All of lore.kernel.org
 help / color / mirror / Atom feed
From: Alexander Stein <alexander.stein@ew.tq-group.com>
To: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: linux-media@vger.kernel.org, Sakari Ailus <sakari.ailus@iki.fi>,
	Manivannan Sadhasivam <mani@kernel.org>,
	Dave Stevenson <dave.stevenson@raspberrypi.com>
Subject: Re: [PATCH v2 14/17] media: i2c: imx290: Initialize runtime PM before subdev
Date: Mon, 16 Jan 2023 13:29:52 +0100	[thread overview]
Message-ID: <4243355.ejJDZkT8p0@steina-w> (raw)
In-Reply-To: <Y8U5wsnA6BS9PNkP@pendragon.ideasonboard.com>

Hi Laurent,

Am Montag, 16. Januar 2023, 12:49:22 CET schrieb Laurent Pinchart:
> Hi Alexander,
> 
> On Mon, Jan 16, 2023 at 12:14:46PM +0100, Alexander Stein wrote:
> > Am Samstag, 14. Januar 2023, 18:17:59 CET schrieb Laurent Pinchart:
> > > Initializing the subdev before runtime PM means that no subdev
> > > initialization can interact with the runtime PM framework. This can be
> > > problematic when modifying controls, as the .s_ctrl() handler commonly
> > > calls pm_runtime_get_if_in_use(). These code paths are not trivial,
> > > making the driver fragile and possibly causing subtle bugs.
> > > 
> > > To make the subdev initialization more robust, initialize runtime PM
> > > first.
> > > 
> > > Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> > 
> > Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> 
> Thank you for giving me my own ack, but yours would be more useful :-)

