No response from original post, and it missed getting taken into -rc7
and -rc8, so repost:
Release acquired lock on error path, fixing potential hang up.
Signed-off-by: Daniel J Blueman <[email protected]>
diff --git a/net/bluetooth/mgmt.c b/net/bluetooth/mgmt.c
index f827fd9..ace8726 100644
--- a/net/bluetooth/mgmt.c
+++ b/net/bluetooth/mgmt.c
@@ -111,8 +111,10 @@ static int read_index_list(struct sock *sk)
body_len = sizeof(*ev) + sizeof(*rp) + (2 * count);
skb = alloc_skb(sizeof(*hdr) + body_len, GFP_ATOMIC);
- if (!skb)
+ if (!skb) {
+ read_unlock(&hci_dev_list_lock);
return -ENOMEM;
+ }
hdr = (void *) skb_put(skb, sizeof(*hdr));
hdr->opcode = cpu_to_le16(MGMT_EV_CMD_COMPLETE);--
--
Daniel J Blueman
Hi Daniel,
* Daniel J Blueman <[email protected]> [2011-03-09 00:07:54 +0800]:
> No response from original post, and it missed getting taken into -rc7
> and -rc8, so repost:
>
> Release acquired lock on error path, fixing potential hang up.
I already have a patch like this one in my tree, thanks anyway. :)
--
Gustavo F. Padovan
http://profusion.mobi
Hi Gustavo,
On 12 March 2011 09:47, Gustavo F. Padovan <[email protected]> wrote:
> Hi Daniel,
>
> * Daniel J Blueman <[email protected]> [2011-03-09 00:07:54 +0800]:
>
>> No response from original post, and it missed getting taken into -rc7
>> and -rc8, so repost:
>>
>> Release acquired lock on error path, fixing potential hang up.
>
> I already have a patch like this one in my tree, thanks anyway. :)
I guess I'm not looking in the right public tree, but I still couldn't
detect the fix in:
http://git.kernel.org/?p=linux/kernel/git/padovan/bluetooth-2.6.git
or:
http://git.kernel.org/?p=linux/kernel/git/padovan/bluetooth-next-2.6.git
Thanks,
Daniel
--
Daniel J Blueman
Hi Daniel,
* Daniel J Blueman <[email protected]> [2011-03-13 11:59:27 +0800]:
> Hi Gustavo,
>
> On 12 March 2011 09:47, Gustavo F. Padovan <[email protected]> wrote:
> > Hi Daniel,
> >
> > * Daniel J Blueman <[email protected]> [2011-03-09 00:07:54 +0800]:
> >
> >> No response from original post, and it missed getting taken into -rc7
> >> and -rc8, so repost:
> >>
> >> Release acquired lock on error path, fixing potential hang up.
> >
> > I already have a patch like this one in my tree, thanks anyway. :)
>
> I guess I'm not looking in the right public tree, but I still couldn't
> detect the fix in:
> http://git.kernel.org/?p=linux/kernel/git/padovan/bluetooth-2.6.git
> or:
> http://git.kernel.org/?p=linux/kernel/git/padovan/bluetooth-next-2.6.git
On bluetooth-next:
commit b2c60d42db0fea1e6c4345739601024863566a13
Author: Jesper Juhl <[email protected]>
Date: Fri Jan 14 00:18:49 2011 +0100
Bluetooth: Fix failure to release lock in read_index_list()
If alloc_skb() fails in read_index_list() we'll return -ENOMEM without
releasing 'hci_dev_list_lock'.
Signed-off-by: Jesper Juhl <[email protected]>
Signed-off-by: Gustavo F. Padovan <[email protected]>
--
Gustavo F. Padovan
http://profusion.mobi