linux-bluetooth.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Gustavo Padovan <padovan@profusion.mobi>
To: Ulisses Furquim <ulisses@profusion.mobi>
Cc: Emeltchenko Andrei <Andrei.Emeltchenko.news@gmail.com>,
	linux-bluetooth@vger.kernel.org
Subject: Re: [RFCv6 02/14] Bluetooth: Revert to mutexes from RCU list
Date: Mon, 20 Feb 2012 20:46:43 -0200	[thread overview]
Message-ID: <20120220224643.GA866@joana> (raw)
In-Reply-To: <CAA37ikZSwSLhDhs69UdDB38RBu3s2F8XmWAUB1vwS7A95xDBtw@mail.gmail.com>

* Ulisses Furquim <ulisses@profusion.mobi> [2012-02-20 15:57:52 -0300]:

> Hi Gustavo,
> 
> On Mon, Feb 20, 2012 at 3:33 PM, Gustavo Padovan <padovan@profusion.mobi> wrote:
> >
> > Hi Andrei,
> >
> > * Emeltchenko Andrei <Andrei.Emeltchenko.news@gmail.com> [2012-02-20 16:21:13 +0200]:
> >
> > > From: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
> > >
> > > Usage of RCU list looks not reasonalbe for a number of reasons:
> > > our code sleep and we have to use socket spinlocks, some parts
> > > of code are updaters thus we need to use mutexes anyway.
> > >
> > > Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
> > > ---
> > >  net/bluetooth/l2cap_core.c |  165 ++++++++++++++++++++++++++------------------
> > >  net/bluetooth/l2cap_sock.c |   10 +++
> > >  2 files changed, 108 insertions(+), 67 deletions(-)add an extra blank line here please.
> 
> <snip>
> 
> > > +unlock:
> > > +     mutex_unlock(&conn->chan_lock);
> > > +
> > > +     return err;
> > >  }
> > >
> > >  static inline void set_default_fcs(struct l2cap_chan *chan)
> > > @@ -2793,6 +2804,7 @@ static inline int l2cap_config_req(struct l2cap_conn *conn, struct l2cap_cmd_hdr
> > >               return -ENOENT;
> > >
> > >       sk = chan->sk;
> > > +     lock_sock(sk);
> > >
> > >       if (chan->state != BT_CONFIG && chan->state != BT_CONNECT2) {
> > >               struct l2cap_cmd_rej_cid rej;
> > > @@ -2905,6 +2917,7 @@ static inline int l2cap_config_rsp(struct l2cap_conn *conn, struct l2cap_cmd_hdr
> > >               return 0;
> > >
> > >       sk = chan->sk;
> > > +     lock_sock(sk);
> > >
> > >       switch (result) {
> > >       case L2CAP_CONF_SUCCESS:
> > > @@ -3006,11 +3019,16 @@ static inline int l2cap_disconnect_req(struct l2cap_conn *conn, struct l2cap_cmd
> > >
> > >       BT_DBG("scid 0x%4.4x dcid 0x%4.4x", scid, dcid);
> > >
> > > -     chan = l2cap_get_chan_by_scid(conn, dcid);
> > > -     if (!chan)
> > > +     mutex_lock(&conn->chan_lock);
> > > +
> > > +     chan = __l2cap_get_chan_by_scid(conn, dcid);
> > > +     if (!chan) {
> > > +             mutex_unlock(&conn->chan_lock);
> > >               return 0;
> >
> > Seems to me we can use l2cap_get_chan_by_scid() here.
> 
> It's better if we don't as we'll call l2cap_chan_del() which might
> remove chan from conn->chan_l that's protected by conn->chan_lock.
> Let's leave it as he sent.

Yes, I didn't realized the l2cap_chan_del() part. So after fix the new line
fix:

Acked-by: Gustavo F. Padovan <padovan@profusion.mobi>

	Gustavo

  reply	other threads:[~2012-02-20 22:46 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-02-20 14:21 [RFCv6 00/14] Bluetooth: Change socket lock to l2cap_chan lock Emeltchenko Andrei
2012-02-20 14:21 ` [RFCv6 01/14] Bluetooth: trivial: Fix long line Emeltchenko Andrei
2012-02-20 16:38   ` Gustavo Padovan
2012-02-20 14:21 ` [RFCv6 02/14] Bluetooth: Revert to mutexes from RCU list Emeltchenko Andrei
2012-02-20 14:44   ` Ulisses Furquim
2012-02-20 18:33   ` Gustavo Padovan
2012-02-20 18:57     ` Ulisses Furquim
2012-02-20 22:46       ` Gustavo Padovan [this message]
2012-02-20 14:21 ` [RFCv6 03/14] Bluetooth: Add l2cap_chan_lock Emeltchenko Andrei
2012-02-20 14:21 ` [RFCv6 04/14] Bluetooth: Add locked and unlocked state_change Emeltchenko Andrei
2012-02-20 14:21 ` [RFCv6 05/14] Bluetooth: Add socket error function Emeltchenko Andrei
2012-02-20 14:21 ` [RFCv6 06/14] Bluetooth: Add unlocked __l2cap_chan_add function Emeltchenko Andrei
2012-02-20 14:21 ` [RFCv6 07/14] Bluetooth: Use chan lock in timers Emeltchenko Andrei
2012-02-20 14:21 ` [RFCv6 08/14] Bluetooth: Use chan lock in L2CAP sig commands Emeltchenko Andrei
2012-02-20 14:21 ` [RFCv6 09/14] Bluetooth: Use chan lock in L2CAP conn start Emeltchenko Andrei
2012-02-20 14:21 ` [RFCv6 10/14] Bluetooth: Use chan lock in receiving data Emeltchenko Andrei
2012-02-20 14:21 ` [RFCv6 11/14] Bluetooth: Change locking logic for conn/chan ready Emeltchenko Andrei
2012-02-20 14:21 ` [RFCv6 12/14] Bluetooth: Change locking logic in security_cfm Emeltchenko Andrei
2012-02-20 14:21 ` [RFCv6 13/14] Bluetooth: Use l2cap chan lock in socket connect Emeltchenko Andrei
2012-02-20 14:21 ` [RFCv6 14/14] Bluetooth: Remove socket lock check Emeltchenko Andrei
2012-02-20 14:29 ` [RFCv6 00/14] Bluetooth: Change socket lock to l2cap_chan lock Marcel Holtmann
2012-02-20 14:44   ` Emeltchenko Andrei
2012-02-20 14:52     ` Ulisses Furquim
2012-02-20 14:53       ` Marcel Holtmann
2012-02-21 10:57         ` Emeltchenko Andrei

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=20120220224643.GA866@joana \
    --to=padovan@profusion.mobi \
    --cc=Andrei.Emeltchenko.news@gmail.com \
    --cc=linux-bluetooth@vger.kernel.org \
    --cc=ulisses@profusion.mobi \
    /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).