26 Feb
2013
26 Feb
'13
5:21 p.m.
In freebsd there's a linux compatibility library/package at least for binaries but
I'd be surprised if there wasn't something similar for compilation. Also freebsd
(not openbsd but they're all a family so port is likely) I used ndis to use the
windows driver for a not supported at the time wireless card in bsd. Not sure if that
would be in line with other drivers as well.
26 Feb
26 Feb
10:15 p.m.
On 02/26/2013 05:21 PM, Sam O'nella wrote:
In freebsd there's a linux compatibility
library/package at least for
binaries but I'd be surprised if there wasn't something similar for
compilation. Also freebsd (not openbsd but they're all a family so
port is likely) I used ndis to use the windows driver for a not
supported at the time wireless card in bsd. Not sure if that would be
in line with other drivers as well.
I recall using the NDIS2 driver in the early days of the Linksys WMP54GS
PCI wireless card on Debian. I'm not sure that OpenBSD has any such thing.
However, some years ago, the Linksys USB200 adapter driver was ported
from FreeBSD to OpenBSD. Armed with that and the Linux
source to the
adapter I have now, I think I may be able to make some sense out of
things.
(Fingers crossed)
--Chuck
27 Feb
27 Feb
6:51 a.m.
On Feb 26, 2013, at 8:21 PM, "Sam O'nella" <barythrin at gmail.com>
wrote:
In freebsd there's a linux compatibility
library/package at least for binaries but I'd be surprised if there wasn't
something similar for compilation. Also freebsd (not openbsd but they're all a family
so port is likely) I used ndis to use the windows driver for a not supported at the time
wireless card in bsd. Not sure if that would be in line with other drivers as well.
That's an emulation for application binaries; it basically mimics
the syscall interface of Linux (or, with other modules, other BSDs
or older kernel versions where the API has changed). It's fine for
applications, but not so useful for kernel drivers, unfortunately.
Unless I'm thinking of something different than you are.
- Dave
11:15 p.m.
Well, a little poking around reveals that the HG20f9 USB-to-ethernet
gizmo, although it uses a Linksys VID (0x066b), isn't a Linksys product
at all.
Rather, it's a Chinese-origin device that uses the ASX88772B chip from
Asix. The 0x20f9 PID doesn't appear to conflict with any existing
Linksys device, so it appears that I'll be okay by simply adding the
PID/VID pair to the existing AXE driver for OpenBSD.
There's an interesting approach to the costly problem of purchasing a
VID license from USB-IF--just steal someone else's! Who's gonna know?
Now why didn't I think of that?
--CHuck
28 Feb
28 Feb
5:23 a.m.
On Feb 28, 2013, at 2:15 AM, Chuck Guzis <cclist at sydex.com> wrote:
Well, a little poking around reveals that the HG20f9
USB-to-ethernet gizmo, although it uses a Linksys VID (0x066b), isn't a Linksys
product at all.
Rather, it's a Chinese-origin device that uses the ASX88772B chip from Asix. The
0x20f9 PID doesn't appear to conflict with any existing Linksys device, so it appears
that I'll be okay by simply adding the PID/VID pair to the existing AXE driver for
OpenBSD.
Is the device not Linksys-badged? It's not terribly uncommon
for large companies (especially as large as Linksys has gotten)
to buy the design from a design house and slap their labels and
their own VID on. For example, I don't know how true it is at
this point, but at least a few years ago basically all Apple
and Microsoft USB mice were Logitech devices, including the
iMac's infamous "hockey puck" mouse (it said so on the PCB).
The VIDs were usually those of the company on the label.
I think the comp.os.bsd.tech post I linked had some useful info
regarding that for OpenBSD; in particular, there was a bugfix
the poster had to import from FreeBSD's driver to fix a PHY
problem. If you're finding communications problems, you might
want to check it (relinked in case the original got buried).
http://comments.gmane.org/gmane.os.openbsd.tech/21147
There's an interesting approach to the costly
problem of purchasing a VID license from USB-IF--just steal someone else's! Who's
gonna know?
Now why didn't I think of that?
That's a common problem with USB (and PCI, and anything else
that requires a centrally-assigned ID system). For example,
to make a USB device that pretends to be a serial port, you
need to either write drivers for everything under the sun
(especially Windows, though there are filter drivers that
make libusb work mostly like it does under Unix) or fake an
FTDI VID/PID and duplicate its command set so you don't have
to worry about distributing a driver with your thing. It
would be nice if there were more generic classes like Mass
Storage and HID, but they're sadly scarce.
- Dave
9:10 a.m.
On 02/28/2013 05:23 AM, David Riley wrote:
Is the device not Linksys-badged? It's not
terribly uncommon
for large companies (especially as large as Linksys has gotten)
to buy the design from a design house and slap their labels and
their own VID on. For example, I don't know how true it is at
this point, but at least a few years ago basically all Apple
and Microsoft USB mice were Logitech devices, including the
iMac's infamous "hockey puck" mouse (it said so on the PCB).
The VIDs were usually those of the company on the label.
No badging at all; just a stick-on barcode label (ZC55232132ZC). The
AXE OpenBSD port from FreeBSD has been around since 2004 and covers 27
different devices using the Asix chip family; from a variety of vendors,
including Apple, Linksys, Acer, Logitech, Netgear and Belkin.
The PID is not mentioned in the USB-IF database for Linksys.
All in all, it look to be a case of "by their chips shall ye know them,
not from their VID". I wonder if that's what USB-IF intended--I thought
it was more a matter of "here's a VID/PID pair that tells you what's
inside and what driver you need". Apparently not--it's just a fancy
barcode for the product.
At any rate, I'll add the VID/PID pair to the AXE driver and see what
happens.
I suppose this might be compared to a "Cheap Rolex Watch" type of thing,
except that the insides really *are* those of a Rolex watch.
A quick check of VID 0001 (Fry's Electronics!) apparently shows that the
practice of counterfeiting IDs is apparently widespread:
http://usb-ids.gowdy.us/read/UD/0001/7778
So why should anyone PAY for an ID from USB-IF? Many Chinese
manufacturers apparently don't.
--Chuck
4098
days inactive
4099
days old
5 comments
3 participants
participants (3)
-
barythrin＠gmail.com -
cclist＠sydex.com -
fraveydank＠gmail.com