Hey all!
Thanks to those who wrote wanting to help me test the usb client/server
code. Here's a quick update:
Originally I didn't have all the files, so I had to hack together some
header files, and then change some code based on missing things I couldn't
reverse engineer. Recently, I got ahold of those header files, and I had
to back up and start again, as they contained structures that were modified.
So, you can pretty much compile what I've got now by getting the files off
the cvs repository at handhelds.org: /linux/kernel/drivers/usb. There are
two Makefiles that go with the package, since the client driver needs to be
cross compiled for the iPaq (using Makefile.2: make -f Makefile.2) in one
kernel tree (a working iPaq compilable tree), and the host driver needs to
be compiled using Makefile.1 (like in Makefile.2) in a working i386 kernel
tree. BEWARE - the Makefile for the host zaps all existing .o files in the
usb driver directory - you may have to recompile those modules again if
they aren't installed into your /lib/modules dir....
The client is made up of the files usbc_3600_*.c, where * = ctl, ep0, dma,
eth, list, recv, and send.
The host (what there is so far) is all in the file usbh_h3600.c/h.
The client seems to work and report all the debugging info, but the host
faults when it tries to set up the ether driver name (nusb#m # = 0, 1,
2...). I'm debugging through that to get the test code to work. It's a
kernel NULL pointer dereference error, and I'm pretty sure it's because
there's unallocated space for structures or a garbled pointer construct.
As far as I can tell, the code currently tries to stream usb bulk data back
and forth a few times. Once this is working, I'll attack hooking the usb
data streams up to the network layer...
Since I can't attach the Makefiles in mail to the list, I'll put them on a
skiff somewhere - let me know if you want it copied to a directory that
you're using on a particular skiff, and I'll copy it there for any and all
who are interested...
Have at it....
-G!
PS It's all adapted from the plusb network driver. Maybe there's another
newer driver we could examine? I'm currently learning from the existing
usb code....
Received on Mon Sep 25 17:17:58 2000
This archive was generated by hypermail 2.1.8 : Tue May 04 2004 - 09:43:42 EDT