If the message length was wrong, the dprintk() after the 'err' label
was bypassed. Fix that, and fix a smatch warning at the same time:
mb86a16.c:1514 mb86a16_send_diseqc_msg() warn: missing unwind goto?
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
If the media bus is unsupported, then return -EINVAL. Instead it
returned 'ret' which happened to be 0.
This fixes a smatch warning:
ov7670.c:1843 ov7670_parse_dt() warn: missing error code? 'ret'
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Fixes: 01b8444828 ("media: v4l2: i2c: ov7670: Implement OF mbus configuration")
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Since 'ret' is known to be 0, just return '0'. This fixes a smatch warning:
vpbe_display.c:1152 vpbe_display_open() warn: missing error code? 'err'
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Cc: "Lad, Prabhakar" <prabhakar.csengg@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Since 'ret' is known to be 0, just return '0'. This fixes a smatch warning:
venc_h264_if.c:568 h264_encode_frame() warn: missing error code? 'ret'
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Cc: Andrew-CT Chen <andrew-ct.chen@mediatek.com>
Cc: Yunfei Dong <yunfei.dong@mediatek.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
An error was returned at one point without going through the
goto label for proper unwinding.
This fixes a smatch warning:
mtk-mdp3-comp.c:1005 mdp_comp_config() warn: missing unwind goto?
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Cc: Moudy Ho <moudy.ho@mediatek.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
In three places there was a return instead of a goto to the unwind
code.
This fixes three smatch warnings:
drxj.c:9542 ctrl_get_qam_sig_quality() warn: missing unwind goto?
drxj.c:10919 ctrl_set_standard() warn: missing unwind goto?
drxj.c:11466 drxj_open() warn: missing unwind goto?
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
The probe function used 'goto out' everywhere, except in one
place where it returned an error. That too should be a 'goto out'.
This fixes a smatch warning:
mmp-driver.c:257 mmpcam_probe() warn: missing unwind goto?
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Since 'ret' is known to be 0, just return '0'. This fixes two smatch warnings:
af9015.c:1168 af9015_rc_query() warn: missing error code? 'ret'
af9015.c:1177 af9015_rc_query() warn: missing error code? 'ret'
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Cc: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Since 'ret' is known to be 0, just return '0'. This fixes two smatch warnings:
cxd2880_tnrdmd.c:2165 cxd2880_tnrdmd_check_internal_cpu_status() warn: missing error code? 'ret'
cxd2880_tnrdmd.c:2169 cxd2880_tnrdmd_check_internal_cpu_status() warn: missing error code? 'ret'
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Cc: Yasunari Takiguchi <Yasunari.Takiguchi@sony.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Since 'ret' is known to be 0, just return '0'. This fixes a smatch warning:
s5c73m3-core.c:439 __s5c73m3_s_stream() warn: missing error code? 'ret'
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Cc: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
DEFINE_RES_IRQ_NAMED already casts to (struct resource), so no
need to do it again.
This fixes a sparse warning:
vpif.c:483:20: warning: cast to non-scalar
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Cc: "Lad, Prabhakar" <prabhakar.csengg@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
This struct can be static. This fixes a sparse warning:
visl-video.c:690:22: warning: symbol 'visl_qops' was not declared. Should it be static?
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Cc: Daniel Almeida <daniel.almeida@collabora.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Inline the single invocation of the macro STATS_ENABLED().
The macro abstraction is not necessary because the logic behind it is only
used once.
Link: https://lore.kernel.org/r/20230120171408.16099-1-bpappas@pappasbrent.com
Signed-off-by: Brent Pappas <bpappas@pappasbrent.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Replace the macros SUBWORD() and INV_SUBWORD() with functions to comply
with Linux coding style standards.
Link: https://lore.kernel.org/r/20230120182625.23227-1-bpappas@pappasbrent.com
Signed-off-by: Brent Pappas <bpappas@pappasbrent.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Replace the function-like macros FPNTBL_BYTES(), SCTBL_BYTES(), and
MORPH_PLANE_BYTES() with functions to comply with Linux coding style
standards.
Replace multiplication with calls to array_size() and array3_size()
to prevent accidental arithmetic overflow.
Link: https://lore.kernel.org/r/20230118160739.26059-1-bpappas@pappasbrent.com
Signed-off-by: Brent Pappas <bpappas@pappasbrent.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Depending on which order userspace makes various v4l2 calls, the sensor
might still be powered down when set_fmt is called.
