| |||||||||||
The weak and strong spot of Linux - Wireless networkingThe weak and strong spot of Linux - Wireless networkingPosted Mar 3, 2004 17:08 UTC (Wed) by nchip (guest, #13292)Parent article: Linux wireless networking (developerWorks)
At the same time Linux is making a massive success in wireless access
(Log in to post comments)
The weak and strong spot of Linux - Wireless networking Posted Mar 3, 2004 17:43 UTC (Wed) by proski (subscriber, #104) [Link] I actually want to ask developers to join forces and create a mailing list for Linux wireless developers. There is a lot of work done in parallel. Wireless Extensions are essentially developed by one person. There is no standard for encapsulating 802.11 frames for non-wireless medium. There is no extensible, hardware-independent, documented standard for passing addition data with sniffed frames to the userspace. There is a lot of experience that the developers of drivers for old chipsets (e.g. Prism) could share with the developers of the new drivers.I'm glad that I'm not the only person who thinks about it.
The weak and strong spot of Linux - Wireless networking Posted Mar 3, 2004 19:19 UTC (Wed) by millette (guest, #8585) [Link] There's a list to discuss openwrt available, although it hasn't really served much yet...
on mailing lists and capture headers Posted Mar 3, 2004 21:16 UTC (Wed) by pizza (subscriber, #46) [Link] There are of course mailing lists based around nearly all of the various drivers themselves, but no overarching list. And of course, the majority of the questions have little to do with the actual development of the drivers.I'd love to see a general-purpose wireless driver developer list. I could probably even host it if necessary (at lists.linux-wlan.com or at shaftnet.org) Unfortunately it could be said that I have a conflict of interests -- I write Linux wireless drivers for a living, but not under F/OSS terms. But beyond that, by participating I may be putting myself into legal hot water thanks to NDAs, which are Evil Evil Evil. Thanks to them I can't even participate in a Free Software project if it remotely involves something covered. On the other hand, I've already run into and had to deal with quite a few problems relating to a shared 802.11 infrastructure/driver suite that currently handles half a dozen chipsets.. and thus have dealt with many of the mentioned problems already. However, the only thing I've spat back out publically was a proposal for a new capture frame header format, which I received almost no feedback on. Incidentally, Ethereal's supported it for some time now. It's worth mentioning that the old "prism header" was technically hardware independent and extremely extensible. But nobody else actually implemented it properly, so it ended up being a giant monolitic header of mostly wasted space. (Besides, what do you mean by "extensible"? Arbitrary opaque data?) But unfortunately real progress is hampered by lack of coopration from chipset companies, either in the form of source or documentation -- or at least non-draconian NDAs.
on mailing lists and capture headers Posted Mar 3, 2004 21:55 UTC (Wed) by proski (subscriber, #104) [Link] I'd love to see a general-purpose wireless driver developer list. I could probably even host it if necessary (at lists.linux-wlan.com or at shaftnet.org)I'd like to see linux-wireless at vger.kernel.org It's worth mentioning that the old "prism header" was technically hardware independent and extremely extensible.Yes. But libpcap thinks that it has fixed length. It's a human communication failure.
|
|||||||||||