- 24 Sep, 2018 6 commits
-
-
Jia-Ju Bai authored
The driver may sleep in a interrupt handler. The function call paths (from bottom to top) in Linux-4.16 are: [FUNC] kzalloc(GFP_KERNEL) drivers/media/pci/ivtv/ivtv-yuv.c, 938: kzalloc in ivtv_yuv_init drivers/media/pci/ivtv/ivtv-yuv.c, 960: ivtv_yuv_init in ivtv_yuv_next_free drivers/media/pci/ivtv/ivtv-yuv.c, 1126: ivtv_yuv_next_free in ivtv_yuv_setup_stream_frame drivers/media/pci/ivtv/ivtv-irq.c, 827: ivtv_yuv_setup_stream_frame in ivtv_irq_dec_data_req drivers/media/pci/ivtv/ivtv-irq.c, 1013: ivtv_irq_dec_data_req in ivtv_irq_handler To fix this bug, GFP_KERNEL is replaced with GFP_ATOMIC. This bug is found by my static analysis tool DSAC. Signed-off-by:
Jia-Ju Bai <baijiaju1990@gmail.com> Signed-off-by:
Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by:
Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
-
Arnd Bergmann authored
A warning that I thought to be solved by a previous patch of mine has resurfaced with gcc-8: media/imx/imx-media-csi.c: In function 'csi_link_validate': media/imx/imx-media-csi.c:1025:20: error: 'upstream_ep' may be used uninitialized in this function [-Werror=maybe-uninitialized] media/imx/imx-media-csi.c:1026:24: error: 'upstream_ep.bus_type' may be used uninitialized in this function [-Werror=maybe-uninitialized] media/imx/imx-media-csi.c:127:19: error: 'upstream_ep.bus.parallel.bus_width' may be used uninitialized in this function [-Werror=maybe-uninitialized] media/imx/imx-media-csi.c: In function 'csi_enum_mbus_code': media/imx/imx-media-csi.c:132:9: error: '*((void *)&upstream_ep+12)' may be used uninitialized in this function [-Werror=maybe-uninitialized] media/imx/imx-media-csi.c:132:48: error: 'upstream_ep.bus.parallel.bus_width' may be used uninitialized in this function [-Werror=maybe-uninitialized] I spent some more time digging in this time, and think I have a better fix, bailing out of the function that either initializes or errors out here, which simplifies the code enough for gcc to figure out what is going on. The earlier partial workaround can be removed now, as the new workaround is better. Fixes: 890f2769 ("media: imx: work around false-positive warning") Signed-off-by:
Arnd Bergmann <arnd@arndb.de> Signed-off-by:
-
Hans Verkuil authored
Move cec_get_edid_phys_addr() to cec-adap.c. It's not worth keeping a separate source for this. Signed-off-by:
-
Hans Verkuil authored
Several CEC functions are actually specific for use with receivers, i.e. they should be part of the V4L2 subsystem, not CEC. These functions deal with validating and modifying EDIDs for (HDMI) receivers, and they do not actually have anything to do with the CEC subsystem and whether or not CEC is enabled. The problem was that if the CEC_CORE config option was not set, then these functions would become stubs, but that's not right: they should always be valid. So replace the cec_ prefix by v4l2_ and move them to v4l2-dv-timings.c. Update all drivers that call these accordingly. Signed-off-by:
Lars-Peter Clausen <lars@metafoo.de> Cc: <stable@vger.kernel.org> # for v4.17 and up Signed-off-by:
-
Hans Verkuil authored
The cec_phys_addr_validate() function will be moved to V4L2, so use a simplified variant of that function in cec-api.c. cec now no longer calls cec_phys_addr_validate() and it can be safely moved to V4L2. Signed-off-by:
-
Hans Verkuil authored
This function is needed by both V4L2 and CEC, so move this to cec.h as a static inline since there are no obvious shared modules between the two subsystems. This patch, together with the following ones, fixes a dependency bug: if CEC_CORE is disabled, then building adv7604 (and other HDMI receivers) will fail because an essential function is now stubbed out. Signed-off-by:
-
- 17 Sep, 2018 34 commits
-
-
Hugues Fruchet authored
Mode setting depends on last mode set, in particular because of exposure calculation when downscale mode change between subsampling and scaling. At stream on the last mode was wrongly set to current mode, so no change was detected and exposure calculation was not made, fix this. Signed-off-by:
