linux-media.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jacopo Mondi <jacopo@jmondi.org>
To: Dave Stevenson <dave.stevenson@raspberrypi.com>
Cc: paul.j.murphy@intel.com, daniele.alessandrelli@intel.com,
	linux-media@vger.kernel.org
Subject: Re: [PATCH 12/16] media: i2c: ov9282: Make V4L2_CID_HBLANK r/w
Date: Thu, 6 Oct 2022 13:53:51 +0200	[thread overview]
Message-ID: <20221006115351.2c7o7xqwl252xtp7@uno.localdomain> (raw)
In-Reply-To: <CAPY8ntCEoEn4J9-gqMwH6L+48DT8fWjnxiBYkjmpzDAn0qcg5w@mail.gmail.com>

Hi Dave,

On Thu, Oct 06, 2022 at 12:33:01PM +0100, Dave Stevenson wrote:
> Hi Jacopo.
>
> On Thu, 6 Oct 2022 at 10:41, Jacopo Mondi <jacopo@jmondi.org> wrote:
> >
> > Hi Dave
> >
> > On Wed, Oct 05, 2022 at 04:28:05PM +0100, Dave Stevenson wrote:
> > > There's no reason why HBLANK has to be read-only as it
> > > only changes the TIMING_HTS register in the sensor.
> > >
> > > Remove the READ_ONLY flag, and add the relevant handling
> > > for it.
> > >
> > > The minimum value also varies based on whether continuous clock
> > > mode is being used or not, so allow hblank_min to depend on
> > > that.
> >
> > Interesting, do you know why they're different and why the continous
> > version is smaller ?
>
> Dropping from HS to LP-11 and back to HS for non-continuous mode takes
> a minimum amount of time, so the time per line on the MIPI bus is
> increased. Whilst there is a FIFO between the pixel array and MIPI
> block, having to expend that time on the MIPI side means that the FIFO
> can overflow.
>
> I haven't found this behaviour documented, but I was trying to work
> out the difference between this driver and our downstream driver [1]
> in max frame rate. The default value for TIMING_HTS of 0x2d8 (doubled
> to 1456 pixels) didn't work with this mainline driver, and it used
> 0x2fd (1530 pixels). The CSI clock mode turned out to be the
> difference.
> The minimum HTS has therefore been left alone for non-continuous clock
> mode, and the lower setting added for continuous clock.
>

