linux-clk.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Sowjanya Komatineni <skomatineni@nvidia.com>
To: Helen Koike <helen.koike@collabora.com>,
	<thierry.reding@gmail.com>, <jonathanh@nvidia.com>,
	<frankc@nvidia.com>, <hverkuil@xs4all.nl>
Cc: <linux-media@vger.kernel.org>, <devicetree@vger.kernel.org>,
	<linux-clk@vger.kernel.org>, <linux-tegra@vger.kernel.org>,
	<linux-kernel@vger.kernel.org>
Subject: Re: [RFC PATCH v1 4/5] media: tegra: Add Tegra Video input driver for Tegra210
Date: Tue, 28 Jan 2020 16:49:04 -0800	[thread overview]
Message-ID: <017ca95e-7dd3-2d04-8d84-9047ac4e548b@nvidia.com> (raw)
In-Reply-To: <162488d0-4e74-963a-3366-e4c1f7cf04ca@nvidia.com>


On 1/28/20 2:13 PM, Sowjanya Komatineni wrote:
>
> On 1/28/20 1:45 PM, Helen Koike wrote:
>> External email: Use caution opening links or attachments
>>
>>
>> Hi Sowjanya,
>>
>> I just took a really quick look, I didn't check the driver in deep, 
>> so just some small comments below.
>>
>> On 1/28/20 4:23 PM, Sowjanya Komatineni wrote:
>>> Tegra210 contains a powerful Video Input (VI) hardware controller
>>> which can support up to 6 MIPI CSI camera sensors.
>>>
>>> Each Tegra CSI port can be one-to-one mapped to VI channel and can
>>> capture from an external camera sensor connected to CSI or from
>>> built-in test pattern generator.
>>>
>>> Tegra210 supports built-in test pattern generator from CSI to VI.
>>>
>>> This patch adds a V4L2 media controller and capture driver support
>>> for Tegra210 built-in CSI to VI test pattern generator.
>>>
>>> Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
>> Could you send us the output of media-ctl --print-dot ? So we can 
>> view the media topology easily?
> root@tegra-ubuntu:/home/ubuntu# ./media-ctl --print-dot
> digraph board {
>         rankdir=TB
>         n00000001 [label="54080000.vi-output-0\n/dev/video0", 
> shape=box, style=filled, fillcolor=yellow]
>         n00000005 [label="54080000.vi-output-1\n/dev/video1", 
> shape=box, style=filled, fillcolor=yellow]
>         n00000009 [label="54080000.vi-output-2\n/dev/video2", 
> shape=box, style=filled, fillcolor=yellow]
>         n0000000d [label="54080000.vi-output-3\n/dev/video3", 
> shape=box, style=filled, fillcolor=yellow]
>         n00000011 [label="54080000.vi-output-4\n/dev/video4", 
> shape=box, style=filled, fillcolor=yellow]
>         n00000015 [label="54080000.vi-output-5\n/dev/video5", 
> shape=box, style=filled, fillcolor=yellow]
>         n00000019 [label="{{} | tpg-0 | {<port0> 0}}", shape=Mrecord, 
> style=filled, fillcolor=green]
>         n00000019:port0 -> n00000001
>         n0000001d [label="{{} | tpg-1 | {<port0> 0}}", shape=Mrecord, 
> style=filled, fillcolor=green]
>         n0000001d:port0 -> n00000005
>         n00000021 [label="{{} | tpg-2 | {<port0> 0}}", shape=Mrecord, 
> style=filled, fillcolor=green]
>         n00000021:port0 -> n00000009
>         n00000025 [label="{{} | tpg-3 | {<port0> 0}}", shape=Mrecord, 
> style=filled, fillcolor=green]
>         n00000025:port0 -> n0000000d
>         n00000029 [label="{{} | tpg-4 | {<port0> 0}}", shape=Mrecord, 
> style=filled, fillcolor=green]
>         n00000029:port0 -> n00000011
>         n0000002d [label="{{} | tpg-5 | {<port0> 0}}", shape=Mrecord, 
> style=filled, fillcolor=green]
>         n0000002d:port0 -> n00000015
> }
>
>>
>>> --- diff --git a/drivers/staging/media/tegra/host1x-video.h 
>>> b/drivers/staging/media/tegra/host1x-video.h
>>> new file mode 100644
>>> index 000000000000..84d28e6f4362
>>> --- /dev/null
>>> +++ b/drivers/staging/media/tegra/host1x-video.h
>>> @@ -0,0 +1,33 @@
>>> +// SPDX-License-Identifier: GPL-2.0-only
>>> +/*
>>> + * Copyright (C) 2020 NVIDIA CORPORATION.  All rights reserved.
