From: Karl Relton <karllinuxtest.relton@ntlworld.com>
To: linux-bluetooth@vger.kernel.org
Subject: [PATCH hidp] Fix Kernel OOPS in hidp_session caused by orphaned sock structures
Date: Tue, 08 Jan 2013 14:21:57 +0000 [thread overview]
Message-ID: <1357654917.5641.7.camel@dellpc> (raw)
This is a proposed fix to the kernel OOPS reported in
https://bugzilla.kernel.org/show_bug.cgi?id=50541
hidp_session() is crashing the kernel typically on a resume from
suspend. My analysis concluded that the sock structure pointed to by
ctrl_sk was being orphaned at some point parallel to the execution of
hidp_session (sometimes even while hidp_session is still in its main
loop). This mean that calls to sk_sleep(ctrl_sk) would return NULL,
leading to ..._wait() or ..._wakeup() calls crashing.
The proposed fix is to store the waitqueue_head structure needed for the
waiting/waking in a local variable. rcu_dereference_raw() [normally
called via sk_sleep()] is still used and required to protect access, but
we ensure we have the proper handle onto the structure rather than
losing it by the sock being orphaned.
Signed-off-by: Karl Relton <karllinuxtest.relton@ntlworld.com>
--- linux-3.7.0.orig/net/bluetooth/hidp/core.c 2013-01-08 13:04:35.945237334 +0000
+++ linux-3.7.0/net/bluetooth/hidp/core.c 2013-01-08 13:06:11.313240959 +0000
@@ -680,16 +680,19 @@ static int hidp_session(void *arg)
struct sock *intr_sk = session->intr_sock->sk;
struct sk_buff *skb;
wait_queue_t ctrl_wait, intr_wait;
+ struct socket_wq *ctrl_wq, *intr_wq;
BT_DBG("session %p", session);
__module_get(THIS_MODULE);
set_user_nice(current, -15);
+ ctrl_wq = ctrl_sk->sk_wq;
+ intr_wq = intr_sk->sk_wq;
init_waitqueue_entry(&ctrl_wait, current);
init_waitqueue_entry(&intr_wait, current);
- add_wait_queue(sk_sleep(ctrl_sk), &ctrl_wait);
- add_wait_queue(sk_sleep(intr_sk), &intr_wait);
+ add_wait_queue(&rcu_dereference_raw(ctrl_wq)->wait, &ctrl_wait);
+ add_wait_queue(&rcu_dereference_raw(intr_wq)->wait, &intr_wait);
session->waiting_for_startup = 0;
wake_up_interruptible(&session->startup_queue);
set_current_state(TASK_INTERRUPTIBLE);
@@ -722,8 +725,8 @@ static int hidp_session(void *arg)
set_current_state(TASK_INTERRUPTIBLE);
}
set_current_state(TASK_RUNNING);
- remove_wait_queue(sk_sleep(intr_sk), &intr_wait);
- remove_wait_queue(sk_sleep(ctrl_sk), &ctrl_wait);
+ remove_wait_queue(&rcu_dereference_raw(intr_wq)->wait, &intr_wait);
+ remove_wait_queue(&rcu_dereference_raw(ctrl_wq)->wait, &ctrl_wait);
clear_bit(HIDP_WAITING_FOR_SEND_ACK, &session->flags);
clear_bit(HIDP_WAITING_FOR_RETURN, &session->flags);
@@ -747,12 +750,15 @@ static int hidp_session(void *arg)
session->intr_sock->sk->sk_err = EUNATCH;
session->ctrl_sock->sk->sk_err = EUNATCH;
- hidp_schedule(session);
+ wake_up_interruptible(&rcu_dereference_raw(ctrl_wq)->wait);
+ wake_up_interruptible(&rcu_dereference_raw(intr_wq)->wait);
fput(session->intr_sock->file);
- wait_event_timeout(*(sk_sleep(ctrl_sk)),
- (ctrl_sk->sk_state == BT_CLOSED), msecs_to_jiffies(500));
+ /* By now ctrl_sk might have been orphaned already */
+ if (ctrl_wq == ctrl_sk->sk_wq)
+ wait_event_timeout(rcu_dereference_raw(ctrl_wq)->wait,
+ (ctrl_sk->sk_state == BT_CLOSED), msecs_to_jiffies(500));
fput(session->ctrl_sock->file);
reply other threads:[~2013-01-08 14:21 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
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=1357654917.5641.7.camel@dellpc \
--to=karllinuxtest.relton@ntlworld.com \
--cc=linux-bluetooth@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;
as well as URLs for NNTP newsgroup(s).