Hugues Fruchet <hugues.fruchet@st.com> Reviewed-by:
Jacopo Mondi <jacopo@jmondi.org> Tested-by:
Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by:
-
Hugues Fruchet authored
When switching from auto to manual mode, V4L2 core is calling g_volatile_ctrl() in manual mode in order to get the manual initial value. Remove the manual mode check/return to not break this behaviour. Signed-off-by:
-
Hugues Fruchet authored
ov5640_set_mode_exposure_calc() is checking binning value but binning value read is buggy, fix this. Rename ov5640_binning_on() to ov5640_get_binning() as per other similar functions. Signed-off-by:
Laurent Pinchart <laurent.pinchart@ideasonboard.com> Reviewed-by:
-
Hugues Fruchet authored
Ensure that auto gain and auto exposure are well restored when changing mode. Signed-off-by:
-
Hugues Fruchet authored
Symptom was black image when capturing HD or 5Mp picture due to manual exposure set to 1 while it was intended to set autoexposure to "manual", fix this. Fixes: bf4a4b51 ("media: ov5640: Don't force the auto exposure state at start time"). Signed-off-by:
-
Jacopo Mondi authored
As of: commit 476dec01 ("media: ov5640: Add horizontal and vertical totals") the timings parameters gets programmed separately from the static register values array. When changing capture mode, the vertical and horizontal totals gets inspected by the set_mode_exposure_calc() functions, and only later programmed with the new values. This means exposure, light banding filter and shutter gain are calculated using the previous timings, and are thus not correct. Fix this by programming timings right after the static register value table has been sent to the sensor in the ov5640_load_regs() function. Fixes: 476dec01 ("media: ov5640: Add horizontal and vertical totals") Tested-by: Steve Longerbeam <slongerbeam@gmail.com> # i.MX6q SabreSD, CSI-2 Tested-by: Loic Poulain <loic.poulain@linaro.org> # Dragonboard-410c, CSI-2 Signed-off-by:
Samuel Bobrowicz <sam@elite-embedded.com> Signed-off-by:
Maxime Ripard <maxime.ripard@bootlin.com> Signed-off-by:
-
Jacopo Mondi authored
Rework the MIPI interface startup sequence with the following changes: - Remove MIPI bus initialization from the initial settings blob - At set_power(1) time power up MIPI Tx/Rx and set data and clock lanes in LP11 during 'sleep' and 'idle' with MIPI clock in non-continuous mode. - At s_stream time enable/disable the MIPI interface output. - Restore default settings at set_power(0) time. Before this commit the sensor MIPI interface was initialized with settings that require a start/stop sequence at power-up time in order to force lanes into LP11 state, as they were initialized in LP00 when in 'sleep mode', which is assumed to be the sensor manual definition for the D-PHY defined stop mode. The stream start/stop was performed by enabling disabling clock gating, and had the side effect to change the lanes sleep mode configuration when stream was stopped. Clock gating/ungating: - ret = ov5640_mod_reg(sensor, OV5640_REG_MIPI_CTRL00, BIT(5), - on ? 0 : BIT(5)); - if (ret) Set lanes in LP11 when in 'sleep mode': - ret = ov5640_write_reg(sensor, OV5640_REG_PAD_OUTPUT00, - on ? 0x00 : 0x70); This commit fixes an issue reported by Jagan Teki on i.MX6 platforms that prevents the host interface from powering up correctly: https://lkml.org/lkml/2018/6/1/38 It also improves MIPI capture operations stability on my testing platform where MIPI capture often failed and returned all-purple frames. Fixes: f22996db ("media: ov5640: add support of DVP parallel interface") Tested-by: Steve Longerbeam <slongerbeam@gmail.com> (i.MX6q SabreSD, CSI-2) Tested-by: Loic Poulain <loic.poulain@linaro.org> (Dragonboard-410c, CSI-2) Reported-by:
Jagan Teki <jagan@amarulasolutions.com> Signed-off-by:
-
Javier Martinez Canillas authored
