When I tested 2.6.31-rc4-wl on my old i386 laptop that uses ifup to
control the network, wireless did not work. The problem was bisected
to the following:
commit f1b3c87b70287ded1338fb03a0a714d185041a96
Author: Johannes Berg <[email protected]>
Date: Tue Jul 7 03:56:05 2009 +0200
cfg80211: fix netdev down problem
We shouldn't be looking at the ssid_len for non-IBSS,
and for IBSS we should also return an error on trying
to leave an IBSS while not in or joining an IBSS.
This fixes an issue where we wouldn't disconnect() on
an interface being taken down since there's no SSID
configured this way.
The patch cannot be reverted as net/wireless.ibss.c has changed.
Reverting the first hunk that modifies net/wireless/core.c does not
make the system work.
Larry
Hi Larry,
On Sat, 2009-08-01 at 20:54 -0500, Larry Finger wrote:
> When I tested 2.6.31-rc4-wl on my old i386 laptop that uses ifup to
> control the network, wireless did not work. The problem was bisected
> to the following:
Thanks for taking the time to bisect the problem, can you be more
specific as to how the problem manifests itself? Are you using ifup to
configure an ad-hoc network?
Thanks,
johannes
Johannes Berg wrote:
> Hi Larry,
>
> On Sat, 2009-08-01 at 20:54 -0500, Larry Finger wrote:
>> When I tested 2.6.31-rc4-wl on my old i386 laptop that uses ifup to
>> control the network, wireless did not work. The problem was bisected
>> to the following:
>
> Thanks for taking the time to bisect the problem, can you be more
> specific as to how the problem manifests itself? Are you using ifup to
> configure an ad-hoc network?
No, the network is in managed made using WPA2 encryption. I'm using
ifup because that laptop is only used for testing, is quite slow, and
does not run X.
When I tried to test the latest version on my main laptop using ifup
instead of the usual NM setup, I found that I had to modify my
configuration file /etc/sysconfig/network/ifcfg-eth1. My udev rules
explicitly rename my Broadcom devices to eth1. When I transferred the
revised file to the other laptop, it also started working. With
testing, it turns out the critical addition was "WIRELESS='yes'"
I do not understand why a change in configuration was necessary at
that point; however, the "regression" is only that a more complete
configuration file is required. Please disregard this thread.
Larry