oh... Sorry. I messed up my copy and paste :(

Acked-by: Alexander Stein <alexander.stein@ew.tq-group.com>

> > > ---
> > > 
> > >  drivers/media/i2c/imx290.c | 57 ++++++++++++++++++++++----------------
> > >  1 file changed, 33 insertions(+), 24 deletions(-)
> > > 
> > > diff --git a/drivers/media/i2c/imx290.c b/drivers/media/i2c/imx290.c
> > > index 6a3e93c10fb1..c3ffb23104bf 100644
> > > --- a/drivers/media/i2c/imx290.c
> > > +++ b/drivers/media/i2c/imx290.c
> > > @@ -581,9 +581,7 @@ static int imx290_set_ctrl(struct v4l2_ctrl *ctrl)
> > > 
> > >  	/*
> > >  	
> > >  	 * Return immediately for controls that don't need to be applied to
> > >  	 the
> > > 
> > > -	 * device. This includes all controls modified in
> > > imx290_ctrl_update(),
> > > -	 * which is called at probe time before runtime PM is initialized, 
so
> > > -	 * we can't proceed to the pm_runtime_get_if_in_use() call below.
> > > +	 * device.
> > > 
> > >  	 */
> > >  	
> > >  	switch (ctrl->id) {
> > > 
> > >  	case V4L2_CID_LINK_FREQ:
> > > @@ -1054,22 +1052,20 @@ static void imx290_subdev_cleanup(struct imx290
> > > *imx290)> > 
> > >   * Power management
> > >   */
> > > 
> > > -static int imx290_power_on(struct device *dev)
> > > +static int imx290_power_on(struct imx290 *imx290)
> > > 
> > >  {
> > > 
> > > -	struct v4l2_subdev *sd = dev_get_drvdata(dev);
> > > -	struct imx290 *imx290 = to_imx290(sd);
> > > 
> > >  	int ret;
> > >  	
> > >  	ret = clk_prepare_enable(imx290->xclk);
> > >  	if (ret) {
> > > 
> > > -		dev_err(dev, "Failed to enable clock\n");
> > > +		dev_err(imx290->dev, "Failed to enable clock\n");
> > > 
> > >  		return ret;
> > >  	
> > >  	}
> > >  	
> > >  	ret = regulator_bulk_enable(ARRAY_SIZE(imx290->supplies),
> > >  	
> > >  				    imx290->supplies);
> > >  	
> > >  	if (ret) {
> > > 
> > > -		dev_err(dev, "Failed to enable regulators\n");
> > > +		dev_err(imx290->dev, "Failed to enable regulators\n");
> > > 
> > >  		clk_disable_unprepare(imx290->xclk);
> > >  		return ret;
> > >  	
> > >  	}
> > > 
> > > @@ -1084,20 +1080,33 @@ static int imx290_power_on(struct device *dev)
> > > 
> > >  	return 0;
> > >  
> > >  }
> > > 
> > > -static int imx290_power_off(struct device *dev)
> > > +static void imx290_power_off(struct imx290 *imx290)
> > > 
> > >  {
> > > 
> > > -	struct v4l2_subdev *sd = dev_get_drvdata(dev);
> > > -	struct imx290 *imx290 = to_imx290(sd);
> > > -
> > > 
> > >  	clk_disable_unprepare(imx290->xclk);
> > >  	gpiod_set_value_cansleep(imx290->rst_gpio, 1);
> > >  	regulator_bulk_disable(ARRAY_SIZE(imx290->supplies),
> > >  	imx290->supplies);
> > > 
> > > +}
> > > +
> > > +static int imx290_runtime_resume(struct device *dev)
> > > +{
> > > +	struct v4l2_subdev *sd = dev_get_drvdata(dev);
> > > +	struct imx290 *imx290 = to_imx290(sd);
> > > +
> > > +	return imx290_power_on(imx290);
> > > +}
> > > +
> > > +static int imx290_runtime_suspend(struct device *dev)
> > > +{
> > > +	struct v4l2_subdev *sd = dev_get_drvdata(dev);
> > > +	struct imx290 *imx290 = to_imx290(sd);
> > > +
> > > +	imx290_power_off(imx290);
> > > 
> > >  	return 0;
> > >  
> > >  }
> > >  
> > >  static const struct dev_pm_ops imx290_pm_ops = {
> > > 
> > > -	SET_RUNTIME_PM_OPS(imx290_power_off, imx290_power_on, NULL)
> > > +	SET_RUNTIME_PM_OPS(imx290_runtime_suspend, imx290_runtime_resume,
> > > NULL)
> > > 
> > >  };
> > >  
> > >  /*
> > >  ----------------------------------------------------------------------
> > >  ------> > 
> > > @@ -1276,20 +1285,15 @@ static int imx290_probe(struct i2c_client
> > > *client)
> > > if (ret)
> > > 
> > >  		return ret;
> > > 
> > > -	/* Initialize the V4L2 subdev. */
> > > -	ret = imx290_subdev_init(imx290);
> > > -	if (ret)
> > > -		return ret;
> > > -
> > > 
> > >  	/*
> > >  	
> > >  	 * Enable power management. The driver supports runtime PM, but 
needs
> > >  	 to
> > >  	 * work when runtime PM is disabled in the kernel. To that end, 
power
> > >  	 * the sensor on manually here.
> > >  	 */
> > > 
> > > -	ret = imx290_power_on(dev);
> > > +	ret = imx290_power_on(imx290);
> > > 
> > >  	if (ret < 0) {
> > >  	
> > >  		dev_err(dev, "Could not power on the device\n");
> > > 
> > > -		goto err_subdev;
> > > +		return ret;
> > > 
> > >  	}
> > >  	
> > >  	/*
> > > 
> > > @@ -1303,6 +1307,11 @@ static int imx290_probe(struct i2c_client
> > > *client)
> > > 
> > >  	pm_runtime_set_autosuspend_delay(dev, 1000);
> > >  	pm_runtime_use_autosuspend(dev);
> > > 
> > > +	/* Initialize the V4L2 subdev. */
> > > +	ret = imx290_subdev_init(imx290);
> > > +	if (ret)
> > > +		goto err_pm;
> > > +
> > > 
> > >  	/*
> > >  	
> > >  	 * Finally, register the V4L2 subdev. This must be done after
> > >  	 * initializing everything as the subdev can be used immediately 
after
> > > 
> > > @@ -1323,12 +1332,12 @@ static int imx290_probe(struct i2c_client
> > > *client)
> > > 
> > >  	return 0;
> > > 
> > > +err_subdev:
> > > +	imx290_subdev_cleanup(imx290);
> > > 
> > >  err_pm:
> > >  	pm_runtime_disable(dev);
> > >  	pm_runtime_put_noidle(dev);
> > > 
> > > -	imx290_power_off(dev);
> > > -err_subdev:
> > > -	imx290_subdev_cleanup(imx290);
> > > +	imx290_power_off(imx290);
> > > 
> > >  	return ret;
> > >  
> > >  }
> > > 
> > > @@ -1346,7 +1355,7 @@ static void imx290_remove(struct i2c_client
> > > *client)
> > > 
> > >  	 */
> > >  	
> > >  	pm_runtime_disable(imx290->dev);
> > >  	if (!pm_runtime_status_suspended(imx290->dev))
> > > 
> > > -		imx290_power_off(imx290->dev);
> > > +		imx290_power_off(imx290);
> > > 
> > >  	pm_runtime_set_suspended(imx290->dev);
> > >  
> > >  }





