- 04 Oct, 2017 40 commits
-
-
Chiranjeevi Rapolu authored
Previously, with crop (0, 0), (4255, 3167), VTS < 0xC9E was resulting in blank frames sometimes. This appeared as video flickering. But we need VTS < 0xC9E to get ~30fps. Omni Vision recommends to use crop (0,8), (4255, 3159) for 4224x3136. With this crop, VTS 0xC8E is supported and yields ~30fps. Signed-off-by:
Chiranjeevi Rapolu <chiranjeevi.rapolu@intel.com> Reviewed-by:
Tomasz Figa <tfiga@chromium.org> Signed-off-by:
Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by:
Mauro Carvalho Chehab <mchehab@osg.samsung.com>
-
Chiranjeevi Rapolu authored
Calculate pixel-rate at run time instead of compile time. Instead of using hardcoded pixels-per-line, extract it from current sensor mode. Signed-off-by:
-
Colin Ian King authored
Don't populate the array reset_seq on the stack, instead make it static. Makes the object code smaller by over 50 bytes: Before: text data bss dec hex filename 11737 6000 64 17801 4589 drivers/media/i2c/ov2640.o After: text data bss dec hex filename 11582 6096 64 17742 454e drivers/media/i2c/ov2640.o Signed-off-by:
Colin Ian King <colin.king@canonical.com> Signed-off-by:
-
Sakari Ailus authored
The active and has_no_links arrays will overrun in media_entity_pipeline_start() if there's an entity which has more than MEDIA_ENTITY_MAX_PAD pads. Ensure in media_entity_init() that there are fewer pads than that. Signed-off-by:
Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by:
-
Markus Elfring authored
Replace the specification of a data structure by pointer dereferences as the parameter for the operator "sizeof" to make size determination a bit safer according to the Linux coding style convention. This issue was detected by using the Coccinelle software. [mchehab@s-opensource.com: merged similar patches] Signed-off-by:
Markus Elfring <elfring@users.sourceforge.net> Signed-off-by:
-
Markus Elfring authored
Omit extra messages for memory allocation failures. This issue was detected by using the Coccinelle software. [mchehab@s-opensource.com: merged similar patches] Signed-off-by:
-
Chiranjeevi Rapolu authored
Previously, the sensor was not streaming after resume from suspend, i.e. on S0->S3->S0 transition. Due to this, camera app preview appeared as stuck. Now, handle streaming state correctly in case of suspend-resume. Signed-off-by:
-
Rajmohan Mani authored
Current v4l2 focus ctrl step value of 16, limits the minimum granularity of focus positions to 16. Setting this value as 1, enables more accurate focus positions. Signed-off-by:
Rajmohan Mani <rajmohan.mani@intel.com> Signed-off-by:
-
Fabio Estevam authored
clk_prepare_enable() may fail, so we should better check its return value and propagate it in the case of error. Signed-off-by:
Fabio Estevam <fabio.estevam@nxp.com> Signed-off-by:
-
Fabio Estevam authored
devm_clk_get() may return different error codes other than -EPROBE_DEFER, so it is better to return the real error code instead. Signed-off-by:
-
Fabio Estevam authored
v4l2_clk_get() may return different error codes other than -EPROBE_DEFER, so it is better to return the real error code instead. Signed-off-by:
-
Chiranjeevi Rapolu authored
Previously, images were relatively darker due to non-optimal settings for black target level and bias. Now, use recommended settings for black target level and output bias as default values. The same default settings apply to all the resolutions. Given these recommeneded settings do not change dynamically, add these to existing mode register settings. Signed-off-by:
-
Hans Verkuil authored
If you have a control handler that does not contain any controls, then currently calling VIDIOC_G/S/TRY_EXT_CTRLS with count == 0 will return -EINVAL in the class_check() function. This is not correct, there is no reason why this should return an error. The purpose of setting count to 0 is to test if the ioctl can mix controls from different control classes. And this is possible. The fact that there are not actually any controls defined is another matter that is unrelated to this test. This caused v4l2-compliance to fail, so that is fixed with this patch applied. Signed-off-by:
Dave Stevenson <dave.stevenson@raspberrypi.org> Tested-by:
-
Hans Verkuil authored