What should really happen here is delay the writing of the mode-related
registers till streaming is started, but for now use the same quick fix
as the atomisp_ov2680 / atomisp_gc0310 code and call power_up() from
set_fmt() in combination with keeping track of the power-state to avoid
doing the power-up sequence twice.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
On ov2722_set_fmt() calls with format->which == V4L2_SUBDEV_FORMAT_TRY,
ov2722_set_fmt() does not talk to the sensor, so there is no need to
lock the dev->input_lock mutex in this case.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
The comment claims the PWDN pin is active when pulled down in other words,
it is /power-down so it needs to be driven high to get the sensor
powered-up (not powered down) and flag is 1 when powering-up the sensor
so the ! is wrong, drop it.
This also matches with the schematics which I have which shows GPIO1 also
enables a 3.3v line to the sensor-module which controls the privacy-LED
and indeed before this patch the privacy LED was inverted from what it
should be (and the sensor did not work).
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Call atomisp_gmin_remove_subdev() on probe failure to properly free
the GPIOs and other resources acquired by the gmin_camera_platform_data()
call earlier.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
The DSDT of all Windows BYT / CHT devices which I have seen has proper
ACPI powermagement for the clk and regulators used by the sensors.
So there is no need for the whole custom atomisp_gmin custom code to
disable the ACPI pm and directly poke at the PMIC for this.
Replace all the atomisp_gmin usage with using the new
atomisp_register_sensor_no_gmin() / atomisp_unregister_subdev()
helpers which allow registering a sensor with the atomisp code
without using any of the atomisp_gmin power-management code.
Note eventually these calls should be replaced by the standard
v4l2_async_register_subdev_sensor() mechanism.
But this first requires a bunch of work to the atomisp main code
to make it set the necessary fwnodes up, similar to how
drivers/media/pci/intel/ipu3/cio2-bridge.c does this.
This has been tested on:
-Trekstor Surftab duo W1 10.1, CHT, AXP288 PMIC, 2x ov2680 sensor
-Asus T101HA, CHT, TI PMIC, 1x ov2680 sensor
-MPMAN Converter 9, BYT, AXP288 PMIC, ov2680 back, gc0310 front sensor
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Use devm_kzalloc() to allocate the sensor data struct. It is always free-ed
as the last step of probe-error-exit or remove, so it can be devm-managed.
This will make unwinding things easier when support is added to the ov2680
code to use standard GPIO APIs instead of the custom atomisp_gmin code.
This also allows dropping the out_free label and use direct return
on errors.
This may seem like a functional change since the out_free label also
did a v4l2_device_unregister_subdev() but at the 2 changed returns
the device is not registered yet, so that always is a no-op and can
be dropped.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Add a dev local variable to probe(), to allow shortening
&client->dev in various places, including further patches
in this series.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Using dev as name for variables pointing to struct ov2680_device is a bit
unfortunate choice.
All the recently added / rewritten code is already using sensor for this,
replace the remaining usages of "struct ov2680_device *dev" with
"struct ov2680_device *sensor".
Note the power_up()/power_down() related functions are not changed as
these will be removed in one of the next patches.
No functional changes.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Add runtime-pm support. This is a preparation patch for letting
ACPI deal with the regulators and clocks instead of the DIY code
in atomisp_gmin_platform.c.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Move the setting of the mode to stream on, this also allows
delaying power-on till streaming is started.
And drop the deprecated s_power callback since this now no long
is necessary.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Remove unused includes and sort the remaining ones alphabetically.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Use the same indentation level for all #define values.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
The ov2680 only supports a single format, there is no need to
use a define for this.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Remove a bunch of unused data-types and defines from ov2680.h.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
3 fixes for the framesize list:
1. Drop modes < 640x480, these are made by significant cropping,
leading to such a small remainig field-of-view that they are
not really usable
2. 1616x1082 is presumably intended to be 1600x1080 + 16 pixels
padding in both dimensions, but the height is wrong.
Change this to 1616x1096.
3. The 800x600 mode is missing the 16 pixels padding and
720x592 is missing 16 pixels padding in its width and
the 720x576 base mode is a mode with non square pixels,
while the sensor has square pixels.
Replace both with 768x576 + 16 pixels padding -> 784x592
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Drop struct ov2680_resolution and the ov2680_res_preview[] array,
this is now only used in ov2680_enum_frame_size() and only
the width + height are used there.
Replace this with a new struct v4l2_frmsize_discrete ov2680_frame_sizes[]
array.
No functional changes.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Since we now calculate timings baded on the desired width and height,
any width and height are valid as long as they don't exceed the sensor's
dimensions.
Drop the v4l2_find_nearest_size() call and instead clamp the requested
width and height.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Fix and simplify ov2680_enum_frame_interval(), the index is not
an index into ov2680_res_preview[], so using N_PREVIEW is wrong.
