- 21 Dec, 2012 30 commits
-
-
Sachin Kamat authored
module_usb_driver eliminates a lot of boilerplate by replacing module_init() and module_exit() calls. Signed-off-by:
Sachin Kamat <sachin.kamat@linaro.org> Signed-off-by:
Hans de Goede <hdegoede@redhat.com> Signed-off-by:
Mauro Carvalho Chehab <mchehab@redhat.com>
-
Jacob Schloss authored
Add the USB ID for the Kinect for Windows RGB camera so it can be used with the gspca_kinect driver. Signed-off-by:
Jacob Schloss <jacob.schloss@unlimitedautomata.com> Signed-off-by:
Antonio Ospite <ospite@studenti.unina.it> Signed-off-by:
-
Hans de Goede authored
When the v4l1 drivers were removed, there docs were forgotten. Signed-off-by:
-
Hans de Goede authored
The stk-webcam module is inserted upside-down in some webcams, so we need to de hflip and vflip by default on some models. Note that this patch inverts the value of the controls as reported by the control API in this case so that for the user they still make sense (iow not doing any flipping from the ctrl API pov results in an upright image). Reported-by:
Jose Gómez <adler@dreamcoder.org> Signed-off-by:
-
Hans de Goede authored
Some cams have their led connected in such a way that on = off and visa versa unfortunately we cannot tell this from the driver in any way, so add a module parameter for this. Reported-by:
Yuri Glushkov <yglushkov@yahoo.com> Signed-off-by:
-
Jean-François Moine authored
Setting the H and V flip controls at webcam connection time prevents the webcam to work correctly. This patch checks if the webcam is streaming before setting the flips. It does not set the flips (nor other controls) at webcam start time. Tested-by:
Philippe ROUBACH <philippe.roubach@free.fr> Signed-off-by:
Jean-François Moine <moinejf@free.fr> Signed-off-by:
-
Andrzej Hajda authored
Some sensors can have more than one pad (case of S5C73M3). In such cases FIMC assumes the last pad of the sensor is the source pad. Signed-off-by:
Andrzej Hajda <a.hajda@samsung.com> Signed-off-by:
Kyungmin Park <kyungmin.park@samsung.com> Signed-off-by:
Sylwester Nawrocki <s.nawrocki@samsung.com> Signed-off-by:
-
Sylwester Nawrocki authored
Add second source media pad for the FIFO data output to FIMC-IS and implement subdev s_stream op for configurations where FIMC-LITE is used as a glue logic between FIMC-IS and MIPI-CSIS or an image sensor. The second source media pad will be linked to the FIMC-LITE video node. For proper configuration the attached image sensor/video encoder properties are needed, like video bus type, signal polarities, etc. For this purpose there is a small routine added that walks the pipeline and returns the sensor subdev. Signed-off-by:
-
Sylwester Nawrocki authored
Add subdev group id definition for FIMC-IS ISP and sensor subdev. While at it rename all group id definitions to start with GRP_ID. Signed-off-by:
-
Sylwester Nawrocki authored
Dump registers contents together with the event counters state in VIDIOC_LOG_STATUS ioctl. Signed-off-by:
-
Sylwester Nawrocki authored
Enable only MIPI CSI-2 data lanes at the DPHY that are actively used, rather than unmasking all unconditionally. Signed-off-by:
-
Sylwester Nawrocki authored
The MIPI CSIS device supports MIPI CSI-2 RAW8, RAW10, RAW12 data types. Add related media bus pixel format definitions. This doesn't cover all possible supported media bus pixel formats. Signed-off-by:
-
Sylwester Nawrocki authored
Add variant data structures for Exynos4212 and Exynos4412 SoC. Add 'const' qualifier for the variant description structures. Also remove has_cam_if flags from FIMC3 on Exynos4210 SoC is it has no interconnections the camera ports. Signed-off-by:
-
Sylwester Nawrocki authored
The FIMC FIFO output is not supported in the driver due to some hardware issues thus we can remove some code as out_path is always FIMC_IO_DMA. Signed-off-by:
-
Sylwester Nawrocki authored
Use v4l2_info() to make it possible to identify which FIMC-LITE device instance the logs refer to. Signed-off-by:
-
Sylwester Nawrocki authored
This patch adds definition of media bus code for YUV pixel format transferred in 30-bit samples where each component has 10 bits width. [mchehab@redhat.com: fix a merge conflict at v4l2-mediabus.h] Signed-off-by:
-
Sylwester Nawrocki authored
The counter field is unsigned so >= 0 condition always evaluates to true. Fix this to log events for which counter is > 0 or for all when in debug mode. Signed-off-by:
-
Sylwester Nawrocki authored
Setting FIMC_REG_CITRGFMT_FLIP_X_MIRROR bit causes X-axis image flip (vertical flip) and thus it corresponds to V4L2_CID_VFLIP. Likewise, setting FIMC_REG_CITRGFMT_FLIP_Y_MIRROR bit causes Y-axis image flip (horizontal flip) and thus it corresponds to V4L2_CID_HFLIP. Currently the driver does X-axis flip when V4L2_CID_HFLIP is set and Y-axis flip for V4L2_CID_VFLIP. Fix this incorrect assignment by setting proper FIMC_REG_CITRGFMT register bits for ctx->hflip and ctx->vflip. Reported-by:
-
Sakari Ailus authored
Correct a typo. v4l2_plane.m.userptr is unsigned long, not __unsigned long. Signed-off-by:
Sakari Ailus <sakari.ailus@iki.fi> Acked-by:
Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by:
-
Sakari Ailus authored