>>> + */
>>> +
>>> +#ifndef HOST1X_VIDEO_H
>>> +#define HOST1X_VIDEO_H 1
>>> +
>>> +#include <linux/host1x.h>
>>> +
>>> +#include <media/media-device.h>
>>> +#include <media/media-entity.h>
>>> +#include <media/v4l2-async.h>
>>> +#include <media/v4l2-ctrls.h>
>>> +#include <media/v4l2-device.h>
>>> +#include <media/v4l2-dev.h>
>>> +#include <media/videobuf2-v4l2.h>
>>> +
>>> +#include "tegra-vi.h"
>>> +#include "csi.h"
>>> +
>>> +struct tegra_camera {
>>> +     struct v4l2_device v4l2_dev;
>>> +     struct media_device media_dev;
>>> +     struct device *dev;
>> You can use cam->media_dev.dev instead of having this pointer.
>>
Will fix in v2
>>> +     struct tegra_vi *vi;
>>> +     struct tegra_csi_device *csi;
>>> +};
>>> +
>>> +
>>> +#define to_tegra_channel(vdev) \
>>> +     container_of(vdev, struct tegra_channel, video)
>> Why not inline instead of define. Inlines has the advantage of 
>> checking types.
Will change in v2
>>
>>> +static int __tegra_channel_try_format(struct tegra_channel *chan,
>>> +                                   struct v4l2_pix_format *pix,
>>> +                                   const struct tegra_video_format 
>>> **vfmt)
>>> +{
>>> +     const struct tegra_video_format *fmt_info;
>>> +     struct v4l2_subdev *subdev;
>>> +     struct v4l2_subdev_format fmt;
>>> +     struct v4l2_subdev_pad_config *pad_cfg;
>>> +
>>> +     subdev = tegra_channel_get_remote_subdev(chan);
>>> +     pad_cfg = v4l2_subdev_alloc_pad_config(subdev);
>>> +     if (!pad_cfg)
>>> +             return -ENOMEM;
>>> +
>>> +     /*
>>> +      * Retrieve format information and select the default format 
>>> if the
>>> +      * requested format isn't supported.
>>> +      */
>>> +     fmt_info = tegra_core_get_format_by_fourcc(chan, 
>>> pix->pixelformat);
>>> +     if (!fmt_info) {
>>> +             pix->pixelformat = chan->format.pixelformat;
>>> +             pix->colorspace = chan->format.colorspace;
>>> +             fmt_info = tegra_core_get_format_by_fourcc(chan,
>>> + pix->pixelformat);
>>> +     }
>>> +
>>> +     /* Change this when start adding interlace format support */
>>> +     pix->field = V4L2_FIELD_NONE;
>>> +     fmt.which = V4L2_SUBDEV_FORMAT_TRY;
>>> +     fmt.pad = 0;
>>> +     v4l2_fill_mbus_format(&fmt.format, pix, fmt_info->code);
>>> +     v4l2_subdev_call(subdev, pad, set_fmt, pad_cfg, &fmt);
>> As fas as I understand, entities formats should be independent, it is 
>> up to link_validate
>> to check formats between entities.
>> The capture shouldn't change the format of the subdevice.
>>
Tegra Built-in TPG on CSI accepts specific TPG sizes and CSI is source 
and VI is sink.

link validation happens only for sink ends of the link.

So with CSI subdev set_fmt sets width/height to default incase if 
width/height is not from one of the supported sizes.

