Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Sat, 5 Feb 2022 06:04:50 +0200
From:   Laurent Pinchart <laurent.pinchart@ideasonboard.com>
To:     Alexander Stein <alexander.stein@ew.tq-group.com>
Cc:     Steve Longerbeam <slongerbeam@gmail.com>,
        Philipp Zabel <p.zabel@pengutronix.de>,
        Mauro Carvalho Chehab <mchehab@kernel.org>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Shawn Guo <shawnguo@kernel.org>,
        Sascha Hauer <s.hauer@pengutronix.de>,
        Pengutronix Kernel Team <kernel@pengutronix.de>,
        Fabio Estevam <festevam@gmail.com>,
        NXP Linux Team <linux-imx@nxp.com>,
        Rui Miguel Silva <rmfrfs@gmail.com>,
        Dorota Czaplejewicz <dorota.czaplejewicz@puri.sm>,
        linux-media@vger.kernel.org, linux-staging@lists.linux.dev,
        linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 3/8] media: imx: Use dedicated format handler for i.MX7/8
Message-ID: <Yf33YviITGDnolHw@pendragon.ideasonboard.com>
References: <20220204121514.2762676-1-alexander.stein@ew.tq-group.com>
 <20220204121514.2762676-4-alexander.stein@ew.tq-group.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <20220204121514.2762676-4-alexander.stein@ew.tq-group.com>
Precedence: bulk

Hi Alexander and Dorota,

Thank you for the patch.

On Fri, Feb 04, 2022 at 01:15:09PM +0100, Alexander Stein wrote:
> From: Dorota Czaplejewicz <dorota.czaplejewicz@puri.sm>
> 
> This splits out a format handler which takes into account
> the capabilities of the i.MX7/8 video device,
> as opposed to the default handler compatible with both i.MX5/6 and i.MX7/8.
> 
> Signed-off-by: Dorota Czaplejewicz <dorota.czaplejewicz@puri.sm>
> Signed-off-by: Alexander Stein <alexander.stein@ew.tq-group.com>
> ---
> Changes in comparison to original commit from Dorota:
> * is_imx56 is used instead of enum
> 
>  drivers/staging/media/imx/imx-media-utils.c | 56 +++++++++++++++++++--
>  1 file changed, 52 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/staging/media/imx/imx-media-utils.c b/drivers/staging/media/imx/imx-media-utils.c
> index 0daa6aad45f4..32aaa2e81bea 100644
> --- a/drivers/staging/media/imx/imx-media-utils.c
> +++ b/drivers/staging/media/imx/imx-media-utils.c
> @@ -516,10 +516,9 @@ void imx_media_try_colorimetry(struct v4l2_mbus_framefmt *tryfmt,
>  }
>  EXPORT_SYMBOL_GPL(imx_media_try_colorimetry);
>  
> -int imx_media_mbus_fmt_to_pix_fmt(struct v4l2_pix_format *pix,
> -				  const struct v4l2_mbus_framefmt *mbus,
> -				  const struct imx_media_pixfmt *cc,
> -				  bool is_imx56)
> +static int imx56_media_mbus_fmt_to_pix_fmt(struct v4l2_pix_format *pix,
> +					   const struct v4l2_mbus_framefmt *mbus,
> +					   const struct imx_media_pixfmt *cc)
>  {
>  	u32 width;
>  	u32 stride;
> @@ -568,6 +567,55 @@ int imx_media_mbus_fmt_to_pix_fmt(struct v4l2_pix_format *pix,
>  
>  	return 0;
>  }
> +
> +static int imx78_media_mbus_fmt_to_pix_fmt(struct v4l2_pix_format *pix,
> +					   const struct v4l2_mbus_framefmt *mbus,
> +					   const struct imx_media_pixfmt *cc)
> +{
> +	int ret;
> +
> +	if (!cc)
> +		cc = imx_media_find_mbus_format(mbus->code, PIXFMT_SEL_ANY);
> +
> +	if (!cc)
> +		return -EINVAL;
> +	/*
> +	 * The hardware can handle line lengths divisible by 4 pixels
> +	 * as long as the whole buffer size ends up divisible by 8 bytes.
> +	 * If not, use the value of 8 pixels recommended in the datasheet.
> +	 */
> +	ret = v4l2_fill_pixfmt(pix, cc->fourcc,
> +			       round_up(mbus->width, 4), mbus->height);
> +	if (ret)
> +		return ret;
> +
> +	/* Only 8bits-per-pixel formats may need to get aligned to 8 pixels,
> +	 * because both 10-bit and 16-bit pixels occupy 2 bytes.
> +	 * In those, 4-pixel aligmnent is equal to 8-byte alignment.
> +	 */
> +	if (pix->sizeimage % 8 != 0)
> +		ret = v4l2_fill_pixfmt(pix, cc->fourcc,
> +				       round_up(mbus->width, 8), mbus->height);

I think you could simplify this by using cc->bpp to figure out the
alignment instead of calling v4l2_fill_pixfmt() twice.

> +
> +	pix->colorspace = mbus->colorspace;
> +	pix->xfer_func = mbus->xfer_func;
> +	pix->ycbcr_enc = mbus->ycbcr_enc;
> +	pix->quantization = mbus->quantization;
> +	pix->field = mbus->field;

Should v4l2_fill_pixfmt() be updated to handle colorspace and field too
instead of doing it manually here ?

> +
> +	return ret;
> +}
> +
> +int imx_media_mbus_fmt_to_pix_fmt(struct v4l2_pix_format *pix,
> +				  const struct v4l2_mbus_framefmt *mbus,
> +				  const struct imx_media_pixfmt *cc,
> +				  bool is_imx56)
> +{
> +	if (is_imx56)
> +		return imx56_media_mbus_fmt_to_pix_fmt(pix, mbus, cc);
> +	else
> +		return imx78_media_mbus_fmt_to_pix_fmt(pix, mbus, cc);
> +}
>  EXPORT_SYMBOL_GPL(imx_media_mbus_fmt_to_pix_fmt);
>  
>  void imx_media_free_dma_buf(struct device *dev,

-- 
Regards,

Laurent Pinchart