Set buffer timestamp flags for videobuf, videobuf2 and drivers that use neither. Signed-off-by:
Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by:
-
Sakari Ailus authored
Convert drivers using wall clock time (CLOCK_REALTIME) to timestamp from the monotonic timer (CLOCK_MONOTONIC). Signed-off-by:
-
Sakari Ailus authored
v4l2_get_timestamp() produces a monotonic timestamp but unlike ktime_get_ts(), it uses struct timeval instead of struct timespec, saving the drivers the conversion job when getting timestamps for v4l2_buffer's timestamp field. Signed-off-by:
-
Sakari Ailus authored
Define video buffer flags for different timestamp types. Everything up to now have used either realtime clock or monotonic clock, without a way to tell which clock the timestamp was taken from. Also document that the clock source of the timestamp in the timestamp field depends on buffer flags. [mchehab@redhat.com: fix a few wrong references to Kernel 3.8 - as this patch is meant for 3.9] Signed-off-by:
-
Shaik Ameer Basha authored
G-Scaler src buffer count as well as destination buffer count is increased to 32. This is required for G-Scaler to interface with MFC, as MFC demands 32 capture buffers for some H264 streams. Signed-off-by:
Shaik Ameer Basha <shaik.ameer@samsung.com> Signed-off-by:
-
Sachin Kamat authored
devm_clk_get() is a device managed function and makes error handling a bit simpler. Signed-off-by:
-
Sylwester Nawrocki authored
Make sure there is no unbalanced clk_unprepare call and add missing clock release in the driver's remove() callback. Signed-off-by:
Sylwester Nawrocki <sylvester.nawrocki@gmail.com> Signed-off-by:
-
Sachin Kamat authored
In case of clk_prepare failure, the function gsc_clk_get also prints "failed to get clock" which is not correct. Hence move the error messages to their respective blocks. While at it, also renamed the labels meaningfully. Signed-off-by:
-
Sachin Kamat authored
Fixes the following warning: WARNING: space prohibited between function name and open parenthesis '(' FILE: media/platform/exynos-gsc/gsc-m2m.c:606: ctx = kzalloc(sizeof (*ctx), GFP_KERNEL); Signed-off-by: -
Shaik Ameer Basha authored
Make gsc-m2m propagate the timestamp field from source to destination buffers. Signed-off-by:
John Sheu <sheu@google.com> Signed-off-by:
-
Shaik Ameer Basha authored
Adding V4L2_PIX_FMT_NV12MT_16X16 to G-Scaler supported formats. If the output or input format is V4L2_PIX_FMT_NV12MT_16X16, configure G-Scaler to use GSC_IN_TILE_MODE. [s.nawrocki: shortened the pixel format description] Signed-off-by:
-
- 20 Dec, 2012 10 commits
-
-
Fabio Estevam authored
Since commit 43829731 (workqueue: deprecate flush[_delayed]_work_sync()), flush_work() should be used instead of flush_work_sync(). Signed-off-by:
Fabio Estevam <fabio.estevam@freescale.com> Acked-by:
Tejun Heo <tj@kernel.org> Signed-off-by:
-
Mauro Carvalho Chehab authored
drivers/media/usb/tm6000/tm6000-video.c: In function '__check_keep_urb': drivers/media/usb/tm6000/tm6000-video.c:1926:1: warning: return from incompatible pointer type [enabled by default] Signed-off-by: -
Mauro Carvalho Chehab authored
drivers/media/platform/sh_veu.c:269:2: warning: format '%x' expects argument of type 'unsigned int', but argument 5 has type 'dma_addr_t' [-Wformat] drivers/media/platform/sh_veu.c:276:2: warning: format '%x' expects argument of type 'unsigned int', but argument 5 has type 'dma_addr_t' [-Wformat] Cc: Guennadi Liakhovetski <g.liakhovetski@gmx.de> Signed-off-by: -
Javier Martin authored
The CSI on i.MX27 has some constraints regarding image width. This patch makes sure those requirements are met in try_fmt(). Signed-off-by:
Javier Martin <javier.martin@vista-silicon.com> [g.liakhovetski@gmx.de: make constraint i.MX27-specific] Signed-off-by:
Guennadi Liakhovetski <g.liakhovetski@gmx.de> Signed-off-by:
-
Guennadi Liakhovetski authored
sh_vou might be better described by "Odd Fixes," but mark it "Maintained" for now. sh_veu is a new driver and might see some development in the future. Signed-off-by:
-
Guennadi Liakhovetski authored
Use managed allocations to simplify error handling and clean up paths. Signed-off-by:
-
Guennadi Liakhovetski authored
Use managed allocations to simplify error handling and clean up paths. Signed-off-by:
-
Guennadi Liakhovetski authored
Explicit checks for the JPEG pixel format in soc_mbus_bytes_per_line() and soc_mbus_image_size() are superfluous, because also without them these functions will perform correctly. The former will return 0 based on packing == SOC_MBUS_PACKING_VARIABLE and the latter will simply multiply the user-provided line length by the image height to obtain a frame buffer size estimate. The original version of the "media: soc_camera: don't clear pix->sizeimage in JPEG mode" patch was correct and my amendment, adding these two checks was superfluous. Signed-off-by:
-
Guennadi Liakhovetski authored
Update soc-camera documentation to reflect the current camera host API and the use of the common V4L2 subdev API. Signed-off-by:
-
Guennadi Liakhovetski authored
In both error and clean up cases there is no need to wait for runtime PM to finish suspending the device. Signed-off-by:
-