>>
>>> +
>>> +     v4l2_fill_pix_format(pix, &fmt.format);
>>> +     tegra_channel_fmt_align(chan, &fmt_info->bpp, &pix->width, 
>>> &pix->height,
>>> +                             &pix->bytesperline);
>>> +     pix->sizeimage = pix->bytesperline * pix->height;
>>> +
>>> +     if (vfmt)
>>> +             *vfmt = fmt_info;
>>> +
>>> +     v4l2_subdev_free_pad_config(pad_cfg);
>>> +
>>> +     return 0;
>>> +}
>>> +
>>> +static int tegra_channel_try_format(struct file *file, void *fh,
>>> +                                 struct v4l2_format *format)
>>> +{
>>> +     struct v4l2_fh *vfh = file->private_data;
>>> +     struct tegra_channel *chan = to_tegra_channel(vfh->vdev);
>>> +
>>> +     return __tegra_channel_try_format(chan, &format->fmt.pix, NULL);
>>> +}
>>> +
>>> +static int tegra_channel_set_format(struct file *file, void *fh,
>>> +                                 struct v4l2_format *format)
>>> +{
>>> +     struct v4l2_fh *vfh = file->private_data;
>>> +     struct tegra_channel *chan = to_tegra_channel(vfh->vdev);
>>> +     const struct tegra_video_format *info;
>>> +     int ret;
>>> +     struct v4l2_subdev_format fmt;
>>> +     struct v4l2_subdev *subdev;
>>> +     struct v4l2_pix_format *pix = &format->fmt.pix;
>>> +
>>> +     if (vb2_is_busy(&chan->queue))
>>> +             return -EBUSY;
>>> +
>>> +     /* get supported format by try_fmt */
>>> +     ret = __tegra_channel_try_format(chan, pix, &info);
>>> +     if (ret)
>>> +             return ret;
>>> +
>>> +     subdev = tegra_channel_get_remote_subdev(chan);
>>> +
>>> +     fmt.which = V4L2_SUBDEV_FORMAT_ACTIVE;
>>> +     fmt.pad = 0;
>>> +     v4l2_fill_mbus_format(&fmt.format, pix, info->code);
>>> +     v4l2_subdev_call(subdev, pad, set_fmt, NULL, &fmt);
>> same here.
>>
Calling subdev set_fmt here for the same reason as explained above.
>>> +
>>> +     v4l2_fill_pix_format(pix, &fmt.format);
>>> +     chan->format = *pix;
>>> +     chan->fmtinfo = info;
>>> +     tegra_channel_update_format(chan, pix->width,
>>> +                                 pix->height, info->fourcc,
>>> +                                 &info->bpp,
>>> +                                 pix->bytesperline);
>>> +     *pix = chan->format;
>>> +
>>> +     return 0;
>>> +}
>>> +
>>> +static int tegra_channel_enum_input(struct file *file, void *fh,
>>> +                                 struct v4l2_input *inp)
>>> +{
>>> +     /* Currently driver supports internal TPG only */
>>> +     if (inp->index != 0)
>> just
>> if (inp->index)
>>
Will update in v2
>>> +             return -EINVAL;
>>> +
>>> +     inp->type = V4L2_INPUT_TYPE_CAMERA;
>>> +     strscpy(inp->name, "Tegra TPG", sizeof(inp->name));
>>> +
>>> +     return 0;
>>> +}
>>> +static const struct tegra_video_format tegra_default_format = {
>>> +     /* RAW 10 */
>>> +     TEGRA_VF_RAW10,
>>> +     10,
>>> +     MEDIA_BUS_FMT_SRGGB10_1X10,
>>> +     {2, 1},
>>> +     TEGRA_IMAGE_FORMAT_DEF,
>>> +     TEGRA_IMAGE_DT_RAW10,
>>> +     V4L2_PIX_FMT_SRGGB10,
>>> +     "RGRG.. GBGB..",
>> It would be more readable to do:
>>
>> .code = TEGRA_VF_RAW10,
>> .width = 10,
>> .code = MEDIA_BUS_FMT_SRGGB10_1X10,
>>
>> and so on
Will update in v2
>>
>>> +};
>>> +
>>> +/*
>>> + * Helper functions
>>> + */
>>> +
>>> +/**
>>> + * tegra_core_get_default_format - Get default format
>>> + *
>>> + * Return: pointer to the format where the default format needs
>>> + * to be filled in.
