usb/119002: add "ucp" driver support

[Steve Franks]

The following reply was made to PR usb/119002; it has been noted by GNATS.

From: "Steve Franks" <stevefranks at ieee.org>
To: bug-followup at FreeBSD.org, stevefranks at ieee.org
Cc:  
Subject: Re: usb/119002: add "ucp" driver support
Date: Tue, 25 Dec 2007 16:41:10 -0700

 Sucess!  I have built and proven that the ucp.ko driver works with all
 my various work and hobby projects using the Silicon Laboratories
 CP2101, CP2102, and CP2103 devices.  There are alot of other
 off-the-shelf OEM devices using this chipset, which allow for
 different VID's & PID's (see the patch file) - I was obviously unable
 to test them as I only have the manufacturer's reference design style
 hardware, but there is no way to modify the operation of the chipset
 except for the VID & PID, so I assume they will work fine if FreeBSD
 recognizes the VID & PID - not a usb expert.  Test was performed with
 Minicom talking to /dev/cuaU0; after kldload ucp, plugging in a CP210x
 yields "ucom0" in dmesg, and cuaU0 in /dev.
 
 I think my earlier problem was due to the patch mucking the system -
 if run from /usr/src/sys, it doesn't create the modules/ucp folder
 like it should, don't know why.  Here's what I did that worked:
 
 1. /usr/src/sys/dev/usb# patch < ~/ucp-0.01.diff
 2. /usr/src/sys/dev/usb# mkdir /usr/src/sys/modules/ucp
 3. /usr/src/sys/dev/usb# mv Makefile /usr/src/sys/modules/ucp
 3a. Once you have proven it works, you might want to remove the -DEBUG
 line from this makefile (I did up front, never tested with debug
 symbols/code on my system)
 4. manually edit /usr/src/sys/dev/usb/usbdevs to fix the patch
 collision and add all the new lines (specifically vendor SILABS 0x10c4
 Silicon labs)
 5. manually edit /usr/src/sys/modules/Makefile to add the line:
             ucp \
     after the line
             ucom \
 6. build & install kernel
 
 Incidentally, since I cvsup'd to clean up my previous mess, I am now
 on 6.3-xxx, but I don't think that was a factor in making it work.
 
 Also, note that step 5 is not in the diff file at all for some reason,
 as opposed to step 4 which simply didn't patch in automatically.
 
 Best,
 
 Steve