[nycbug-talk] Any Ideas on wireless cards

Miles Nordin carton at Ivy.NET
Fri Jun 29 19:10:39 EDT 2007


>>>>> "ms" == Marc Spitzer <mspitzer at gmail.com> writes:

    ms> Any other ideas?

given ideals you've stated in the past, I would think you'd use
NDISAPI/ndisgen.  supposedly the binary drivers perform better under
crappy radio conditions or on busy bands like at conferences.

I got an old Atheros card made by NEC from a Yahoo! store for cheap.
It does work.  Old cards might be a good idea.

finally, FWIW, Theo raves about the Ralink cards on URL's that have
been posted here in the past such as:

 http://www.openbsd.org/papers/opencon06-drivers/mgp00024.html

because they are not really that crappy, have simple drivers with more
hope of being bug-free (laptops will probably be exploited through
their wireless card drivers relatively commonly in the future), and
they are cooperative about releasing documentation without NDA.  so
that's what I would get.

My understanding is, Atheros would suck less if we were able to take
release engineering of the hal away from Sam Leffler.  He pushes out
highly unstable versions and uses the community as his private beta
testers, and he has a vested interest in making sure Reyk's driver
stays out of FreeBSD because, since he's signed the NDA, he's crossed
over, and he can't work on Reyk's driver because he's signed an
agreement that he's tainted.  so if Reyk's HAL is imported, Sam will
be excluded from wireless work.  I find the whole thing disgusting,
mostly since OpenBSD is the only group seeing the big issue and doing
something about it---everyone else, Linux NetBSD FreeBSD Solaris, took
the bait.

so, for you Marc, use the NDIS driver.  And let us know which Windows
installer .EXE gives you files that work well under FreeBSD.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 304 bytes
Desc: not available
URL: <http://lists.nycbug.org/pipermail/talk/attachments/20070629/dec1c63e/attachment.bin>


More information about the talk mailing list