From e2e414d92397c366396d13f627a98a20be92e509 Mon Sep 17 00:00:00 2001 From: Johannes Berg <johannes@sipsolutions.net> Date: Fri, 10 Jul 2009 11:38:14 +0200 Subject: mac80211: disable mesh My kvm instance was complaining a lot about sleeping in atomic contexts in the mesh code, and it turns out that both mesh_path_add() and mpp_path_add() need to be able to sleep (they even use synchronize_rcu()!). I put in a might_sleep() to annotate that, but I see no way, at least right now, of actually making sure those functions are only called from process context since they are both called during TX and RX and the mesh code itself even calls them with rcu_read_lock() "held". Therefore, let's disable it completely for now. It's possible that I'm only seeing this because the hwsim's beaconing is broken and thus the peers aren't discovered right away, but it is possible that this happens even if beaconing is working, for a peer that doesn't exist or so. It should be possible to solve this by deferring the freeing of the tables to call_rcu() instead of using synchronize_rcu(), and also using atomic allocations, but maybe it makes more sense to rework the code to not call these from atomic contexts and defer more of the work to the workqueue. Right now, I can't work on either of those solutions though. Signed-off-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: John W. Linville <linville@tuxdriver.com> --- net/mac80211/mesh_pathtbl.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'net/mac80211/mesh_pathtbl.c') diff --git a/net/mac80211/mesh_pathtbl.c b/net/mac80211/mesh_pathtbl.c index 3c72557df45..02f8709a181 100644 --- a/net/mac80211/mesh_pathtbl.c +++ b/net/mac80211/mesh_pathtbl.c @@ -175,6 +175,8 @@ int mesh_path_add(u8 *dst, struct ieee80211_sub_if_data *sdata) int err = 0; u32 hash_idx; + might_sleep(); + if (memcmp(dst, sdata->dev->dev_addr, ETH_ALEN) == 0) /* never add ourselves as neighbours */ return -ENOTSUPP; @@ -265,6 +267,7 @@ int mpp_path_add(u8 *dst, u8 *mpp, struct ieee80211_sub_if_data *sdata) int err = 0; u32 hash_idx; + might_sleep(); if (memcmp(dst, sdata->dev->dev_addr, ETH_ALEN) == 0) /* never add ourselves as neighbours */ -- cgit v1.2.3-70-g09d2 From 35946a571099a50d2595c8866f07617d29558f53 Mon Sep 17 00:00:00 2001 From: Javier Cardona <javier@cozybit.com> Date: Mon, 13 Jul 2009 17:00:10 -0700 Subject: mac80211: use correct address for mesh Path Error For forwarded frames, we save the precursor address in addr1 in case it needs to be used to send a Path Error. mesh_path_discard_frame, however, was using addr2 instead of addr1 to send Path Error frames, so correct that and also make the comment regarding this more clear. Signed-off-by: Andrey Yurovsky <andrey@cozybit.com> Signed-off-by: John W. Linville <linville@tuxdriver.com> --- net/mac80211/mesh_pathtbl.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'net/mac80211/mesh_pathtbl.c') diff --git a/net/mac80211/mesh_pathtbl.c b/net/mac80211/mesh_pathtbl.c index 02f8709a181..479597e8858 100644 --- a/net/mac80211/mesh_pathtbl.c +++ b/net/mac80211/mesh_pathtbl.c @@ -494,8 +494,10 @@ void mesh_path_tx_pending(struct mesh_path *mpath) * @skb: frame to discard * @sdata: network subif the frame was to be sent through * - * If the frame was beign forwarded from another MP, a PERR frame will be sent - * to the precursor. + * If the frame was being forwarded from another MP, a PERR frame will be sent + * to the precursor. The precursor's address (i.e. the previous hop) was saved + * in addr1 of the frame-to-be-forwarded, and would only be overwritten once + * the destination is successfully resolved. * * Locking: the function must me called within a rcu_read_lock region */ @@ -510,7 +512,7 @@ void mesh_path_discard_frame(struct sk_buff *skb, u8 *ra, *da; da = hdr->addr3; - ra = hdr->addr2; + ra = hdr->addr1; mpath = mesh_path_lookup(da, sdata); if (mpath) dsn = ++mpath->dsn; -- cgit v1.2.3-70-g09d2