These formats store the color information of the image
in a geometrical representation. The colors are mapped into a
cylinder, where the angle is the HUE, the height is the VALUE
and the distance to the center is the SATURATION. This is a very
useful format for image segmentation algorithms.
Signed-off-by: NRicardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Delay suspending the device by 1000 ms by default. This is done on
explicit power off through s_power() callback, through releasing the
file descriptor, NVM read or when the probe finishes.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Switch to runtime PM in sensor power management. The internal power count
is thus removed.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Use the suspend and resume ops for freeze, thaw, poweroff and restore
callbacks as well.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

The dev field of the v4l2_subdev was left NULL for the pixel array and
binner sub-devices. Fix this.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Checking that the mutex is not acquired is unnecessary for user processes
are stopped by this point. Drop the check.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Don't complain about a failure to compute the pre_pll divisor. The
function is used to determine whether a particular combination of bits per
sample value and a link frequency can be used, in which case there are
lots of unnecessary driver messages. During normal operation the failure
generally does not happen. Use dev_dbg() instead.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

The first time the sensor is powered on, the information is not yet
available.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

The media bus code obtained for try format may have been a code that the
sensor did not even support. Use a supported code with the current pixel
order.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

The code probably has been unindented at some point but rewrapping has not
been done. Do it now.

Also remove a useless memory allocation failure message.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Provide more debugging information on reading the frame layout.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Besides the image data, SMIA++ compliant sensors also provide embedded
data in form of registers used to capture the image. Store this
information for later use in frame descriptor and routing.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Replace plain value 2 with SMIAPP_PADS when referring to the number of
pads.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

The initialisation of the source sub-device is somewhat different as it's
not created by the smiapp driver itself. Remove redundancy in initialising
the two kind of sub-devices.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

The information gathered during frame format reading will be required
earlier in the initialisation when it was available. Also return an error
if frame format cannot be obtained.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

The smiapp_probe() is the sole caller of smiapp_init(). Unify the two.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

If the registered() callback failed, resources were left unaccounted for.
Fix this, as well as add unregistering the sub-devices in driver
unregistered() callback.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Initialise the sensor in probe. The reason why it wasn't previously done
in case of platform data was that the probe() of the driver that provided
the clock through the set_xclk() callback would need to finish before the
probe() function of the smiapp driver. The set_xclk() callback no longer
exists.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Instead, calculate how much is needed and then allocate the memory
dynamically.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

The same pixel array size is required for the active format of each
sub-device sink pad and try format of each sink pad of each opened file
handle as well as for the native size rectangle.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Remove the loop in sub-device registration and create each sub-device
explicitly instead.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

This allows determining the number of pads in the entity based on the
sensor.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Define the number of pads explicitly in initialising the sub-devices.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Simplify smiapp_init() by moving the initialisation of individual
sub-devices to a separate function.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NSebastian Reichel <sre@kernel.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Fix smatch error:

media-git/drivers/media/platform/mtk-mdp/mtk_mdp_vpu.c:100 mtk_mdp_vpu_send_msg() error: double unlock 'mutex:&ctx->mdp_dev->vpulock'
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Improve the MT21C documentation, making it clearer that this format
requires the MDP for further processing.

Also fix the fourcc (it was a fivecc :-) )
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

This patch fix build error without CONFIG_PM_RUNTIME
and CONFIG_PM_SLEEP
Signed-off-by: NMinghsiu Tsai <minghsiu.tsai@mediatek.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

This patch fix build warning in arch x86
Signed-off-by: NMinghsiu Tsai <minghsiu.tsai@mediatek.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Add Minghsiu Tsai, Houlong Wei and Andrew-CT Chen as
maintainers for Mediatek MDP driver
Signed-off-by: NMinghsiu Tsai <minghsiu.tsai@mediatek.com>
Signed-off-by: NHoulong Wei <houlong.wei@mediatek.com>
Signed-off-by: NAndrew-CT Chen <andrew-ct.chen@mediatek.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Add V4L2_PIX_FMT_MT21C in format list.

[mchehab@s-opensource.org: re-add mtk_mdp_size_align]
Signed-off-by: NMinghsiu Tsai <minghsiu.tsai@mediatek.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

The mtk_mdp_dbg() is empty if !DEBUG. This causes the following
warnings:

	drivers/media/platform/mtk-mdp/mtk_mdp_m2m.c: In function ‘mtk_mdp_try_fmt_mplane’:
	drivers/media/platform/mtk-mdp/mtk_mdp_m2m.c:231:52: warning: suggest braces around empty body in an ‘if’ statement [-Wempty-body]
	        org_w, org_h, pix_mp->width, pix_mp->height);
                                                    ^
	drivers/media/platform/mtk-mdp/mtk_mdp_m2m.c: In function ‘mtk_mdp_m2m_start_streaming’:
	drivers/media/platform/mtk-mdp/mtk_mdp_m2m.c:414:21: warning: suggest braces around empty body in an ‘if’ statement [-Wempty-body]
	        ctx->id, ret);
	                     ^

With could actually make the code to do something wrong. So,
add an empty block to make it be parsed ok.
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Add MDP node for MT8173
Signed-off-by: NMinghsiu Tsai <minghsiu.tsai@mediatek.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Add MDP driver for MT8173
Signed-off-by: NMinghsiu Tsai <minghsiu.tsai@mediatek.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Add a DT binding documentation of MDP for the MT8173 SoC
from Mediatek
Signed-off-by: NMinghsiu Tsai <minghsiu.tsai@mediatek.com>
Acked-by: NRob Herring <robh@kernel.org>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

VPU driver add mdp support

[mchehab@s-opensource.com: fix a merge conflict and make checkpatch happy]
Signed-off-by: NMinghsiu Tsai <minghsiu.tsai@mediatek.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Add video decoder node for MT8173
Signed-off-by: NTiffany Lin <tiffany.lin@mediatek.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Add V4L2_PIX_FMT_MT21C support
Signed-off-by: NTiffany Lin <tiffany.lin@mediatek.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Add V4L2_PIX_FMT_MT21C documentation
Signed-off-by: NTiffany Lin <tiffany.lin@mediatek.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

Add V4L2_PIX_FMT_MT21C format used on MT8173 driver.
It is compressed format and need MT8173 MDP driver to transfer to other
standard format.
Signed-off-by: NTiffany Lin <tiffany.lin@mediatek.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>

pdev is being null checked, however, prior to that it is being
dereferenced by platform_get_drvdata.  Move the assignments of
vpu and run to after the pdev null check to avoid a potential
null pointer dereference.
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Reviewed-by: NTiffany Lin <tiffany.lin@mediatek.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@s-opensource.com>