| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <YSaVbS59js1xVFKC@pendragon.ideasonboard.com>
Date: Wed, 25 Aug 2021 22:09:33 +0300
From: Laurent Pinchart <laurent.pinchart@...asonboard.com>
To: Pedro Terra Delboni <pedro@...raco.de>
Cc: Helen Koike <helen.koike@...labora.com>,
Hans Verkuil <hverkuil-cisco@...all.nl>,
Dafna Hirschfeld <dafna.hirschfeld@...labora.com>,
Mauro Carvalho Chehab <mchehab@...nel.org>,
Shuah Khan <skhan@...uxfoundation.org>,
linux-media@...r.kernel.org, linux-kernel@...r.kernel.org,
lkcamp@...ts.libreplanetbr.org,
Gabriela Bittencourt <gabrielabittencourt00@...il.com>,
Gabriel Francisco Mandaji <gfmandaji@...il.com>
Subject: Re: [PATCH] [PATCH v5] media: vimc: Enable set resolution at the
scaler src pad
Hi Pedro,
On Wed, Aug 25, 2021 at 03:07:18PM -0300, Pedro Terra Delboni wrote:
> Hi Laurent,
>
> Thanks a lot for the reply and feedback.
> I have made the proposed changes and would like to verify some things
> before sending the next version.
>
> On Thu, Aug 19, 2021 at 7:27 PM Laurent Pinchart wrote:
> > On Sun, Aug 08, 2021 at 11:32:02PM -0300, Pedro Terra wrote:
> > > Modify the scaler subdevice to accept setting the resolution of the source
> > > pad (previously the source resolution would always be 3 times the sink for
> > > both dimensions). Now any resolution can be set at src (even smaller ones)
> > > and the sink video will be scaled to match it.
> > >
> > > Test example: With the vimc module up (using the default vimc topology)
> > > media-ctl -d /dev/media0 -V '"Sensor A":0[fmt:SBGGR8_1X8/640x480]'
> > > media-ctl -d /dev/media0 -V '"Debayer A":0[fmt:SBGGR8_1X8/640x480]'
> > > media-ctl -d /dev/media0 -V '"Scaler":0[fmt:RGB888_1X24/640x480]'
> > > media-ctl -d /dev/media0 -V '"Scaler":0[crop:(100,50)/400x150]'
> > > media-ctl -d /dev/media0 -V '"Scaler":1[fmt:RGB888_1X24/300x700]'
> > > v4l2-ctl -d /dev/video2 -v width=300,height=700
> > > v4l2-ctl -d /dev/video0 -v pixelformat=BA81
> >
> > Did you mean video2 here ? Not the exact device matters much as it
> > depends on the video devices present in the system, but being consistent
> > is best. But even better, I'd write
> >
> > v4l2-ctl -z platform:vimc -d "RGB/YUV Capture" ...
> >
> > to avoid depending on a particular device number. Same for the line
> > below. Similarly, the '-d /dev/media0' argument to media-ctl could be
> > replaced with '-d platform:vimc'.
> >
> > > v4l2-ctl --stream-mmap --stream-count=10 -d /dev/video2 \
> > > --stream-to=test.raw
> > >
> > > The result will be a cropped stream that can be checked with the command
> > > ffplay -loglevel warning -v info -f rawvideo -pixel_format rgb24 \
> > > -video_size "300x700" test.raw
> > >
> > > Co-developed-by: Gabriela Bittencourt <gabrielabittencourt00@...il.com>
> > > Signed-off-by: Gabriela Bittencourt <gabrielabittencourt00@...il.com>
> > > Co-developed-by: Gabriel Francisco Mandaji <gfmandaji@...il.com>
> > > Signed-off-by: Gabriel Francisco Mandaji <gfmandaji@...il.com>
> > > Signed-off-by: Pedro "pirate" Terra <pirate@...raco.de>
> > >
> > > ---
> > >
> > > Changes in V5:
> > > * Fixed code aliment mistake
> > > * Renamed some variables to make the code more readable
> > > * Propagate sink pad formatting to src resetting the 1:1 scaling ratio.
> > > (the crop is also reset when this is done).
> > >
> > > Changes in V4:
> > > * Rebased with media/master
> > > * Scaling is now compatible with crop
> > > * Updated test example at the commit message
> > > * Add vimc prefix to the pad enumeration
> > >
> > > Changes in V3:
> > > * Corrections suggested by Hans:
> > > - Default scaling factor is now 1 (we removed the define and
> > > set the source format equals the sink).
> > > - Removed SCA_COUNT (enum that represents the number of pads)
> > > as there always 2
> > > - Swapped the per byte pixel copy to memcpy.
> > > * Corrections suggested by Dafna:
> > > - Removed from the documentation the old scaler parameter which
> > > isn't necessary anymore.
> > > * Added a thank you note at the end of the email
> > >
> > > Changes in V2:
> > > * Patch was not sent to media list mail for some reason (even though it
> > > was on the Cc list), trying again.
