linux-media.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jack Mitchell <ml@embed.me.uk>
To: Philipp Zabel <p.zabel@pengutronix.de>, linux-media@vger.kernel.org
Cc: Steve Longerbeam <steve_longerbeam@mentor.com>,
	Marek Vasut <marex@denx.de>, Hans Verkuil <hverkuil@xs4all.nl>,
	Gary Bisson <gary.bisson@boundarydevices.com>,
	kernel@pengutronix.de, Sascha Hauer <s.hauer@pengutronix.de>,
	Marc Kleine-Budde <mkl@pengutronix.de>
Subject: Re: [PATCH v2 08/21] [media] imx: Add i.MX IPUv3 capture driver
Date: Mon, 17 Oct 2016 12:32:15 +0100	[thread overview]
Message-ID: <a5a06050-f6e7-2031-4b14-312f085c5644@embed.me.uk> (raw)
In-Reply-To: <1476466481-24030-9-git-send-email-p.zabel@pengutronix.de>

Hi Philipp,

I'm looking at how I would enable a parallel greyscale camera using this 
set of drivers and am a little bit confused. Do you have an example 
somewhere of a devicetree with an input node. I also have a further note 
below:

<snip>

> +
> +static int ipu_capture_start_streaming(struct vb2_queue *vq, unsigned int count)
> +{
> +	struct ipu_capture *priv = vb2_get_drv_priv(vq);
> +	struct v4l2_subdev *csi_sd = priv->csi_sd;
> +	u32 width = priv->format.fmt.pix.width;
> +	u32 height = priv->format.fmt.pix.height;
> +	struct device *dev = priv->dev;
> +	int burstsize;
> +	struct ipu_capture_buffer *buf;
> +	int nfack_irq;
> +	int ret;
> +	const char *irq_name[2] = { "CSI0", "CSI1" };
> +	bool raw;
> +
> +	ret = ipu_capture_get_resources(priv);
> +	if (ret < 0) {
> +		dev_err(dev, "Failed to get resources: %d\n", ret);
> +		goto err_dequeue;
> +	}
> +
> +	ipu_cpmem_zero(priv->ipuch);
> +
> +	nfack_irq = ipu_idmac_channel_irq(priv->ipu, priv->ipuch,
> +					  IPU_IRQ_NFACK);
> +	ret = request_threaded_irq(nfack_irq, NULL,
> +				   ipu_capture_new_frame_handler, IRQF_ONESHOT,
> +				   irq_name[priv->id], priv);
> +	if (ret) {
> +		dev_err(dev, "Failed to request NFACK interrupt: %d\n", nfack_irq);
> +		goto put_resources;
> +	}
> +
> +	dev_dbg(dev, "width: %d height: %d, %.4s\n",
> +		width, height, (char *)&priv->format.fmt.pix.pixelformat);
> +
> +	ipu_cpmem_set_resolution(priv->ipuch, width, height);
> +
> +	raw = false;
> +
> +	if (raw && priv->smfc) {

How does this ever get used? If I were to set 1X8 greyscale it wouldn't 
ever take this path, correct?

> +		/*
> +		 * raw formats. We can only pass them through to memory
> +		 */
> +		u32 fourcc = priv->format.fmt.pix.pixelformat;
> +		int bytes;
> +
> +		switch (fourcc) {
> +		case V4L2_PIX_FMT_GREY:
> +			bytes = 1;
> +			break;
> +		case V4L2_PIX_FMT_Y10:
> +		case V4L2_PIX_FMT_Y16:
> +		case V4L2_PIX_FMT_UYVY:
> +		case V4L2_PIX_FMT_YUYV:
> +			bytes = 2;
> +			break;
> +		}
> +
> +		ipu_cpmem_set_stride(priv->ipuch, width * bytes);
> +		ipu_cpmem_set_format_passthrough(priv->ipuch, bytes * 8);
> +		/*
> +		 * According to table 37-727 (SMFC Burst Size), burstsize should
> +		 * be set to NBP[6:4] for PFS == 6. Unfortunately, with a 16-bit
> +		 * bus any value below 4 doesn't produce proper images.
> +		 */
> +		burstsize = (64 / bytes) >> 3;
> +	} else {
> +		/*
> +		 * formats we understand, we can write it in any format not requiring
> +		 * colorspace conversion.
> +		 */
> +		u32 fourcc = priv->format.fmt.pix.pixelformat;
> +
> +		switch (fourcc) {
> +		case V4L2_PIX_FMT_RGB32:
> +			ipu_cpmem_set_stride(priv->ipuch, width * 4);
> +			ipu_cpmem_set_fmt(priv->ipuch, fourcc);
> +			break;
> +		case V4L2_PIX_FMT_UYVY:
> +		case V4L2_PIX_FMT_YUYV:
> +			ipu_cpmem_set_stride(priv->ipuch, width * 2);
> +			ipu_cpmem_set_yuv_interleaved(priv->ipuch, fourcc);
> +			break;
> +		case V4L2_PIX_FMT_YUV420:
> +		case V4L2_PIX_FMT_YVU420:
> +		case V4L2_PIX_FMT_NV12:
> +		case V4L2_PIX_FMT_YUV422P:
> +			ipu_cpmem_set_stride(priv->ipuch, width);
> +			ipu_cpmem_set_fmt(priv->ipuch, fourcc);
> +			ipu_cpmem_set_yuv_planar(priv->ipuch, fourcc,
> +						 width, height);
> +			burstsize = 16;
> +			break;
> +		default:
> +			dev_err(dev, "invalid color format: %4.4s\n",
> +				(char *)&fourcc);
> +			ret = -EINVAL;
> +			goto free_irq;
> +		}
> +	}
> +