>>> + */
>>> +const struct tegra_video_format *tegra_core_get_default_format(void)
>>> +{
>>> +     return &tegra_default_format;
>>> +}
>> This is only used in tegra-channel.c, why not to declare it there as 
>> static?
>>
Will move all video format retrieval helper functions to corresponding 
file as static in v2
>>> + +static struct v4l2_frmsize_discrete tegra_csi_tpg_sizes[] = {
>>> +     {1280, 720},
>>> +     {1920, 1080},
>>> +     {3840, 2160},
>>> +};
>>> +
>>> +/*
>>> + * V4L2 Subdevice Pad Operations
>>> + */
>>> +static int tegra_csi_get_format(struct v4l2_subdev *subdev,
>>> +                             struct v4l2_subdev_pad_config *cfg,
>>> +                             struct v4l2_subdev_format *fmt)
>>> +{
>>> +     struct tegra_csi_channel *csi_chan = to_csi_chan(subdev);
>>> +
>>> +     mutex_lock(&csi_chan->format_lock);
>> Do you need this lock? I think there is already a serialization in 
>> the ioctls in place (to be confirmed).
>>
This is on CSI v4l2 subdevice side during format updates
>>> +     memcpy(fmt, &csi_chan->ports->format,
>>> +            sizeof(struct v4l2_mbus_framefmt));
>> I would prefer just:
>> *fmt = *csi_chan->ports->format;
>>
>> I think it is easier to read IMHO.
>> same in tegra_csi_set_format().
>>
Will fix in v2
>>> + mutex_unlock(&csi_chan->format_lock);
>>> +
>>> +     return 0;
>>> +}
>>> +
>>> +static void tegra_csi_try_mbus_fmt(struct v4l2_subdev *subdev,
>>> +                                struct v4l2_mbus_framefmt *mfmt)
>>> +{
>>> +     struct tegra_csi_channel *csi_chan = to_csi_chan(subdev);
>>> +     struct tegra_csi_device *csi = csi_chan->csi;
>>> +     const struct v4l2_frmsize_discrete *sizes;
>>> +     int i, j;
>> unsigned
>>
Will fix in v2
>>> +
>>> +     for (i = 0; i < ARRAY_SIZE(tegra_csi_tpg_fmts); i++) {
>>> +             struct v4l2_mbus_framefmt *mbus_fmt = 
>>> &tegra_csi_tpg_fmts[i];
>>> +
>>> +             if (mfmt->code == mbus_fmt->code) {
>>> +                     for (j = 0; j < 
>>> ARRAY_SIZE(tegra_csi_tpg_sizes); j++) {
>>> +                             sizes = &tegra_csi_tpg_sizes[j];
>>> +                             if (mfmt->width == sizes->width &&
>>> +                                 mfmt->height == sizes->height) {
>>> +                                     return;
>>> +                             }
>>> +                     }
>>> +             }
>>> +
>>> +             dev_info(csi->dev, "using Tegra default RAW10 video 
>>> format\n");
>>> +     }
>>> +
>>> +     dev_info(csi->dev, "using Tegra default WIDTH X HEIGHT 
>>> (1920x1080)\n");
>>> +     memcpy(mfmt, tegra_csi_tpg_fmts, sizeof(struct 
>>> v4l2_mbus_framefmt));
>>> +}
>>> +
>>> + 
>>

  parent reply	other threads:[~2020-01-29  0:49 UTC|newest]

Thread overview: 53+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-01-28 18:23 [RFC PATCH v1 0/5] Add Tegra driver for video capture Sowjanya Komatineni
2020-01-28 18:23 ` [RFC PATCH v1 1/5] dt-bindings: clock: tegra: Add clk id for CSI TPG clock Sowjanya Komatineni
2020-02-05 19:23   ` Stephen Boyd