The function not only does initialization but also registers the subdevice so change its name to make this more clear. Signed-off-by:
Javier Martinez Canillas <javierm@redhat.com> Signed-off-by:
-
Javier Martinez Canillas authored
The driver registers the v4l2 subdevice before attempting to power on the chip and checking its ID. This means that a media device driver that it's waiting for this subdevice to be bound, will prematurely expose its media device node to userspace because if something goes wrong the media entity will be cleaned up again on the ov2680 probe function. This also simplifies the probe function error path since no initialization is made before attempting to enable the resources or checking the chip ID. Fixes: 3ee47cad ("media: ov2680: Add Omnivision OV2680 sensor driver") Signed-off-by:
-
Dan Carpenter authored
This code doesn't check for NULL consistently and it generates a Smatch warning: drivers/media/i2c/sr030pc30.c:575 sr030pc30_base_config() error: we previously assumed 'info->pdata' could be null (see line 572) Fortunately, "info->pdata" can't be NULL to that check can be removed. The other thing is that if "ret" is an error code here, then we don't want to do the next call to cam_i2c_write(), so actually let's flip that test around and return the error. This is more of a theoretical issue than something which is likely to affect real life. Signed-off-by:Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by:
-
Sakari Ailus authored
The sub-device name is set right after in v4l2_i2c_subdev_init(). Remove the redundant strcpy() call. Signed-off-by:
-
Sakari Ailus authored
Use v4l2_i2c_subdev_set_name() to set the name of the smiapp driver's sub-devices. There is no functional change. Signed-off-by:
-
Sakari Ailus authored
v4l2_i2c_subdev_set_name() can be used to assign a name to a sub-device. This way uniform names can be formed easily without having to resort to things such as snprintf in drivers. Signed-off-by:
-
Luca Ceresoli authored
Signed-off-by:
Luca Ceresoli <luca@lucaceresoli.net> Signed-off-by:
-
Luca Ceresoli authored
Currently 2-bytes and 3-bytes registers are read one byte at a time, doing the needed shift & mask each time. Replace all of this code by a unique helper function that calls regmap_bulk_read(), which has two advantages: - reads all the bytes in a unique I2C transaction - simplifies code to read multibyte registers Signed-off-by:
-
Luca Ceresoli authored
This parameter holds the number of bytes, not bits. Signed-off-by:
-
Luca Ceresoli authored
A mix of "mode", "format" and "frmfmt" is used to refer to the sensor readout mode. Use the term "mode" for all of them. Now "format" is only used in the V4L2 meaning. Signed-off-by:
-
Luca Ceresoli authored
Forcibly setting the subdev name to DRIVER_NAME (i.e. "IMX274") makes it non-unique and less informative. Let the driver use the default name from i2c, e.g. "IMX274 2-001a". Signed-off-by:
-
Luca Ceresoli authored
Rearrange the imx274_start_<N> register tables to better match the datasheet and slightly simplify code: - collapes tables 1 and 2, they are applied one after each other and together they implement the fixed part 1 of the startup procedure in the datasheet - while there, cleanup comments - rename tables 3 and 4 -> 2 and 3, coherently with the datasheet Signed-off-by:
-
Luca Ceresoli authored
The "mode" has been renamed to "binning" in commit 39dd23dc ("media: imx274: add cropping support via SELECTION API"), but this define has not been updated. Signed-off-by:
-
Sakari Ailus authored
The device is already suspended when it's the driver's remove function is called. Remove redundant pm_runtime_set_suspended call. Signed-off-by:
-
Sakari Ailus authored
Add the DT binding documentation for dw9714 and dw9807-vcm to the MAINTAINERS file. The dw9807-vcm binding documentation file is renamed to match the dw9807's VCM bit's compatible string. Signed-off-by:
-
Kieran Bingham authored