> > > * Updating documentation.
> > >
> > > Running
> > > /usr/local/bin/v4l2-compliance -m /dev/media0
> > > Gave the following result:
> > > v4l2-compliance SHA: c86aab9cc7f1 2021-07-28 11:52:45
> > > Grand Total for vimc device /dev/media0: 473, Succeeded: 473, Failed: 0, Warnings: 0
> > > ---
> > > Documentation/admin-guide/media/vimc.rst | 18 +-
> > > drivers/media/test-drivers/vimc/vimc-scaler.c | 287 +++++++-----------
> > > 2 files changed, 115 insertions(+), 190 deletions(-)
> >
> > I like the diffstat :-)
> >
> > > diff --git a/Documentation/admin-guide/media/vimc.rst b/Documentation/admin-guide/media/vimc.rst
> > > index 211cc8972410..b6a123e22544 100644
> > > --- a/Documentation/admin-guide/media/vimc.rst
> > > +++ b/Documentation/admin-guide/media/vimc.rst
> > > @@ -65,6 +65,11 @@ vimc-scaler:
> > > 1920x1440 image. (this value can be configured, see at
> > > `Module options`_).
> > > Exposes:
> > > + Re-size the image to meet the source pad resolution. E.g.: if the sync pad
> > > +is configured to 360x480 and the source to 1280x720, the image will be stretched
> > > +to fit the source resolution. Works for any resolution within the vimc
> > > +limitations (even shrinking the image if necessary).
> > > +Exposes:
> >
> > This doesn't look right. Were you trying to replace the existing text
> > instead ?
> >
> > >
> > > * 1 Pad sink
> > > * 1 Pad source
> > > @@ -75,16 +80,3 @@ vimc-capture:
> > >
> > > * 1 Pad sink
> > > * 1 Pad source
> > > -
> > > -
> > > -Module options
> > > ---------------
> > > -
> > > -Vimc has a module parameter to configure the driver.
> > > -
> > > -* ``sca_mult=<unsigned int>``
> > > -
> > > - Image size multiplier factor to be used to multiply both width and
> > > - height, so the image size will be ``sca_mult^2`` bigger than the
> > > - original one. Currently, only supports scaling up (the default value
> > > - is 3).
> > > diff --git a/drivers/media/test-drivers/vimc/vimc-scaler.c b/drivers/media/test-drivers/vimc/vimc-scaler.c
> > > index 06880dd0b6ac..745316a50459 100644
> > > --- a/drivers/media/test-drivers/vimc/vimc-scaler.c
> > > +++ b/drivers/media/test-drivers/vimc/vimc-scaler.c
> > > @@ -6,6 +6,7 @@
> > > */
> > >
> > > #include <linux/moduleparam.h>
> > > +#include <linux/string.h>
> > > #include <linux/vmalloc.h>
> > > #include <linux/v4l2-mediabus.h>
> > > #include <media/v4l2-rect.h>
> > > @@ -13,11 +14,11 @@
> > >
> > > #include "vimc-common.h"
> > >
> > > -static unsigned int sca_mult = 3;
> > > -module_param(sca_mult, uint, 0000);
> > > -MODULE_PARM_DESC(sca_mult, " the image size multiplier");
> > > -
> > > -#define MAX_ZOOM 8
> > > +/* Pad identifier */
> > > +enum vic_sca_pad {
> > > + VIMC_SCA_SINK = 0,
> > > + VIMC_SCA_SRC = 1,
> > > +};
> > >
> > > #define VIMC_SCA_FMT_WIDTH_DEFAULT 640
> > > #define VIMC_SCA_FMT_HEIGHT_DEFAULT 480
> > > @@ -25,14 +26,11 @@ MODULE_PARM_DESC(sca_mult, " the image size multiplier");
> > > struct vimc_sca_device {
> > > struct vimc_ent_device ved;
> > > struct v4l2_subdev sd;
> > > - /* NOTE: the source fmt is the same as the sink
> > > - * with the width and hight multiplied by mult
> > > - */
> > > - struct v4l2_mbus_framefmt sink_fmt;
> > > struct v4l2_rect crop_rect;
> > > + /* Frame format for both sink and src pad */
> > > + struct v4l2_mbus_framefmt fmt[2];
> > > /* Values calculated when the stream starts */
> > > u8 *src_frame;
> > > - unsigned int src_line_size;
> > > unsigned int bpp;
> > > struct media_pad pads[2];
> > > };
> > > @@ -72,17 +70,6 @@ vimc_sca_get_crop_bound_sink(const struct v4l2_mbus_framefmt *sink_fmt)
> > > return r;
> > > }
> > >
> > > -static void vimc_sca_adjust_sink_crop(struct v4l2_rect *r,
> > > - const struct v4l2_mbus_framefmt *sink_fmt)
> > > -{
> > > - const struct v4l2_rect sink_rect =
> > > - vimc_sca_get_crop_bound_sink(sink_fmt);
> > > -
> > > - /* Disallow rectangles smaller than the minimal one. */
> > > - v4l2_rect_set_min_size(r, &crop_rect_min);
> > > - v4l2_rect_map_inside(r, &sink_rect);
> > > -}
> > > -
> > > static int vimc_sca_init_cfg(struct v4l2_subdev *sd,
> > > struct v4l2_subdev_state *sd_state)
> > > {
> > > @@ -90,17 +77,15 @@ static int vimc_sca_init_cfg(struct v4l2_subdev *sd,
> > > struct v4l2_rect *r;
> > > unsigned int i;
> > >
> > > - mf = v4l2_subdev_get_try_format(sd, sd_state, 0);
> > > + mf = v4l2_subdev_get_try_format(sd, sd_state, VIMC_SCA_SINK);
> > > *mf = sink_fmt_default;
> > >
> > > - r = v4l2_subdev_get_try_crop(sd, sd_state, 0);
> > > + r = v4l2_subdev_get_try_crop(sd, sd_state, VIMC_SCA_SINK);
> > > *r = crop_rect_default;
> > >
> > > for (i = 1; i < sd->entity.num_pads; i++) {
> > > mf = v4l2_subdev_get_try_format(sd, sd_state, i);
> > > *mf = sink_fmt_default;
> > > - mf->width = mf->width * sca_mult;
> > > - mf->height = mf->height * sca_mult;
> > > }
> >
> > I think you can simplify this as
> >
> > for (i = 0; i < sd->entity.num_pads; i++) {
> > mf = v4l2_subdev_get_try_format(sd, sd_state, i);
> > *mf = sink_fmt_default;
> > }
> >
> > r = v4l2_subdev_get_try_crop(sd, sd_state, VIMC_SCA_SINK);
> > *r = crop_rect_default;
> >
> > By the way, sink_fmt_default should be renamed to fmt_default as it's
> > not about the sink pad only anymore.
> >
> > >
> > > return 0;
> > > @@ -144,13 +129,8 @@ static int vimc_sca_enum_frame_size(struct v4l2_subdev *sd,
> > > fse->min_width = VIMC_FRAME_MIN_WIDTH;
> > > fse->min_height = VIMC_FRAME_MIN_HEIGHT;
> > >
> > > - if (VIMC_IS_SINK(fse->pad)) {
> > > - fse->max_width = VIMC_FRAME_MAX_WIDTH;
> > > - fse->max_height = VIMC_FRAME_MAX_HEIGHT;
> > > - } else {
> > > - fse->max_width = VIMC_FRAME_MAX_WIDTH * MAX_ZOOM;
> > > - fse->max_height = VIMC_FRAME_MAX_HEIGHT * MAX_ZOOM;
> > > - }
> > > + fse->max_width = VIMC_FRAME_MAX_WIDTH;
> > > + fse->max_height = VIMC_FRAME_MAX_HEIGHT;
> > >
> > > return 0;
> > > }
> > > @@ -160,94 +140,82 @@ static int vimc_sca_get_fmt(struct v4l2_subdev *sd,
> > > struct v4l2_subdev_format *format)
> > > {
> > > struct vimc_sca_device *vsca = v4l2_get_subdevdata(sd);
> > > - struct v4l2_rect *crop_rect;
> > > -
> > > - /* Get the current sink format */
> > > - if (format->which == V4L2_SUBDEV_FORMAT_TRY) {
> > > - format->format = *v4l2_subdev_get_try_format(sd, sd_state, 0);
> > > - crop_rect = v4l2_subdev_get_try_crop(sd, sd_state, 0);
> > > - } else {
> > > - format->format = vsca->sink_fmt;
> > > - crop_rect = &vsca->crop_rect;
> > > - }
> > >
> > > - /* Scale the frame size for the source pad */
> > > - if (VIMC_IS_SRC(format->pad)) {
> > > - format->format.width = crop_rect->width * sca_mult;
> > > - format->format.height = crop_rect->height * sca_mult;
> > > - }
> > > + if (format->which == V4L2_SUBDEV_FORMAT_TRY)
> > > + format->format = *v4l2_subdev_get_try_format(sd, sd_state,
> > > + format->pad);
> > > + else
> > > + format->format = vsca->fmt[format->pad];
> > >
> > > return 0;
> > > }
> > >
> > > -static void vimc_sca_adjust_sink_fmt(struct v4l2_mbus_framefmt *fmt)
> > > +static void vimc_sca_adjust_fmt(struct v4l2_mbus_framefmt *fmt[], __u32 pad)
> >
> > s/__u32/u32/ (the former is used for userspace-facing code in
> > include/uapi/ to avoid collisions, inside the kernel we use the latter).
> >
> > > {
> > > - const struct vimc_pix_map *vpix;
> > > + unsigned int src_width, src_height;
> > >
> > > - /* Only accept code in the pix map table in non bayer format */
> > > - vpix = vimc_pix_map_by_code(fmt->code);
> > > - if (!vpix || vpix->bayer)
> > > - fmt->code = sink_fmt_default.code;
> > > + if (pad == VIMC_SCA_SINK) {
> > > + const struct vimc_pix_map *vpix;
> > >
> > > - fmt->width = clamp_t(u32, fmt->width, VIMC_FRAME_MIN_WIDTH,
> > > - VIMC_FRAME_MAX_WIDTH) & ~1;
> > > - fmt->height = clamp_t(u32, fmt->height, VIMC_FRAME_MIN_HEIGHT,
> > > - VIMC_FRAME_MAX_HEIGHT) & ~1;
> > > + /* Only accept code in the pix map table in non bayer format */
> > > + vpix = vimc_pix_map_by_code(fmt[VIMC_SCA_SINK]->code);
> > > + if (!vpix || vpix->bayer)
> > > + fmt[VIMC_SCA_SINK]->code = sink_fmt_default.code;
> > > + if (fmt[VIMC_SCA_SINK]->field == V4L2_FIELD_ANY)
> > > + fmt[VIMC_SCA_SINK]->field = sink_fmt_default.field;
> > > +
> > > + vimc_colorimetry_clamp(fmt[VIMC_SCA_SINK]);
> > > + }
> > >
> > > - if (fmt->field == V4L2_FIELD_ANY)
> > > - fmt->field = sink_fmt_default.field;
> > > + fmt[pad]->width = clamp_t(u32, fmt[pad]->width, VIMC_FRAME_MIN_WIDTH,
> > > + VIMC_FRAME_MAX_WIDTH) & ~1;
> > > + fmt[pad]->height = clamp_t(u32, fmt[pad]->height, VIMC_FRAME_MIN_HEIGHT,
> > > + VIMC_FRAME_MAX_HEIGHT) & ~1;
> > >
> > > - vimc_colorimetry_clamp(fmt);
> > > + /* Resets the src pad to match the sink, then restore dimensions if
> > > + * the pad formatted was the src one
> > > + */
> > > + src_width = fmt[pad]->width;
> > > + src_height = fmt[pad]->height;
> > > + *fmt[VIMC_SCA_SRC] = *fmt[VIMC_SCA_SINK];
> > > + fmt[VIMC_SCA_SRC]->width = src_width;
> > > + fmt[VIMC_SCA_SRC]->height = src_height;
> >
> > This is hard to follow, I wonder if we could do better. Please see below
> > for a proposal.
>
> Indeed, it was. Thanks a lot for the proposal.
>
> > > }
> > >
> > > static int vimc_sca_set_fmt(struct v4l2_subdev *sd,
> > > struct v4l2_subdev_state *sd_state,
> > > - struct v4l2_subdev_format *fmt)
> > > + struct v4l2_subdev_format *format)
> > > {
> > > struct vimc_sca_device *vsca = v4l2_get_subdevdata(sd);
> > > - struct v4l2_mbus_framefmt *sink_fmt;
> > > + struct v4l2_mbus_framefmt *fmt[2];
> > > struct v4l2_rect *crop_rect;
> > >
> > > - if (fmt->which == V4L2_SUBDEV_FORMAT_ACTIVE) {
> > > + if (format->which == V4L2_SUBDEV_FORMAT_ACTIVE) {
> > > /* Do not change the format while stream is on */
> > > if (vsca->src_frame)
> > > return -EBUSY;
> > >
> > > - sink_fmt = &vsca->sink_fmt;
> > > + fmt[VIMC_SCA_SINK] = &vsca->fmt[VIMC_SCA_SINK];
> > > + fmt[VIMC_SCA_SRC] = &vsca->fmt[VIMC_SCA_SRC];
> > > crop_rect = &vsca->crop_rect;
> > > } else {
> > > - sink_fmt = v4l2_subdev_get_try_format(sd, sd_state, 0);
> > > - crop_rect = v4l2_subdev_get_try_crop(sd, sd_state, 0);
> > > + fmt[VIMC_SCA_SINK] = v4l2_subdev_get_try_format(sd, sd_state,
> > > + VIMC_SCA_SINK);
> > > + fmt[VIMC_SCA_SRC] = v4l2_subdev_get_try_format(sd, sd_state,
> > > + VIMC_SCA_SRC);
> > > + crop_rect = v4l2_subdev_get_try_crop(sd, sd_state, VIMC_SCA_SINK);
> > > }
> > >
> > > - /*
> > > - * Do not change the format of the source pad,
> > > - * it is propagated from the sink
> > > - */
> > > - if (VIMC_IS_SRC(fmt->pad)) {
> > > - fmt->format = *sink_fmt;
> > > - fmt->format.width = crop_rect->width * sca_mult;
> > > - fmt->format.height = crop_rect->height * sca_mult;
> > > - } else {
> > > - /* Set the new format in the sink pad */
> > > - vimc_sca_adjust_sink_fmt(&fmt->format);
> > > -
> > > - dev_dbg(vsca->ved.dev, "%s: sink format update: "
> > > - "old:%dx%d (0x%x, %d, %d, %d, %d) "
> > > - "new:%dx%d (0x%x, %d, %d, %d, %d)\n", vsca->sd.name,
> > > - /* old */
> > > - sink_fmt->width, sink_fmt->height, sink_fmt->code,
> > > - sink_fmt->colorspace, sink_fmt->quantization,
> > > - sink_fmt->xfer_func, sink_fmt->ycbcr_enc,
> > > - /* new */
> > > - fmt->format.width, fmt->format.height, fmt->format.code,
> > > - fmt->format.colorspace, fmt->format.quantization,
> > > - fmt->format.xfer_func, fmt->format.ycbcr_enc);
> > > -
> > > - *sink_fmt = fmt->format;
> > > + *fmt[format->pad] = format->format;
> > > + vimc_sca_adjust_fmt(fmt, format->pad);
> > > + format->format = *fmt[format->pad];
> > >
> > > - /* Do the crop, but respect the current bounds */
> > > - vimc_sca_adjust_sink_crop(crop_rect, sink_fmt);
> > > + if (format->pad == VIMC_SCA_SINK) {
> > > + /* Reset the crop to match the sink pad */
> > > + crop_rect->width = fmt[VIMC_SCA_SINK]->width;
> > > + crop_rect->height = fmt[VIMC_SCA_SINK]->height;
> > > + crop_rect->top = 0;
> > > + crop_rect->left = 0;
> > > }
> > >
> > > return 0;
> >
> > Here's the proposal:
> >
> > struct v4l2_mbus_framefmt *
> > vimc_sca_pad_format(struct vimc_sca_device *vsca,
> > struct v4l2_subdev_state *sd_state, u32 pad,
> > enum v4l2_subdev_format_whence which)
> > {
> > if (which == V4L2_SUBDEV_FORMAT_TRY)
> > return v4l2_subdev_get_try_format(&vsca->sd, sd_state, pad);
> > else
> > return &vsca->fmt[pad];
> > }
> >
> > struct v4l2_rect *
> > vimc_sca_pad_crop(struct vimc_sca_device *vsca,
> > struct v4l2_subdev_state *sd_state,
> > enum v4l2_subdev_format_whence which)
> > {
> > if (which == V4L2_SUBDEV_FORMAT_TRY)
> > return v4l2_subdev_get_try_crop(&vsca->sd, sd_state,
> > VIMC_SCA_SINK);
> > else
> > return &vsca->crop_rect;
> > }
> >
> > static int vimc_sca_set_fmt(struct v4l2_subdev *sd,
> > struct v4l2_subdev_state *sd_state,
> > struct v4l2_subdev_format *format)
> > {
> > struct vimc_sca_device *vsca = v4l2_get_subdevdata(sd);
> > struct v4l2_mbus_framefmt *fmt;
> >
> > /* Do not change the active format while stream is on. */
> > if (format->which == V4L2_SUBDEV_FORMAT_ACTIVE) {
> > if (vsca->src_frame)
> > return -EBUSY;
> > }
> >
> > fmt = vimc_sca_pad_format(vsca, sd_state, format->pad, format->which);
> >
> > /*
> > * The media bus code and colorspace can only be changed on the sink
> > * pad, the source pad only follows.
> > */
> > if (format->pad == VIMC_SCA_SINK) {
> > const struct vimc_pix_map *vpix;
> >
> > /* Only accept code in the pix map table in non bayer format. */
> > vpix = vimc_pix_map_by_code(format->format.code);
> > if (vpix && !vpix->bayer)
> > fmt->code = format->format.code;
> > else
> > fmt->code = sink_fmt_default.code;
> >
> > /* Clamp the colorspace to valid values. */
> > fmt->colorspace = format->format.colorspace;
> > fmt->ycbcr_enc = format->format.ycbcr_enc;
> > fmt->quantization = format->format.quantization;
> > fmt->xfer_func = format->format.xfer_func;
> > vimc_colorimetry_clamp(fmt);
>
> I was instructed that copying the whole structure would be good in
> order to propagate any eventual
> extensions in it, and then change only the necessary fields like:
>
> *fmt = format->format;
> vimc_colorimetry_clamp(fmt);
>
> I do see that this would make redundant writes at fmt->width and
> fmt->height, as they are assigned right in the following lines.
>
> Should I keep the assignments of the structure fields instead of the
> whole struct?
It's a good point. The argument that copying the whole structure is more
future-proof makes sense at first sight, but I think it's actually not
true. For one thing, we don't know what those new parameters will be, so
there's no way to know if they should be copied or not. It may be that a
particular parameter would require the driver to be updated to be
properly taken into account, in which case copying the reserved fields
blindly today will give a false impression to userspace tomorrow that a
new parameter has taken effect when it hasn't. It's very common for
extensions to be designed in a way that makes 0 being treated as
preserving the existing behaviour, and userspace can rely on the driver
returning a zero value when provided a non-zero value to check if a
feature is supported. I thus think we should only copy parameters that
we explicitly support.
> > }
> >
> > /* Clamp and align the width and height. */
> > fmt->width = clamp_t(u32, format->format.width, VIMC_FRAME_MIN_WIDTH,
> > VIMC_FRAME_MAX_WIDTH) & ~1;
> > fmt->height = clamp_t(u32, format->format.height, VIMC_FRAME_MIN_HEIGHT,
> > VIMC_FRAME_MAX_HEIGHT) & ~1;
> >
> > /*
> > * Propagate the sink pad format to the crop rectangle and the source
> > * pad.
> > */
> > if (format->pad == VIMC_SCA_SINK) {
> > struct v4l2_mbus_framefmt *src_fmt;
> > struct v4l2_rect *crop;
> >
> > crop = vimc_sca_pad_crop(vsca, sd_state, format->which);
> > crop->width = fmt->width;
> > crop->height = fmt->height;
> > crop->top = 0;
> > crop->left = 0;
> >
> > src_fmt = vimc_sca_pad_format(vsca, sd_state, VIMC_SCA_SRC,
> > format->which);
> > *src_fmt = *fmt;
> > }
> >
> > format->format = *fmt;
> >
> > return 0;
> > }
> >
> > And vimc_sca_get_fmt() could be simplified as
> >
> > static int vimc_sca_get_fmt(struct v4l2_subdev *sd,
> > struct v4l2_subdev_state *sd_state,
> > struct v4l2_subdev_format *format)
> > {
> > struct vimc_sca_device *vsca = v4l2_get_subdevdata(sd);
> >
> > format->format = *vimc_sca_pad_format(vsca, sd_state, format->pad,
> > format->which);
> > return 0;
> > }
> >
> > > @@ -265,11 +233,13 @@ static int vimc_sca_get_selection(struct v4l2_subdev *sd,
> > > return -EINVAL;
> > >
> > > if (sel->which == V4L2_SUBDEV_FORMAT_ACTIVE) {
> > > - sink_fmt = &vsca->sink_fmt;
> > > + sink_fmt = &vsca->fmt[VIMC_SCA_SINK];
> > > crop_rect = &vsca->crop_rect;
> > > } else {
> > > - sink_fmt = v4l2_subdev_get_try_format(sd, sd_state, 0);
> > > - crop_rect = v4l2_subdev_get_try_crop(sd, sd_state, 0);
> > > + sink_fmt = v4l2_subdev_get_try_format(sd, sd_state,
> > > + VIMC_SCA_SINK);
> > > + crop_rect = v4l2_subdev_get_try_crop(sd, sd_state,
> > > + VIMC_SCA_SINK);
> > > }
> > >
> > > switch (sel->target) {
> >
> > And then, with the helper functions introduced above,
> >
> > static int vimc_sca_get_selection(struct v4l2_subdev *sd,
> > struct v4l2_subdev_state *sd_state,
> > struct v4l2_subdev_selection *sel)
> > {
> > struct vimc_sca_device *vsca = v4l2_get_subdevdata(sd);
> > struct v4l2_mbus_framefmt *sink_fmt;
> >
> > if (VIMC_IS_SRC(sel->pad))
> > return -EINVAL;
> >
> > switch (sel->target) {
> > case V4L2_SEL_TGT_CROP:
> > sel->r = *vimc_sca_pad_crop(vsca, sd_state, sel->which);
> > break;
> > case V4L2_SEL_TGT_CROP_BOUNDS:
> > sink_fmt = vimc_sca_pad_format(vsca, sd_state, VIMC_SCA_SINK,
> > sel->which);
> > sel->r = vimc_sca_get_crop_bound_sink(sink_fmt);
> > break;
> > default:
> > return -EINVAL;
> > }
> >
> > return 0;
> > }
> >
> > > @@ -286,6 +256,17 @@ static int vimc_sca_get_selection(struct v4l2_subdev *sd,
> > > return 0;
> > > }
> > >
> > > +static void vimc_sca_adjust_sink_crop(struct v4l2_rect *r,
> > > + const struct v4l2_mbus_framefmt *sink_fmt)
> > > +{
> > > + const struct v4l2_rect sink_rect =
> > > + vimc_sca_get_crop_bound_sink(sink_fmt);
> > > +
> > > + /* Disallow rectangles smaller than the minimal one. */
> > > + v4l2_rect_set_min_size(r, &crop_rect_min);
> > > + v4l2_rect_map_inside(r, &sink_rect);
> > > +}
> > > +
> > > static int vimc_sca_set_selection(struct v4l2_subdev *sd,
> > > struct v4l2_subdev_state *sd_state,
> > > struct v4l2_subdev_selection *sel)
> > > @@ -303,10 +284,12 @@ static int vimc_sca_set_selection(struct v4l2_subdev *sd,
> > > return -EBUSY;
> >
> > The check here should be removed, as we should support digital zoom.
> > This can come in a separate patch though.
>
> Would it be troublesome if I removed the check in this patch already?
Removing this check will require adding locking, I'd rather split that
to a separate patch.
> > >
> > > crop_rect = &vsca->crop_rect;
> > > - sink_fmt = &vsca->sink_fmt;
> > > + sink_fmt = &vsca->fmt[VIMC_SCA_SINK];
> > > } else {
> > > - crop_rect = v4l2_subdev_get_try_crop(sd, sd_state, 0);
> > > - sink_fmt = v4l2_subdev_get_try_format(sd, sd_state, 0);
> > > + crop_rect = v4l2_subdev_get_try_crop(sd, sd_state,
> > > + VIMC_SCA_SINK);
> > > + sink_fmt = v4l2_subdev_get_try_format(sd, sd_state,
> > > + VIMC_SCA_SINK);
> > > }
> >
> > This can also be simplified as
> >
> > sink_fmt = vimc_sca_pad_format(vsca, sd_state, VIMC_SCA_SINK,
> > sel->which);
> > crop_rect = vimc_sca_pad_crop(vsca, sd_state, sel->which);
> >
> > and maybe the switch/case below could be removed, replaced with a
> >
> > if (self->target != V4L2_SEL_TGT_CROP)
> > return -EINVAL;
> >
> > at the beginning of the function, just keeping
> >
> > /* Do the crop, but respect the current bounds */
> > vimc_sca_adjust_sink_crop(&sel->r, sink_fmt);
> > *crop_rect = sel->r;
> >
> > below.
>
> I have joined this check with the VIMC_IS_SRC(sel->pad) one, as they
> return the same error, like so:
>
> if (VIMC_IS_SRC(sel->pad) | | (sel->target != V4L2_SEL_TGT_CROP))
> return -EINVAL;
>
> Is that alright?
Sure, if you drop the space in the middle of | | :-) You can also drop
the parentheses around the second condition, they're not needed.
> > >
> > > switch (sel->target) {
> > > @@ -344,16 +327,12 @@ static int vimc_sca_s_stream(struct v4l2_subdev *sd, int enable)
> > > return 0;
> > >
> > > /* Save the bytes per pixel of the sink */
> > > - vpix = vimc_pix_map_by_code(vsca->sink_fmt.code);
> > > + vpix = vimc_pix_map_by_code(vsca->fmt[VIMC_SCA_SINK].code);
> > > vsca->bpp = vpix->bpp;
> > >
> > > - /* Calculate the width in bytes of the src frame */
> > > - vsca->src_line_size = vsca->crop_rect.width *
> > > - sca_mult * vsca->bpp;
> > > -
> > > /* Calculate the frame size of the source pad */
> > > - frame_size = vsca->src_line_size * vsca->crop_rect.height *
> > > - sca_mult;
> > > + frame_size = vsca->fmt[VIMC_SCA_SRC].width
> > > + * vsca->fmt[VIMC_SCA_SRC].height * vsca->bpp;
> >
> > Could you align * under = ?
> >
> > >
> > > /* Allocate the frame buffer. Use vmalloc to be able to
> > > * allocate a large amount of memory
> > > @@ -382,77 +361,30 @@ static const struct v4l2_subdev_ops vimc_sca_ops = {
> > > .video = &vimc_sca_video_ops,
> > > };
> > >
> > > -static void vimc_sca_fill_pix(u8 *const ptr,
> > > - const u8 *const pixel,
> > > - const unsigned int bpp)
> > > -{
> > > - unsigned int i;
> > > -
> > > - /* copy the pixel to the pointer */
> > > - for (i = 0; i < bpp; i++)
> > > - ptr[i] = pixel[i];
> > > -}
> > > -
> > > -static void vimc_sca_scale_pix(const struct vimc_sca_device *const vsca,
> > > - unsigned int lin, unsigned int col,
> > > - const u8 *const sink_frame)
> > > -{
> > > - const struct v4l2_rect crop_rect = vsca->crop_rect;
> > > - unsigned int i, j, index;
> > > - const u8 *pixel;
> > > -
> > > - /* Point to the pixel value in position (lin, col) in the sink frame */
> > > - index = VIMC_FRAME_INDEX(lin, col,
> > > - vsca->sink_fmt.width,
> > > - vsca->bpp);
> > > - pixel = &sink_frame[index];
> > > -
> > > - dev_dbg(vsca->ved.dev,
> > > - "sca: %s: --- scale_pix sink pos %dx%d, index %d ---\n",
> > > - vsca->sd.name, lin, col, index);
> > > -
> > > - /* point to the place we are going to put the first pixel
> > > - * in the scaled src frame
> > > - */
> > > - lin -= crop_rect.top;
> > > - col -= crop_rect.left;
> > > - index = VIMC_FRAME_INDEX(lin * sca_mult, col * sca_mult,
> > > - crop_rect.width * sca_mult, vsca->bpp);
