Re: ipw2100 wireless driver
From: James Ketrenos
Date: Wed Aug 11 2004 - 11:25:29 EST
Christoph Hellwig wrote:
On Wed, Aug 11, 2004 at 05:16:35AM -0500, James Ketrenos wrote:
We're currently working to clean up ipw2100 and ieee80211 code for submission to
netdev for discussion and hopefully inclusion in the future. The ieee80211 code
is still being heavily developed, but its usable. If anyone wants to help out,
or if folks feel its ready as-is to get pulled into wireless-2.6, let me know.
Maybe we should switch to your ieee802.11 for a generic wireless stack then
instead of the original hostap code. At least it seems more actively
maintained right now and supports two drivers already.
This would be ideal for those working on the projects using that stack. If
others agree I'll put together patches that introduce the ieee80211* module into
wireless-2.6 once I get the next driver snapshots out for ipw2100 and ipw2200.
Btw, I've looked at the ipw2100 and have to concerns regarding the firmware,
a) yo'ure not using the proper firmware loader but some horrible
handcrafted code using sys_open/sys_read & co that's not namespace
safe at all
The driver supports (and defaults to) using firmware_class for loading the
firmware. The driver also supports a legacy loading approach for folks that
have problems with using hotplug to load the firmware (which represents a fair
number of users).
b) the firmware has an extremly complicated and hard to comply with license,
I'm not sure we want a driver that can't work without a so strangely
licensed blob in the kernel. Can you talk to intel lawyers and put it on
simple redristribution and binary modification for allowed for all purposes
license please?
The firmware license supports redistribution, and complying with the license
shouldn't be too hard (I agree it may not be worded the most clearly, but few
legal documents are). If you have issues or questions about specific terms
please email me offlist and I can try and address them.
Just to re-answer some others may be wondering regarding the firmware:
1) the firmware does not use nor is it dependent (at all) on the kernel. no
part of the firmware executes on the host CPU.
2) the firmware is loaded from disk vs. having to have non-volatile storage on
the NIC and requiring a firmware flashing utility, etc.
3) the firmware, as per its license, can be redistributed by OSDs, ISVs, etc.
Thanks,
James
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/