  reply	other threads:[~2023-01-16 12:30 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-01-14 17:17 [PATCH v2 00/17] media: i2c: imx290: Miscellaneous improvements Laurent Pinchart
2023-01-14 17:17 ` [PATCH v2 01/17] media: i2c: imx290: Group functions in sections Laurent Pinchart
2023-01-14 17:17   ` [PATCH v2 02/17] media: i2c: imx290: Factor out subdev init and cleanup to functions Laurent Pinchart
2023-01-16  8:09     ` Alexander Stein
2023-01-14 17:17   ` [PATCH v2 03/17] media: i2c: imx290: Factor out control update code to a function Laurent Pinchart
2023-01-16  9:20     ` Alexander Stein
2023-01-14 17:17   ` [PATCH v2 04/17] media: i2c: imx290: Access link_freq_index directly Laurent Pinchart
2023-01-14 17:17   ` [PATCH v2 05/17] media: i2c: imx290: Pass format and mode to imx290_calc_pixel_rate() Laurent Pinchart
2023-01-16 10:22     ` Alexander Stein
2023-01-16 10:52       ` Laurent Pinchart
2023-01-14 17:17   ` [PATCH v2 06/17] media: i2c: imx290: Compute pixel rate and blanking in one place Laurent Pinchart
2023-01-14 17:17   ` [PATCH v2 07/17] media: i2c: imx290: Factor out black level setting to a function Laurent Pinchart
2023-01-14 17:17   ` [PATCH v2 08/17] media: i2c: imx290: Factor out DT parsing to separate function Laurent Pinchart
2023-01-16 10:33     ` Alexander Stein
2023-01-16 10:48       ` Laurent Pinchart
2023-01-16 11:19         ` Alexander Stein
2023-01-14 17:17   ` [PATCH v2 09/17] media: i2c: imx290: Use dev_err_probe() Laurent Pinchart
2023-01-14 17:17   ` [PATCH v2 10/17] media: i2c: imx290: Factor out clock initialization to separate function Laurent Pinchart
2023-01-14 17:17   ` [PATCH v2 11/17] media: i2c: imx290: Use V4L2 subdev active state Laurent Pinchart
2023-01-16 11:04     ` Alexander Stein
2023-01-16 11:58       ` Laurent Pinchart
2023-01-16 14:07         ` Laurent Pinchart
2023-01-14 17:17   ` [PATCH v2 12/17] media: i2c: imx290: Rename, extend and expand usage of imx290_pixfmt Laurent Pinchart
2023-01-14 17:17   ` [PATCH v2 13/17] media: i2c: imx290: Use runtime PM autosuspend Laurent Pinchart
2023-01-16 11:11     ` Alexander Stein
2023-01-14 17:17   ` [PATCH v2 14/17] media: i2c: imx290: Initialize runtime PM before subdev Laurent Pinchart
2023-01-16 11:14     ` Alexander Stein
2023-01-16 11:49       ` Laurent Pinchart
2023-01-16 12:29         ` Alexander Stein [this message]
2023-01-14 17:18   ` [PATCH v2 15/17] media: i2c: imx290: Configure data lanes at start time Laurent Pinchart
2023-01-14 17:18   ` [PATCH v2 16/17] media: i2c: imx290: Simplify imx290_set_data_lanes() Laurent Pinchart
2023-01-16 11:16     ` Alexander Stein
2023-01-14 17:18   ` [PATCH v2 17/17] media: i2c: imx290: Handle error from imx290_set_data_lanes() Laurent Pinchart
2023-01-16 11:17     ` Alexander Stein

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=4243355.ejJDZkT8p0@steina-w \
    --to=alexander.stein@ew.tq-group.com \
    --cc=dave.stevenson@raspberrypi.com \
    --cc=laurent.pinchart@ideasonboard.com \
    --cc=linux-media@vger.kernel.org \
    --cc=mani@kernel.org \
    --cc=sakari.ailus@iki.fi \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.