On Mon, Mar 09, 2009 at 10:48:52PM -0400, Luis R. Rodriguez wrote: > There is something awefully wrong with this patch and after some > good review I cannot determine what is wrong with it. The issue is > I end up following bug complaint saying I'm sleeping while I'm atomic.. > but this is all from a workqueue so what gives? You hold a spinlock while trying to alloc memory with GFP_KERNEL.. > [14462.644641] [<ffffffff804a40c4>] __alloc_skb+0x44/0x140 > [14462.644652] [<ffffffffa051a8b3>] nl80211_send_beacon_hint_event+0x33/0x370 [cfg80211] > [14462.644660] [<ffffffff8026960e>] ? put_lock_stats+0xe/0x30 > [14462.644671] [<ffffffffa0511f19>] handle_reg_beacon+0xa9/0xc0 [cfg80211] > [14462.644685] [<ffffffffa0512500>] ? reg_todo+0x0/0x5c0 [cfg80211] > [14462.644696] [<ffffffff80541a1d>] ? _spin_lock_bh+0x6d/0x80 That would be spin_lock_bh() from reg_process_pending_beacon_hints(), I would assume.. nl80211_send_beacon_hint_event() cannot use GFP_KERNEL on such a call path. -- Jouni Malinen PGP id EFC895FA -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