Makes sense! Thanks for the explanation

Please add my tag to v2
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>

Thanks
  j

>   Dave
>
> [1] https://github.com/raspberrypi/linux/blob/rpi-5.15.y/drivers/media/i2c/ov9281.c
> based on https://github.com/rockchip-linux/kernel/blob/develop-4.4/drivers/media/i2c/ov9281.c
>
> > >
> > > Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
> > > ---
> > >  drivers/media/i2c/ov9282.c | 34 +++++++++++++++++++++-------------
> > >  1 file changed, 21 insertions(+), 13 deletions(-)
> > >
> > > diff --git a/drivers/media/i2c/ov9282.c b/drivers/media/i2c/ov9282.c
> > > index 12cbe401fd78..8e86aa7e4b2a 100644
> > > --- a/drivers/media/i2c/ov9282.c
> > > +++ b/drivers/media/i2c/ov9282.c
> > > @@ -22,6 +22,9 @@
> > >  #define OV9282_MODE_STANDBY  0x00
> > >  #define OV9282_MODE_STREAMING        0x01
> > >
> > > +#define OV9282_REG_TIMING_HTS        0x380c
> > > +#define OV9282_TIMING_HTS_MAX        0x7fff
> > > +
> > >  /* Lines per frame */
> > >  #define OV9282_REG_LPFR              0x380e
> > >
> > > @@ -99,7 +102,8 @@ struct ov9282_reg_list {
> > >   * struct ov9282_mode - ov9282 sensor mode structure
> > >   * @width: Frame width
> > >   * @height: Frame height
> > > - * @hblank: Horizontal blanking in lines
> > > + * @hblank_min: Minimum horizontal blanking in lines for non-continuous[0] and
> > > + *           continuous[1] clock modes
> > >   * @vblank: Vertical blanking in lines
> > >   * @vblank_min: Minimum vertical blanking in lines
> > >   * @vblank_max: Maximum vertical blanking in lines
> > > @@ -109,7 +113,7 @@ struct ov9282_reg_list {
> > >  struct ov9282_mode {
> > >       u32 width;
> > >       u32 height;
> > > -     u32 hblank;
> > > +     u32 hblank_min[2];
> > >       u32 vblank;
> > >       u32 vblank_min;
> > >       u32 vblank_max;
> > > @@ -249,8 +253,6 @@ static const struct ov9282_reg mode_1280x720_regs[] = {
> > >       {0x3809, 0x00},
> > >       {0x380a, 0x02},
> > >       {0x380b, 0xd0},
> > > -     {0x380c, 0x02},
> > > -     {0x380d, 0xfd},
> > >       {0x3810, 0x00},
> > >       {0x3811, 0x08},
> > >       {0x3812, 0x00},
> > > @@ -273,7 +275,7 @@ static const struct ov9282_mode supported_modes[] = {
> > >       {
> > >               .width = 1280,
> > >               .height = 720,
> > > -             .hblank = 250,
> > > +             .hblank_min = { 250, 176 },
> > >               .vblank = 1022,
> > >               .vblank_min = 41,
> > >               .vblank_max = 51540,
> > > @@ -399,15 +401,17 @@ static int ov9282_write_regs(struct ov9282 *ov9282,
> > >  static int ov9282_update_controls(struct ov9282 *ov9282,
> > >                                 const struct ov9282_mode *mode)
> > >  {
> > > +     u32 hblank_min;
> > >       int ret;
> > >
> > >       ret = __v4l2_ctrl_s_ctrl(ov9282->link_freq_ctrl, mode->link_freq_idx);
> > >       if (ret)
> > >               return ret;
> > >
> > > -     ret = __v4l2_ctrl_s_ctrl(ov9282->hblank_ctrl, mode->hblank);
> > > -     if (ret)
> > > -             return ret;
> > > +     hblank_min = mode->hblank_min[ov9282->noncontinuous_clock ? 0 : 1];
> > > +     ret =  __v4l2_ctrl_modify_range(ov9282->hblank_ctrl, hblank_min,
> > > +                                     OV9282_TIMING_HTS_MAX - mode->width, 1,
> > > +                                     hblank_min);
> > >
> > >       return __v4l2_ctrl_modify_range(ov9282->vblank_ctrl, mode->vblank_min,
> > >                                       mode->vblank_max, 1, mode->vblank);
> > > @@ -539,6 +543,10 @@ static int ov9282_set_ctrl(struct v4l2_ctrl *ctrl)
> > >       case V4L2_CID_VFLIP:
> > >               ret = ov9282_set_ctrl_vflip(ov9282, ctrl->val);
> > >               break;
> > > +     case V4L2_CID_HBLANK:
> > > +             ret = ov9282_write_reg(ov9282, OV9282_REG_TIMING_HTS, 2,
> > > +                                    (ctrl->val + ov9282->cur_mode->width) >> 1);
> > > +             break;
> > >       default:
> > >               dev_err(ov9282->dev, "Invalid control %d", ctrl->id);
> > >               ret = -EINVAL;
> > > @@ -1033,6 +1041,7 @@ static int ov9282_init_controls(struct ov9282 *ov9282)
> > >       struct v4l2_ctrl_handler *ctrl_hdlr = &ov9282->ctrl_handler;
> > >       const struct ov9282_mode *mode = ov9282->cur_mode;
> > >       struct v4l2_fwnode_device_properties props;
> > > +     u32 hblank_min;
> > >       u32 lpfr;
> > >       int ret;
> > >
> > > @@ -1091,14 +1100,13 @@ static int ov9282_init_controls(struct ov9282 *ov9282)
> > >       if (ov9282->link_freq_ctrl)
> > >               ov9282->link_freq_ctrl->flags |= V4L2_CTRL_FLAG_READ_ONLY;
> > >
> > > +     hblank_min = mode->hblank_min[ov9282->noncontinuous_clock ? 0 : 1];
> > >       ov9282->hblank_ctrl = v4l2_ctrl_new_std(ctrl_hdlr,
> > >                                               &ov9282_ctrl_ops,
> > >                                               V4L2_CID_HBLANK,
> > > -                                             OV9282_REG_MIN,
> > > -                                             OV9282_REG_MAX,
> > > -                                             1, mode->hblank);
> > > -     if (ov9282->hblank_ctrl)
> > > -             ov9282->hblank_ctrl->flags |= V4L2_CTRL_FLAG_READ_ONLY;
> > > +                                             hblank_min,
> > > +                                             OV9282_TIMING_HTS_MAX - mode->width,
> > > +                                             1, hblank_min);
> > >
> > >       ret = v4l2_fwnode_device_parse(ov9282->dev, &props);
> > >       if (!ret) {
> > > --
> > > 2.34.1
> > >

  reply	other threads:[~2022-10-06 11:54 UTC|newest]

