From: Thomas Pedersen <thomas@cozybit.com>
To: johannes@sipsolutions.net
Cc: devel@lists.open80211s.org, linux-wireless@vger.kernel.org,
Thomas Pedersen <thomas@cozybit.com>
Subject: [PATCH] mac80211: don't allocate mesh peer under RCU
Date: Tue, 23 Oct 2012 11:55:07 -0700 [thread overview]
Message-ID: <1351018508-6390-1-git-send-email-thomas@cozybit.com> (raw)
mesh_plink_alloc() was being called with the RCU read lock held, which
triggered warnings because various things inside this function must
sleep. It doesn't make much sense to hold an RCU lock on a pointer we
don't yet have anyway, so do rcu_read_unlock() if allocating and relock
on insertion. This also means there is no need for
cfg80211_notify_new_peer_candidate() to be atomic.
Also balance the RCU lock if we are returning early.
This warning was only visible with CONFIG_DEBUG_ATOMIC_SLEEP enabled.
Reported-by: Bob Copeland <me@bobcopeland.com>
Signed-off-by: Thomas Pedersen <thomas@cozybit.com>
---
net/mac80211/mesh_plink.c | 14 +++++++++-----
1 file changed, 9 insertions(+), 5 deletions(-)
diff --git a/net/mac80211/mesh_plink.c b/net/mac80211/mesh_plink.c
index 234fe75..427c87d 100644
--- a/net/mac80211/mesh_plink.c
+++ b/net/mac80211/mesh_plink.c
@@ -346,7 +346,7 @@ static struct sta_info *mesh_peer_init(struct ieee80211_sub_if_data *sdata,
enum ieee80211_band band = ieee80211_get_sdata_band(sdata);
struct ieee80211_supported_band *sband;
u32 rates, basic_rates = 0;
- struct sta_info *sta;
+ struct sta_info *sta = NULL;
bool insert = false;
sband = local->hw.wiphy->bands[band];
@@ -354,18 +354,19 @@ static struct sta_info *mesh_peer_init(struct ieee80211_sub_if_data *sdata,
sta = sta_info_get(sdata, addr);
if (!sta) {
+ rcu_read_unlock();
/* Userspace handles peer allocation when security is enabled */
if (sdata->u.mesh.security & IEEE80211_MESH_SEC_AUTHED) {
cfg80211_notify_new_peer_candidate(sdata->dev, addr,
elems->ie_start,
elems->total_len,
- GFP_ATOMIC);
- return NULL;
+ GFP_KERNEL);
+ goto lock_out;
}
sta = mesh_plink_alloc(sdata, addr);
if (!sta)
- return NULL;
+ goto lock_out;
insert = true;
}
@@ -397,10 +398,13 @@ static struct sta_info *mesh_peer_init(struct ieee80211_sub_if_data *sdata,
rate_control_rate_init(sta);
spin_unlock_bh(&sta->lock);
- if (insert && sta_info_insert(sta))
+ if (insert && sta_info_insert_rcu(sta))
return NULL;
return sta;
+lock_out:
+ rcu_read_lock();
+ return sta;
}
void mesh_neighbour_update(struct ieee80211_sub_if_data *sdata,
--
1.7.9.5
next reply other threads:[~2012-10-23 18:55 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-10-23 18:55 Thomas Pedersen [this message]
2012-10-23 19:33 ` [PATCH] mac80211: don't allocate mesh peer under RCU Javier Cardona
2012-10-23 19:54 ` Johannes Berg
2012-10-23 20:22 ` Thomas Pedersen
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1351018508-6390-1-git-send-email-thomas@cozybit.com \
--to=thomas@cozybit.com \
--cc=devel@lists.open80211s.org \
--cc=johannes@sipsolutions.net \
--cc=linux-wireless@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox