From: Peter Hurley <peter@hurleysoftware.com>
To: Gianluca Anzolin <gianluca@sottospazio.it>
Cc: gustavo@padovan.org, linux-bluetooth@vger.kernel.org,
marcel@holtmann.org
Subject: Re: [PATCH 3/8] Move device initialization and shutdown to tty_port_operations
Date: Tue, 16 Jul 2013 16:48:22 -0400 [thread overview]
Message-ID: <51E5B196.7010704@hurleysoftware.com> (raw)
In-Reply-To: <1373661649-1385-3-git-send-email-gianluca@sottospazio.it>
On 07/12/2013 04:40 PM, Gianluca Anzolin wrote:
> Move the device initialization from rfcomm_tty_open to the .activate function of
> tty_port.
>
> At the same time move also the device shutdown from rfcomm_tty_close to the
> .shutdown function of tty_port.
This patch should be merged with patch 5/8 to maintain bisectability.
Otherwise, the .activate() and .shutdown() methods will not be called.
> Signed-off-by: Gianluca Anzolin <gianluca@sottospazio.it>
> ---
> net/bluetooth/rfcomm/tty.c | 120 ++++++++++++++++++++++++++++-----------------
> 1 file changed, 75 insertions(+), 45 deletions(-)
>
> diff --git a/net/bluetooth/rfcomm/tty.c b/net/bluetooth/rfcomm/tty.c
> index ff60171..9b0b064 100644
> --- a/net/bluetooth/rfcomm/tty.c
> +++ b/net/bluetooth/rfcomm/tty.c
> @@ -149,8 +149,80 @@ static void rfcomm_tty_copy_pending(struct rfcomm_dev *dev)
> tty_flip_buffer_push(&dev->port);
> }
>
> +/*
> + * Do the device-specific initialization by opening the DLC and waiting for
> + * a connection.
> + */
> +static int rfcomm_dev_activate(struct tty_port *port, struct tty_struct *tty)
> +{
> + DECLARE_WAITQUEUE(wait, current);
> + struct rfcomm_dev *dev = container_of(port, struct rfcomm_dev, port);
> + struct rfcomm_dlc *dlc = dev->dlc;
> + int err;
> +
> + err = rfcomm_dlc_open(dlc, &dev->src, &dev->dst, dev->channel);
> + if (err < 0)
> + goto error_no_dlc;
> +
> + /* Wait for DLC to connect */
> + add_wait_queue(&dev->wait, &wait);
> + while (1) {
> + set_current_state(TASK_INTERRUPTIBLE);
> +
> + if (dlc->state == BT_CLOSED) {
> + err = -dev->err;
> + break;
> + }
> +
> + if (dlc->state == BT_CONNECTED)
> + break;
Please consider moving these dlc->state tests into a
.carrier_raised() port method (this is what the gsm
driver does). Then this wait loop could go away.
> +
> + if (signal_pending(current)) {
> + err = -EINTR;
> + break;
> + }
> +
> + tty_unlock(tty);
> + schedule();
> + tty_lock(tty);
> + }
> + set_current_state(TASK_RUNNING);
> + remove_wait_queue(&dev->wait, &wait);
> +
> + if (err < 0)
> + goto error_no_connection;
> +
> + device_move(dev->tty_dev, rfcomm_get_device(dev),
> + DPM_ORDER_DEV_AFTER_PARENT);
> +
> + rfcomm_tty_copy_pending(dev);
> + rfcomm_dlc_unthrottle(dlc);
> + return 0;
> +
> +error_no_connection:
> + rfcomm_dlc_close(dlc, err);
> +error_no_dlc:
> + return err;
> +}
> +
> +/*
> + * Undo the device initialization by closing the DLC
> + */
> +static void rfcomm_dev_shutdown(struct tty_port *port)
> +{
> + struct rfcomm_dev *dev = container_of(port, struct rfcomm_dev, port);
> +
> + if (dev->tty_dev->parent)
> + device_move(dev->tty_dev, NULL, DPM_ORDER_DEV_LAST);
> +
> + /* Close DLC */
> + rfcomm_dlc_close(dev->dlc, 0);
> +}
> +
> static const struct tty_port_operations rfcomm_port_ops = {
> .destruct = rfcomm_dev_destruct,
> + .activate = rfcomm_dev_activate,
> + .shutdown = rfcomm_dev_shutdown,
> };
>
> static struct rfcomm_dev *__rfcomm_dev_get(int id)
> @@ -647,11 +719,10 @@ static void rfcomm_dev_modem_status(struct rfcomm_dlc *dlc, u8 v24_sig)
> /* ---- TTY functions ---- */
> static int rfcomm_tty_open(struct tty_struct *tty, struct file *filp)
> {