<snip>

  reply	other threads:[~2016-10-17 11:31 UTC|newest]

Thread overview: 61+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-10-14 17:34 [PATCH v2 00/21] Basic i.MX IPUv3 capture support Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 01/21] [media] v4l2-async: move code out of v4l2_async_notifier_register into v4l2_async_test_nofity_all Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 02/21] [media] v4l2-async: allow subdevices to add further subdevices to the notifier waiting list Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 03/21] [media] v4l: of: add v4l2_of_subdev_registered Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 04/21] [media] v4l2-async: add new subdevices to the tail of subdev_list Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 05/21] [media] imx: Add i.MX SoC wide media device driver Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 06/21] [media] imx: Add IPUv3 media common code Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 07/21] [media] imx-ipu: Add i.MX IPUv3 CSI subdevice driver Philipp Zabel
2016-11-08  7:18   ` Ying Liu
2016-10-14 17:34 ` [PATCH v2 08/21] [media] imx: Add i.MX IPUv3 capture driver Philipp Zabel
2016-10-17 11:32   ` Jack Mitchell [this message]
2016-10-17 11:35     ` Marek Vasut
2016-10-17 11:40       ` Jack Mitchell
2016-10-17 12:12     ` Philipp Zabel
2016-10-19 16:22       ` Jack Mitchell
2016-10-19 19:25         ` Marek Vasut
2016-10-26 11:29           ` Jack Mitchell
2016-11-08  7:21   ` Ying Liu
2016-10-14 17:34 ` [PATCH v2 09/21] [media] platform: add video-multiplexer subdevice driver Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 10/21] [media] imx: Add i.MX MIPI CSI-2 " Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 11/21] [media] tc358743: put lanes in STOP state before starting streaming Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 12/21] ARM: dts: imx6qdl: Add capture-subsystem node Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 13/21] ARM: dts: imx6qdl: Add mipi_ipu1/2 multiplexers, mipi_csi, and their connections Philipp Zabel
2016-11-08  8:23   ` Ying Liu
2016-10-14 17:34 ` [PATCH v2 14/21] ARM: dts: imx6qdl: Add MIPI CSI-2 D-PHY compatible and clocks Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 15/21] ARM: dts: nitrogen6x: Add dtsi for BD_HDMI_MIPI HDMI to MIPI CSI-2 receiver board Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 16/21] gpu: ipuv3: add ipu_csi_set_downsize Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 17/21] [media] imx-ipuv3-csi: support downsizing Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 18/21] [media] add mux and video interface bridge entity functions Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 19/21] [media] video-multiplexer: set entity function to mux Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 20/21] [media] imx: Set i.MX MIPI CSI-2 entity function to bridge Philipp Zabel
2016-10-14 17:34 ` [PATCH v2 21/21] [media] tc358743: set entity function to video interface bridge Philipp Zabel
2016-10-17 10:18 ` [PATCH v2 00/21] Basic i.MX IPUv3 capture support Gary Bisson
2016-10-17 12:41   ` Philipp Zabel
2016-10-19 21:30 ` Sakari Ailus
     [not found]   ` <CAH-u=807nRYzza0kTfOMv1AiWazk6FGJyz6W5_bYw7v9nOrccA@mail.gmail.com>
2016-12-29 15:10     ` Jean-Michel Hautbois
2016-12-29 20:51     ` Robert Schwebel
2016-12-30 19:06       ` Marek Vasut
2016-12-30 20:26         ` Steve Longerbeam
2017-01-02 13:51           ` Jean-Michel Hautbois
2017-01-02 14:45             ` Hans Verkuil
2017-01-02 14:59               ` Jean-Michel Hautbois
2017-01-02 19:19                 ` Fabio Estevam
2017-01-02 19:20                 ` Steve Longerbeam
2017-01-02 17:04               ` Fabio Estevam
2017-01-02 19:17               ` Steve Longerbeam
2017-01-09 19:43           ` Philipp Zabel
2017-01-10  0:15             ` Steve Longerbeam
2017-01-10 23:52               ` Steve Longerbeam
2017-01-11  9:12                 ` Jean-Michel Hautbois
2017-01-11 12:10                 ` Philipp Zabel
2017-01-12 23:22                   ` Steve Longerbeam
2017-01-12 23:43                     ` Steve Longerbeam
2017-01-13  2:22                       ` Steve Longerbeam
2017-01-13 11:18                         ` Philipp Zabel
2017-01-13 11:04                     ` Philipp Zabel
2017-01-13 11:05                     ` Philipp Zabel
2017-01-14 20:26                       ` Steve Longerbeam
2017-01-16 11:49                         ` Philipp Zabel
2017-01-11 12:11               ` Philipp Zabel
2017-01-24 18:11             ` Philipp Zabel

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=a5a06050-f6e7-2031-4b14-312f085c5644@embed.me.uk \
    --to=ml@embed.me.uk \
    --cc=gary.bisson@boundarydevices.com \
    --cc=hverkuil@xs4all.nl \
    --cc=kernel@pengutronix.de \
    --cc=linux-media@vger.kernel.org \
    --cc=marex@denx.de \
    --cc=mkl@pengutronix.de \
    --cc=p.zabel@pengutronix.de \
    --cc=s.hauer@pengutronix.de \
    --cc=steve_longerbeam@mentor.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).