Instead it is an index indexing the different framerates for
the resolution specified in fie->width, fie->height.
Since the ov2680 code only supports a single fixed 30 fps,
index must always be 0 and we don't need to check the other
fie input values.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
The res member of struct ov2680_device isn't read anywhere anymore,
drop it.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
The fps, lines-per-frame and skip-frames values are the same for all
resolutions, use defines for these.
The ov2680_res_preview[] incorrectly sets fps to 60 for some low-res
modes, this is incorrect with the current fixed (resolution independent)
lines-per-frame value.
Note this not drop the now no longer used fps, lines-per-frame and
skip-frames struct ov2680_resolution members. The entire struct is going
away in the next patches so that would just cause unnecessary changes.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Instead of using a long fixed register settings list for each resolution,
calculate the register settings based on the requested width + height.
This will allow future enhancements like adding hblank and vblank controls
and adding selection support.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
By default the ov2680 automatically sets the window to match the outputsize
and automatically adjusts it to keep the bayer pattern stable when enabling
hflip/vflip.
This does not work for the 1616x1216 mode because there is no room to
adjust the window there. To make flipping work in the 1616 wide modes the
register lists for those modes set bit 0 of 0x5708 (manual_win_en) to 1 and
ov2680_set_bayer_order() updates the bayer-order on the pad to match.
But ov2680_set_bayer_order() is always called, so when enabling flipping
on modes with a width of less then 1616 now results in the wrong bayer
order being reported on the pad since the sensor is auto-adjusting
the window in this case.
Specify the correct (== output-size) window-size in all resolutions
register-list and always set the manual_win_en bit, so that the bayer order
is changed on hflip/vflip enable on all resolutions.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Add a test pattern control. This is a 1:1 copy of the test pattern
control in the main drivers/media/i2c/ov2680.c driver.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Add exposure and gain controls. This allows controlling
the exposure and gain through standard v4l2 IOCTLs.
Note the register defines for the exposure and gain registers
are renamed to match the datasheet.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Exposure and gain control should use standard v4l2 controls,
not a custom ioctl.
The next patch in this series will re-add support as standard controls,
this is split into 2 patches for easier reviewing.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Rework the flip ctrls to be more like those of mainline (non staging)
drivers.
This is modelled after the main ov2680 and ov5693 drivers. This also
introduces __ov2680_get_pad_format() to make the ov2680 code more compliant
with the mainline v4l2-subdev APIs.
Note the OV2680_FLIP_REG and OV2680_MIRROR_REG defines are renamed to
OV2680_REG_FORMAT1 and OV2680_REG_FORMAT2 to match the datasheet.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Use the new ov_16bit_addr_reg_helpers.h instead of duplicating
the ovxxxx sensor I2C register access helpers found in many different
sensor drivers.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
The following drivers under drivers/media/i2c: ov08x40.c, ov13858.c,
ov13b10.c, ov2680.c, ov2685.c, ov2740.c, ov4689.c, ov5670.c,
ov5675.c, ov5695.c, ov8856.c, ov9282.c and ov9734.c,
as well as various "atomisp" sensor drivers in drivers/staging, *all*
use register access helpers with the following function prototypes:
int ovxxxx_read_reg(struct ovxxxx_dev *sensor, u16 reg,
unsigned int len, u32 *val);
int ovxxxx_write_reg(struct ovxxxx_dev *sensor, u16 reg,
unsigned int len, u32 val);
To read/write registers on Omnivision OVxxxx image sensors wich expect
a 16 bit register address in big-endian format and which have 1-3 byte
wide registers, in big-endian format (for the higher width registers).
Add a new ov_16bit_addr_reg_helpers.h header file with static inline
versions of these register access helpers, so that this code duplication
can be removed.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Don't use MEDIA_ENT_F_V4L2_SUBDEV_UNKNOWN to initialize the function field
of various media-entity links.
This fixes the following warnings showing up in dmesg:
atomisp-isp2 0000:00:03.0: Entity type for entity ATOM ISP CSI2-port0 was not initialized!
atomisp-isp2 0000:00:03.0: Entity type for entity ATOM ISP CSI2-port1 was not initialized!
atomisp-isp2 0000:00:03.0: Entity type for entity ATOM ISP CSI2-port2 was not initialized!
atomisp-isp2 0000:00:03.0: Entity type for entity tpg_subdev was not initialized!
atomisp-isp2 0000:00:03.0: Entity type for entity ATOMISP_SUBDEV_0 was not initialized!
atomisp-isp2 0000:00:03.0: Entity type for entity ATOMISP_SUBDEV_1 was not initialized!
