSD 3.99.14 (PB1000) #0: Wed Dec 21 10:35:17 PST 2005  garrett@zippy:/home/garrett/work/meteor/netbsd/src/sys/arch/evbmips/compile/obj/PB1000 evbmips

>Description:
According to the platform majors.evbmips, aucom is assigned major number 74.  However, when trying to identify a suitable device name in /dev, I can't find one:

cabernet# ls -la /dev/ | grep 73
crw-------   1 uucp  wheel        0, 15373 Nov  1 19:52 dtyCY013
crw-------   1 uucp  wheel        0, 15673 Nov  1 19:52 dtyCZ0057
brw-r-----   1 root  operator    16,    73 Nov  1 19:52 raid4j
crw-r-----   1 root  operator    43,    73 Nov  1 19:52 rraid4j
crw-r-----   1 root  operator     8,    73 Nov  1 19:52 rsd4j
brw-r-----   1 root  operator     8,    73 Nov  1 19:52 sd4j
cabernet#

>How-To-Repeat:
See description.
>Fix:
I think MAKEDEV.conf for evbmips needs something like this in it (which will create devices named ttyA[0-3]

aucom*)
    unit=${i#aucom}
    mkdev ttyA$unit c %aucom_chr% $(($unit + $dialin  )) "" "" $u_uucp
    mkdev dtyA$unit c %aucom_chr% $(($unit + $dialout )) "" "" $u_uucp
    ;;

And then in the md_all) code it should have this added to md_all:

    makedev aucom0 aucom1 aucom2 aucom3

(I am not aware of any alchemy parts with more than 4 serial ports.)

Of course, the makedev manual should probably also be updated.

Ultimately, in the long run, the aucom driver should probably go away and be replaced with a bus-specific attachment for com, but this requires adding some changes to com to cope with a couple of small register differences.