On Mon, 2009-08-17 at 11:12 -0400, Bob Copeland wrote:
> Haven't looked at it yet, I just hit this one while riding the
> train this morning and connecting to someone's adhoc station.
>
> [43559.227490] wlan0: Selected IBSS BSSID 2a:51:0a:35:5c:a1 based on configured SSID
> [43559.258660] ath5k phy0: bf=efb0b680 bf_skb=(null)
> [43559.291501] phy0: Adding new IBSS station 00:21:e9:94:5e:93 (dev=wlan0)
> [43559.291518] phy0: Allocated STA 00:21:e9:94:5e:93
> [43559.292327] phy0: Inserted STA 00:21:e9:94:5e:93
> [43633.646820] phy0: Adding new IBSS station 00:1e:c2:3b:f4:e3 (dev=wlan0)
> [43633.646838] phy0: Allocated STA 00:1e:c2:3b:f4:e3
> [43633.646859]
> [43633.646860] =========================================================
> [43633.646864] [ INFO: possible irq lock inversion dependency detected ]
> [43633.646867] 2.6.31-rc6 #179
> [43633.646870] ---------------------------------------------------------
> [43633.646873] cc1/18756 just changed the state of lock:
> [43633.646876] (&local->sta_lock){..-...}, at: [<fa244acd>] sta_info_insert+0x91/0x174 [mac80211]
> [43633.646900] but this lock took another, SOFTIRQ-unsafe lock in the past:
> [43633.646903] (todo_lock){+.+...}
> [43633.646906]
> [43633.646907] and interrupts could create inverse lock ordering between them.
Yeah, commit 6565a1c6d2e0754b5f1e8160252f6dc9493ed6c4 needs to be
backported. I suspect it applies as-is, John mind trying that? If not,
let me know and I'll port it.
johannes
Attachment:
signature.asc
Description: This is a digitally signed message part