Thread overview: 52+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-10-05 15:27 [PATCH 00/16] Updates to ov9282 sensor driver Dave Stevenson
2022-10-05 15:27 ` [PATCH 01/16] media: i2c: ov9282: Remove duplication of registers Dave Stevenson
2022-10-06  9:14   ` Jacopo Mondi
2022-10-05 15:27 ` [PATCH 02/16] media: i2c: ov9282: Split registers into common and mode specific Dave Stevenson
2022-10-06  9:15   ` Jacopo Mondi
2022-10-05 15:27 ` [PATCH 03/16] media: i2c: ov9282: Remove format code from the mode Dave Stevenson
2022-10-06  9:15   ` Jacopo Mondi
2022-10-05 15:27 ` [PATCH 04/16] media: i2c: ov9282: Remove pixel rate from mode definition Dave Stevenson
2022-10-06  9:17   ` Jacopo Mondi
2022-10-06 11:51     ` Dave Stevenson
2022-10-05 15:27 ` [PATCH 05/16] media: i2c: ov9281: Support more than 1 mode Dave Stevenson
2022-10-06  9:18   ` Jacopo Mondi
2022-10-26  7:22   ` Sakari Ailus
2022-10-05 15:27 ` [PATCH 06/16] media: i2c: ov9282: Correct HTS register for configured pixel rate Dave Stevenson
2022-10-06  9:23   ` Jacopo Mondi
2022-10-06 13:01     ` Dave Stevenson
2022-10-05 15:28 ` [PATCH 07/16] media: i2c: ov9282: Reduce vblank_min values based on testing Dave Stevenson
2022-10-06 11:56   ` Jacopo Mondi
2022-10-06 13:02     ` Dave Stevenson
2022-10-05 15:28 ` [PATCH 08/16] media: i2c: ov9282: Add selection for CSI2 clock mode Dave Stevenson
2022-10-06  9:24   ` Jacopo Mondi
2022-10-26  7:21   ` Sakari Ailus
2022-10-28 12:57     ` Dave Stevenson
2022-10-28 14:30       ` Sakari Ailus
2022-10-28 15:03         ` Dave Stevenson
2022-10-31 13:06           ` Sakari Ailus
2022-10-05 15:28 ` [PATCH 09/16] media: i2c: ov9282: Add the properties from fwnode Dave Stevenson
2022-10-06 11:57   ` Jacopo Mondi
2022-10-05 15:28 ` [PATCH 10/16] media: i2c: ov9282: Action CID_VBLANK when set Dave Stevenson
2022-10-06  9:29   ` Jacopo Mondi
2022-10-06 13:21     ` Dave Stevenson
2022-10-05 15:28 ` [PATCH 11/16] media: i2c: ov9282: Add HFLIP and VFLIP support Dave Stevenson
2022-10-06  9:38   ` Jacopo Mondi
2022-10-06 14:21     ` Dave Stevenson
2022-10-05 15:28 ` [PATCH 12/16] media: i2c: ov9282: Make V4L2_CID_HBLANK r/w Dave Stevenson
2022-10-06  9:41   ` Jacopo Mondi
2022-10-06 11:33     ` Dave Stevenson
2022-10-06 11:53       ` Jacopo Mondi [this message]
2022-10-05 15:28 ` [PATCH 13/16] media: i2c: ov9282: Add selection API calls for cropping info Dave Stevenson
2022-10-06  9:43   ` Jacopo Mondi
2022-10-06 11:39     ` Dave Stevenson
2022-10-06 11:54       ` Jacopo Mondi
2022-10-05 15:28 ` [PATCH 14/16] media: i2c: ov9282: Add support for 1280x800 and 640x400 modes Dave Stevenson
2022-10-06  9:48   ` Jacopo Mondi
2022-10-06 11:46     ` Dave Stevenson
2022-10-05 15:28 ` [PATCH 15/16] media: i2c: ov9282: Add support for 8bit readout Dave Stevenson
2022-10-06  9:57   ` Jacopo Mondi
2022-10-06 12:20     ` Dave Stevenson
2022-10-05 15:28 ` [PATCH 16/16] media: i2c: ov9282: Support event handlers Dave Stevenson
2022-10-06  9:59   ` Jacopo Mondi
2022-10-07 10:22     ` Dave Stevenson
2022-10-07 12:57       ` Jacopo Mondi

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=20221006115351.2c7o7xqwl252xtp7@uno.localdomain \
    --to=jacopo@jmondi.org \
    --cc=daniele.alessandrelli@intel.com \
    --cc=dave.stevenson@raspberrypi.com \
    --cc=linux-media@vger.kernel.org \
    --cc=paul.j.murphy@intel.com \
    /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).