The UDS is currently restricted based on a partition size of 256 pixels. Document the actual restrictions, but don't increase the implementation. The extended partition algorithm may later choose to utilise a larger partition size to support overlapping partitions which will improve the quality of the output images. Signed-off-by:
Kieran Bingham <kieran.bingham+renesas@ideasonboard.com> Signed-off-by:
Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com> Signed-off-by:
-
Kieran Bingham authored
The SRU is currently restricted to 256 pixels as part of the current partition algorithm. Document that the actual capability of this component is 288 pixels, but don't increase the implementation. The extended partition algorithm may later choose to utilise a larger input to support overlapping partitions which will improve the quality of the output images. Signed-off-by:
-
Kieran Bingham authored
The style of this driver uses periods at the end of sentences in comments, but it is applied inconsitently. Update a selection of comments which were discovered to be missing their period. Also fix the spelling of one usage of 'instantiate' Signed-off-by:
-
Kieran Bingham authored
The VSP1 has a minimum width and height of a single pixel, with the exception of pixel formats with sub-sampling. Remove the artificial minimum width and minimum height limitation, and instead clamp the minimum dimensions based upon the sub-sampling parameter of that dimension. Signed-off-by:
-
Koji Matsuoka authored
YCbCr planar formats can have different pitch values for the luma and chroma planes. This isn't taken into account in the driver. Fix it. Based on a BSP patch from Koji Matsuoka <koji.matsuoka.xm@renesas.com>. Fixes: 7863ac50 ("drm: rcar-du: Add tri-planar memory formats support") [Updated documentation of the struct vsp1_du_atomic_config pitch field] Signed-off-by:
Koji Matsuoka <koji.matsuoka.xm@renesas.com> Signed-off-by:
-
Laurent Pinchart authored
The LIF module has a data buffer to accommodate clock rate differences between the DU and the VSP. Several programmable threshold values control DU start of frame notification by the VSP and VSP clock stop/resume. The R-Car Gen2 and Gen3 datasheets recommend values for the different SoCs. Update the driver to use the recommended values for optimal operation. Based on a BSP patch from Koji Matsuoka <koji.matsuoka.xm@renesas.com>, with Gen2 and V3H/V3M updates. Signed-off-by:
-
Laurent Pinchart authored
All source files of the vsp1 driver are licensed under the GPLv2+ except for vsp1_regs.h which is licensed under GPLv2. This is caused by a bad copy&paste that dates back from the initial version of the driver. Fix it. Acked-by:
Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se> Acked-by:
Wolfram Sang <wsa+renesas@sang-engineering.com> Acked-by:
Nobuhiro Iwamatsu <iwamatsu@nigauri.org> Signed-off-by:
-
Kieran Bingham authored
Add myself as a co-maintainer for the Renesas VSP driver. Signed-off-by:
-
Mauro Carvalho Chehab authored
The recent tvp5150 patchset added two new warnings: drivers/media/i2c/tvp5150.c: In function 'tvp5150_querystd': drivers/media/i2c/tvp5150.c:829:18: warning: unused variable 'decoder' [-Wunused-variable] struct tvp5150 *decoder = to_tvp5150(sd); ^~~~~~~ drivers/media/i2c/tvp5150.c:1522:6: warning: no previous prototype for 'tvp5150_volatile_reg' [-Wmissing-prototypes] bool tvp5150_volatile_reg(struct device *dev, unsigned int reg) ^~~~~~~~~~~~~~~~~~~~ Get rid of them. Signed-off-by: -
Mauro Carvalho Chehab authored
When irq is used, the lock is set via IRQ code. When it isn't, the driver just assumes it is always locked. Instead, read the lock status from the status register. Signed-off-by: -
Philipp Zabel authored
Add the querystd video_op and make it return V4L2_STD_UNKNOWN while the TVP5150 is not locked to a signal. Signed-off-by:
Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by:
Marco Felsch <m.felsch@pengutronix.de> Signed-off-by:
-
Marco Felsch authored
Add callback to retrieve the current set norm. Signed-off-by:
-