> - DECLARE_WAITQUEUE(wait, current);
> struct rfcomm_dev *dev;
> struct rfcomm_dlc *dlc;
> unsigned long flags;
> - int err, id;
> + int id;
>
> id = tty->index;
>
> @@ -685,44 +756,7 @@ static int rfcomm_tty_open(struct tty_struct *tty, struct file *filp)
> rfcomm_dlc_unlock(dlc);
> set_bit(RFCOMM_TTY_ATTACHED, &dev->flags);
>
> - err = rfcomm_dlc_open(dlc, &dev->src, &dev->dst, dev->channel);
> - if (err < 0)
> - return err;
> -
> - /* Wait for DLC to connect */
> - add_wait_queue(&dev->wait, &wait);
> - while (1) {
> - set_current_state(TASK_INTERRUPTIBLE);
> -
> - if (dlc->state == BT_CLOSED) {
> - err = -dev->err;
> - break;
> - }
> -
> - if (dlc->state == BT_CONNECTED)
> - break;
> -
> - if (signal_pending(current)) {
> - err = -EINTR;
> - break;
> - }
> -
> - tty_unlock(tty);
> - schedule();
> - tty_lock(tty);
> - }
> - set_current_state(TASK_RUNNING);
> - remove_wait_queue(&dev->wait, &wait);
> -
> - if (err == 0)
> - device_move(dev->tty_dev, rfcomm_get_device(dev),
> - DPM_ORDER_DEV_AFTER_PARENT);
> -
> - rfcomm_tty_copy_pending(dev);
> -
> - rfcomm_dlc_unthrottle(dev->dlc);
> -
> - return err;
> + return 0;
> }
>
> static void rfcomm_tty_close(struct tty_struct *tty, struct file *filp)
> @@ -739,12 +773,8 @@ static void rfcomm_tty_close(struct tty_struct *tty, struct file *filp)
> spin_lock_irqsave(&dev->port.lock, flags);
> if (!--dev->port.count) {
> spin_unlock_irqrestore(&dev->port.lock, flags);
> - if (dev->tty_dev->parent)
> - device_move(dev->tty_dev, NULL, DPM_ORDER_DEV_LAST);
> -
> - /* Close DLC and dettach TTY */
> - rfcomm_dlc_close(dev->dlc, 0);
>
> + /* detach the TTY */
> clear_bit(RFCOMM_TTY_ATTACHED, &dev->flags);
>
> rfcomm_dlc_lock(dev->dlc);
>
next prev parent reply other threads:[~2013-07-16 20:48 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-07-12 20:40 [PATCH 1/8] Take proper tty references in net/bluetooth/rfcomm/tty.c Gianluca Anzolin
2013-07-12 20:40 ` [PATCH 2/8] Move functions before the definition of rfcomm_port_ops Gianluca Anzolin
2013-07-16 15:14 ` Peter Hurley
2013-07-12 20:40 ` [PATCH 3/8] Move device initialization and shutdown to tty_port_operations Gianluca Anzolin
2013-07-16 20:48 ` Peter Hurley [this message]
2013-07-20 7:10 ` Gianluca Anzolin
2013-07-20 14:11 ` Peter Hurley
2013-07-21 8:08 ` Gianluca Anzolin
2013-07-21 17:04 ` Peter Hurley
2013-07-21 17:31 ` Gianluca Anzolin
2013-07-12 20:40 ` [PATCH 4/8] Move tty initialization and cleanup out of open/close Gianluca Anzolin
2013-07-16 19:07 ` Peter Hurley
2013-07-12 20:40 ` [PATCH 5/8] Use the tty_port_* functions in tty_open/tty_close/tty_hangup Gianluca Anzolin
2013-07-16 20:51 ` Peter Hurley
2013-07-17 8:03 ` Gianluca Anzolin
2013-07-12 20:40 ` [PATCH 6/8] Fix the reference counting of tty_port Gianluca Anzolin
2013-07-17 14:02 ` Peter Hurley
2013-07-17 17:05 ` Gianluca Anzolin
2013-07-17 18:10 ` Peter Hurley
2013-07-18 12:45 ` Peter Hurley
2013-07-18 14:13 ` Gianluca Anzolin
2013-07-18 15:19 ` Peter Hurley
2013-07-12 20:40 ` [PATCH 7/8] Avoid a circular dependency between dev and dev->dlc Gianluca Anzolin
2013-07-12 20:40 ` [PATCH 8/8] Add module_put in rfcomm_dev_add error path Gianluca Anzolin
2013-07-17 15:20 ` Peter Hurley
2013-07-16 14:53 ` [PATCH 1/8] Take proper tty references in net/bluetooth/rfcomm/tty.c Peter Hurley
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=51E5B196.7010704@hurleysoftware.com \
--to=peter@hurleysoftware.com \
--cc=gianluca@sottospazio.it \
--cc=gustavo@padovan.org \
--cc=linux-bluetooth@vger.kernel.org \
--cc=marcel@holtmann.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).