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Looking at csi2_link_setup(), this function can never work,
it does a switch-case like this:
switch (local->index | is_media_entity_v4l2_subdev(remote->entity))
with cases like this:
case ATOMISP_SUBDEV_PAD_SOURCE | MEDIA_ENT_F_V4L2_SUBDEV_UNKNOWN
where ATOMISP_SUBDEV_PAD_SOURCE matches an index (0-1) and
MEDIA_ENT_F_V4L2_SUBDEV_UNKNOWN is 0x00020000, but
is_media_entity_v4l2_subdev(remote->entity) does not return
MEDIA_ENT_F_* values, it return a bool, so 0 or 1 which means
that non of the cases can ever match the input value.
Looking at the rest of the function all it ever does (if it
would actually hit one of the cases) is set the atomisp_mipi_csi2_device
struct's output member.
And checking the rest of the atomisp code that member is never
read. Also userspace does not actually setup media-controller
links when using the atomisp /dev/video$ nodes since all the links
are fixed. So csi2_link_setup() never runs.
Remove the unnecessary and broken csi2_link_setup() function
and also remove the unused atomisp_mipi_csi2_device struct's
output member.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Looking at isp_subdev_link_setup(), this function can never work,
it does a switch-case like this:
switch (local->index | is_media_entity_v4l2_subdev(remote->entity))
with cases like this:
case ATOMISP_SUBDEV_PAD_SINK | MEDIA_ENT_F_V4L2_SUBDEV_UNKNOWN
where ATOMISP_SUBDEV_PAD_SINK matches an index (0-4) and
MEDIA_ENT_F_V4L2_SUBDEV_UNKNOWN is 0x00020000, but
is_media_entity_v4l2_subdev(remote->entity) does not return
MEDIA_ENT_F_* values, it return a bool, so 0 or 1 which means
that non of the cases can ever match the input value.
Looking at the rest of the function all it ever does (if it
would actually hit one of the cases) is set the atomisp_sub_device
struct's input member.
And checking the rest of the atomisp code that member is never
read. Also userspace does not actually setup media-controller
links when using the atomisp /dev/video$ nodes since all the links
are fixed. So isp_subdev_link_setup() never runs.
Remove the unnecessary and broken isp_subdev_link_setup() function
and also remove the unused atomisp_sub_device struct's input member.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
The device core will call ACPI to turn the device (i2c_client) for a sensor
on / put it in D0 before calling its probe() method.
This takes a reference on all of the ACPI power-resources belonging to
the device. Since the atomisp_gmin_platform code disables ACPI
power-resource management and does its own pm, this reference never gets
released.
This is a problem for ACPI power-resources which are shared with other
devices since those now never get turned off again (nor back on again).
Explicitly put the device in D3 before disabling the ACPI power-resource
management.
Note that atomisp_register_i2c_module() runs near the end of the sensor
driver's probe() function, after the driver is done with probing the hw.
So the power-resouces (the same resources as directly controlled by
the atomisp platform code) getting turned off (a second time, as they are
already off) is not a problem.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
atomisp_set_fmt() first does:
v4l2_fill_mbus_format(&vformat.format, ...);
vformat.format.height += padding_h;
vformat.format.width += padding_w;
ret = v4l2_subdev_call(isp->inputs[asd->input_curr].camera, pad,
set_fmt, NULL, &vformat);
if (ret)
return ret;
f->fmt.pix.width = vformat.format.width - padding_w;
f->fmt.pix.height = vformat.format.height - padding_h;
this happens with the original padding w/h = 16 values and then later
on it calls:
ret = atomisp_set_fmt_to_snr(vdev, &s_fmt,
f->fmt.pix.pixelformat, padding_w,
padding_h, dvs_env_w, dvs_env_h);
Which repeats the above structure. If at that point padding w/h are
changed to 12 then it will now request a different output-size of
the sensor driver.
The sensor drivers so far have actually been ignoring this since they use
v4l2_find_nearest_size() on a fixed resolution list and the nearest
resolution will be the one from the earlier calls where padding w/h
was 16.
But there really is no reason for sensor drivers to use a fixed
resolution list. They make lower resolutions using cropping so they
can make any resolution as long as width/height are even numbers.
Dropping the fixed-resolution list limit from sensors on BYT results
in trying to start streaming failing because the resolution set to
the sensor now no longer matches with the resolution used during
the initial part of the configuration done by atomisp_set_fmt().
Drop the BYT specific overriding of the padding_w/h to 12, so that
the padding in the first and second s_fmt calls made to the sensor
matches, to fix stream start failing when the fixed resolution list
is dropped.
Reviewed-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
