On Saturday 27 October 2007 08:58:55 Johannes Berg wrote:

Oops, I stand corrected. Still I don't think that the message patch is
worth doing in stable, although the regression potential is pretty much
nil.

According to the stable kernel rules, this fix doesn't look appropriate. 
(though it is practically zero risk) Also, the sta_bss_list changes do not 
qualify due to "It cannot be bigger than 100 lines, with context". The 
radiotap get_unaligned patch doesn't qualify - it's not important enough. The 
IE parsing fix is also too big. I'm not certain that the SSID scanning patch 
qualifies either.. from the patch's POV, it's adding a feature. So, the way 
I'm reading things, no mac80211 patches seem to qualify according to stable 
kernel rules.

FWIW, I think 
[PATCH] mac80211: reorder association debug output
[PATCH] mac80211: honor IW_SCAN_THIS_ESSID in siwscan ioctl (plus fix)
[PATCH] mac80211: make ieee802_11_parse_elems return void
do enough good that they should be pushed to stable. The IE parsing fix is a 
little big but it's trivial, reordering the debug output is safe and reduces 
confusion, and the SSID scanning patch makes it easier to connect to hidden 
APs. In addition, the SSID len match patch should also be pushed to stable 
since I think it's a one line, fairly obvious patch that fixes a problem in 
auto AP selection that can prevent association.

-Michael Wu