netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* WiFi and changed MAC addrs
@ 2007-09-05 15:20 Ben Woodard
  2007-09-05 15:34 ` Johannes Berg
  0 siblings, 1 reply; 2+ messages in thread
From: Ben Woodard @ 2007-09-05 15:20 UTC (permalink / raw)
  To: netdev

I just spent basically a day of my weekend grappling with this problem 
http://wiki.xensource.com/xenwiki/XenWifi
before I understood it well enough to realize that the fact that Xen 
changed the MAC addr for my WiFi card was the root cause of the problem. 
This got me thinking:

If a particular piece of hardware will refuse to send frames after the 
MAC address has been changed, then shouldn't the card's driver refuse to 
allow a change of the MAC address and return an error in response to the 
IOCTL that tries to change it? I can't think of a situation where this 
wouldn't be the right thing to do. However, before I submit a patch to 
fix this, does anyone else know of any special cases that I might need 
to deal with?

Also is behavior of refusing to send frames with a non-hardware 
specified MAC address part of the WiFi spec or is it particular to 
certain WiFi cards? IOW at what level do I need to fix it? Should I just 
fix it for the Intel cards that I've tested or do I need to fix it for 
all WiFi cards?

-ben

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: WiFi and changed MAC addrs
  2007-09-05 15:20 WiFi and changed MAC addrs Ben Woodard
@ 2007-09-05 15:34 ` Johannes Berg
  0 siblings, 0 replies; 2+ messages in thread
From: Johannes Berg @ 2007-09-05 15:34 UTC (permalink / raw)
  To: Ben Woodard; +Cc: netdev

[-- Attachment #1: Type: text/plain, Size: 1254 bytes --]

On Wed, 2007-09-05 at 08:20 -0700, Ben Woodard wrote:

> If a particular piece of hardware will refuse to send frames after the 
> MAC address has been changed, then shouldn't the card's driver refuse to 
> allow a change of the MAC address and return an error in response to the 
> IOCTL that tries to change it?

Yes.

> I can't think of a situation where this 
> wouldn't be the right thing to do. However, before I submit a patch to 
> fix this, does anyone else know of any special cases that I might need 
> to deal with?

That wouldn't make sense now, would it?

> Also is behavior of refusing to send frames with a non-hardware 
> specified MAC address part of the WiFi spec or is it particular to 
> certain WiFi cards? IOW at what level do I need to fix it? Should I just 
> fix it for the Intel cards that I've tested or do I need to fix it for 
> all WiFi cards?

Most wifi cards don't care at all what MAC address you send, in fact
allowing you to send with various MAC addresses. So this is something
intel specific. Oh are we talking about iwlwifi? If so, please report to
their maintainers immediately, they're lazy bastards and don't even
upload the new MAC address to the card even if they could.

johannes

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 190 bytes --]

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2007-09-05 15:33 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-09-05 15:20 WiFi and changed MAC addrs Ben Woodard
2007-09-05 15:34 ` Johannes Berg

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).