2020-01-28 18:23 ` [RFC PATCH v1 2/5] clk: tegra: Add Tegra210 CSI TPG clock gate Sowjanya Komatineni
2020-02-05 19:23   ` Stephen Boyd
2020-01-28 18:23 ` [RFC PATCH v1 3/5] dt-binding: tegra: Add VI and CSI bindings Sowjanya Komatineni
2020-01-28 20:32   ` Helen Koike
2020-01-28 21:04     ` Sowjanya Komatineni
2020-01-28 18:23 ` [RFC PATCH v1 4/5] media: tegra: Add Tegra Video input driver for Tegra210 Sowjanya Komatineni
     [not found]   ` <3cdea635-a9ca-7b9c-3c99-8f489f4d669a@collabora.com>
     [not found]     ` <162488d0-4e74-963a-3366-e4c1f7cf04ca@nvidia.com>
2020-01-29  0:49       ` Sowjanya Komatineni [this message]
2020-01-29  1:05         ` Helen Koike
2020-01-29  2:11           ` Sowjanya Komatineni
2020-01-29  5:59             ` Sowjanya Komatineni
2020-01-29 10:31             ` Helen Koike
2020-01-29 17:49               ` Sowjanya Komatineni
2020-01-29 18:15                 ` Sowjanya Komatineni
2020-01-29 18:46                   ` Helen Koike
2020-01-29 22:40                     ` Sowjanya Komatineni
2020-01-29 18:29                 ` Helen Koike
2020-01-29 18:46                   ` Sowjanya Komatineni
2020-01-29 10:09       ` Thierry Reding
2020-01-29 16:25         ` Sowjanya Komatineni
2020-01-29 11:13   ` Thierry Reding
2020-01-29 17:23     ` Sowjanya Komatineni
2020-01-30 12:20       ` Thierry Reding
2020-01-30 17:02         ` Sowjanya Komatineni
2020-01-29 14:16   ` Hans Verkuil (hansverk)
2020-01-29 17:27     ` Sowjanya Komatineni
2020-01-30 14:45   ` Hans Verkuil
2020-02-05 19:26   ` Stephen Boyd
2020-02-05 19:54     ` Sowjanya Komatineni
2020-01-28 18:23 ` [RFC PATCH v1 5/5] arm64: tegra: Add Tegra VI CSI suppport in device tree Sowjanya Komatineni
2020-01-29  9:46   ` Thierry Reding
2020-01-29 16:22     ` Sowjanya Komatineni
2020-01-30 12:36       ` Thierry Reding
2020-01-30 17:18         ` Sowjanya Komatineni
2020-01-30 17:58           ` Thierry Reding
2020-01-30 18:58             ` Sowjanya Komatineni
2020-01-30 20:18               ` Sowjanya Komatineni
2020-01-31  2:57                 ` Sowjanya Komatineni
2020-01-30 14:41 ` [RFC PATCH v1 0/5] Add Tegra driver for video capture Hans Verkuil
2020-01-30 15:42   ` Thierry Reding
2020-01-31 14:29     ` Hans Verkuil
2020-01-31 17:03       ` Thierry Reding
2020-01-31 17:37         ` Hans Verkuil
2020-01-31 20:31           ` Thierry Reding
2020-02-04  9:50         ` Hans Verkuil
2020-01-30 17:20   ` Sowjanya Komatineni
2020-02-04 12:53 ` Hans Verkuil
2020-02-04 16:42   ` Sowjanya Komatineni
2020-02-04 17:22     ` Hans Verkuil
2020-02-04 19:02       ` Sowjanya Komatineni
2020-02-05  7:57         ` Hans Verkuil

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=017ca95e-7dd3-2d04-8d84-9047ac4e548b@nvidia.com \
    --to=skomatineni@nvidia.com \
    --cc=devicetree@vger.kernel.org \
    --cc=frankc@nvidia.com \
    --cc=helen.koike@collabora.com \
    --cc=hverkuil@xs4all.nl \
    --cc=jonathanh@nvidia.com \
    --cc=linux-clk@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-media@vger.kernel.org \
    --cc=linux-tegra@vger.kernel.org \
    --cc=thierry.reding@gmail.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).