> > > -
> > > - dev_dbg(vsca->ved.dev, "sca: %s: scale_pix src pos %dx%d, index %d\n",
> > > - vsca->sd.name, lin * sca_mult, col * sca_mult, index);
> > > -
> > > - /* Repeat this pixel mult times */
> > > - for (i = 0; i < sca_mult; i++) {
> > > - /* Iterate through each beginning of a
> > > - * pixel repetition in a line
> > > - */
> > > - for (j = 0; j < sca_mult * vsca->bpp; j += vsca->bpp) {
> > > - dev_dbg(vsca->ved.dev,
> > > - "sca: %s: sca: scale_pix src pos %d\n",
> > > - vsca->sd.name, index + j);
> > > -
> > > - /* copy the pixel to the position index + j */
> > > - vimc_sca_fill_pix(&vsca->src_frame[index + j],
> > > - pixel, vsca->bpp);
> > > - }
> > > -
> > > - /* move the index to the next line */
> > > - index += vsca->src_line_size;
> > > - }
> > > -}
> > > -
> > > static void vimc_sca_fill_src_frame(const struct vimc_sca_device *const vsca,
> > > const u8 *const sink_frame)
> > > {
> > > - const struct v4l2_rect r = vsca->crop_rect;
> > > - unsigned int i, j;
> > > -
> > > - /* Scale each pixel from the original sink frame */
> > > - /* TODO: implement scale down, only scale up is supported for now */
> > > - for (i = r.top; i < r.top + r.height; i++)
> > > - for (j = r.left; j < r.left + r.width; j++)
> > > - vimc_sca_scale_pix(vsca, i, j, sink_frame);
> > > + const struct v4l2_mbus_framefmt *src_fmt = &vsca->fmt[VIMC_SCA_SRC];
> > > + const struct v4l2_rect *r = &vsca->crop_rect;
> > > + unsigned int snk_width = vsca->fmt[VIMC_SCA_SINK].width;
> > > + unsigned int src_lin, src_col;
> >
> > You could name those variables y and x respectively, it's widely used as
> > coordinates, and could keep lines shorter below.
> >
> > > + u8 *walker = vsca->src_frame;
> > > +
> > > + /* Set each pixel at the src_frame to its sink_frame equivalent */
> > > + for (src_lin = 0; src_lin < src_fmt->height; src_lin++) {
> > > + for (src_col = 0; src_col < src_fmt->width; src_col++) {
> > > + unsigned int snk_lin, snk_col, index;
> > > +
> > > + snk_lin = (src_lin * r->height) / src_fmt->height
> > > + + r->top;
> >
> > This line doesn't depend on src_col, so you can move it outside of the
> > inner loop.
> >
> > > + snk_col = (src_col * r->width) / src_fmt->width
> > > + + r->left;
> > > + index = VIMC_FRAME_INDEX(snk_lin, snk_col, snk_width,
> > > + vsca->bpp);
> >
> > This can probably also be optimized by splitting the calculation behind
> > the VIMC_FRAME_INDEX macro into the part that doesn't depend on the
> > column, and moving that outside of the inner loop, but that's also a
> > candidate for a separate patch.
>
> I have done so, and if you'd like I can include it in this patch.
Sure.
> > > + memcpy(walker, &sink_frame[index], vsca->bpp);
> > > + walker += vsca->bpp;
> > > + }
> > > + }
> > > }
> > >
> > > static void *vimc_sca_process_frame(struct vimc_ent_device *ved,
> > > @@ -492,8 +424,8 @@ static struct vimc_ent_device *vimc_sca_add(struct vimc_device *vimc,
> > > return ERR_PTR(-ENOMEM);
> > >
> > > /* Initialize ved and sd */
> > > - vsca->pads[0].flags = MEDIA_PAD_FL_SINK;
> > > - vsca->pads[1].flags = MEDIA_PAD_FL_SOURCE;
> > > + vsca->pads[VIMC_SCA_SINK].flags = MEDIA_PAD_FL_SINK;
> > > + vsca->pads[VIMC_SCA_SRC].flags = MEDIA_PAD_FL_SOURCE;
> > >
> > > ret = vimc_ent_sd_register(&vsca->ved, &vsca->sd, v4l2_dev,
> > > vcfg_name,
> > > @@ -508,7 +440,8 @@ static struct vimc_ent_device *vimc_sca_add(struct vimc_device *vimc,
> > > vsca->ved.dev = vimc->mdev.dev;
> > >
> > > /* Initialize the frame format */
> > > - vsca->sink_fmt = sink_fmt_default;
> > > + vsca->fmt[VIMC_SCA_SINK] = sink_fmt_default;
> > > + vsca->fmt[VIMC_SCA_SRC] = sink_fmt_default;
> > >
> > > /* Initialize the crop selection */
> > > vsca->crop_rect = crop_rect_default;
--
Regards,
Laurent Pinchart
Powered by blists - more mailing lists