Only send HPD_LOW/HIGH event if the gpio actually changed value. Signed-off-by:
-
Johan Hovold authored
Make sure to check that we actually have an Interface Association Descriptor before dereferencing it during probe to avoid dereferencing a NULL-pointer. Fixes: e0d3bafd ("V4L/DVB (10954): Add cx231xx USB driver") Cc: stable <stable@vger.kernel.org> # 2.6.30 Reported-by:
Andrey Konovalov <andreyknvl@google.com> Signed-off-by:
Johan Hovold <johan@kernel.org> Tested-by:
-
Kees Cook authored
This converts to use setup_timer() to set callback and data, though it doesn't look like this would have worked with timer checking enabled since no init_timer() was ever called before. Cc: Mats Randgaard <matrandg@cisco.com> Cc: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by:
Kees Cook <keescook@chromium.org> Signed-off-by:
-
Hans Verkuil authored
Add support for 10 and 12 bit luma formats. Signed-off-by:
Hans Verkuil <hansverk@cisco.com> Signed-off-by:
-
Hans Verkuil authored
Support the 10 and 12 bit luma formats. Signed-off-by:
-
Arnd Bergmann authored
Older compilers like gcc-4.6 may run into a case that returns an uninitialized variable from rcar_drif_enable_rx() if that function was ever called with an empty cur_ch_mask: drivers/media/platform/rcar_drif.c:658:2: error: ‘ret’ may be used uninitialized in this function [-Werror=uninitialized] Newer compilers don't have that problem as they optimize the 'ret' variable away and just return zero in that case. This changes the function to return -EINVAL for this particular failure, to make it consistent across all compiler versions. In case gcc gets changed to report a warning for it in the future, it's also a good idea to shut it up now. Link: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82203Signed-off-by:
Arnd Bergmann <arnd@arndb.de> Acked-by:
Ramesh Shanmugasundaram <ramesh.shanmugasundaram@bp.renesas.com> Signed-off-by:
-
Bhumika Goyal authored
Make these const as they are not modified in the file referencing them. They are only used when their function pointer fields invokes a function and therefore none of the structure fields are getting modified. Also, add a const to the declaration in the header. Signed-off-by:
Bhumika Goyal <bhumirks@gmail.com> Signed-off-by:
-
Colin Ian King authored
Don't populate const arrays on the stack, instead make them static. Makes the object code smaller by over 5200 bytes: Before: text data bss dec hex filename 58259 8880 128 67267 106c3 ov519.o After: text data bss dec hex filename 52155 9776 128 62059 f26b ov519.o Signed-off-by:
-
Simon Yuan authored
The video standard was not mapped to the corresponding value of the internal video standard in adv748x_afe_querystd, causing the wrong video standard to be selected. Fixes: 3e89586a ("media: i2c: adv748x: add adv748x driver") [Kieran: Obtain the std from the afe->curr_norm] Signed-off-by:
Simon Yuan <simon.yuan@navico.com> Signed-off-by:
Kieran Bingham <kieran.bingham+renesas@ideasonboard.com> Signed-off-by:
-
Christophe JAILLET authored
If this memory allocation fails, we must release some resources, as already done in the code below and above. Signed-off-by:
Christophe JAILLET <christophe.jaillet@wanadoo.fr> Signed-off-by:
-
Markus Elfring authored
Replace the specification of data structures by pointer dereferences as the parameter for the operator "sizeof" to make the corresponding size determination a bit safer according to the Linux coding style convention. This issue was detected by using the Coccinelle software. Signed-off-by:
Sean Young <sean@mess.org> Signed-off-by:
-
Marc Gonzalez authored
ir_setkeytable() and ir_create_table() print the same debug message. Delete the one in ir_setkeytable() Signed-off-by:
Marc Gonzalez <marc_gonzalez@sigmadesigns.com> Signed-off-by:
-
Stephen Hemminger authored
The Linus policy on Kconfig is that the default should be no for all new devices. I.e the user rebuild a new kernel from an old config should not by default get a larger kernel. Fixes: b4c184e5 ("[media] media: reorganize the main Kconfig items") Signed-off-by:
Stephen Hemminger <sthemmin@microsoft.com> Acked-by:
Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by:
-
Thomas Meyer authored
Replace self coded binary search, by existing library version. Signed-off-by:
Thomas Meyer <thomas@m3y3r.de> Signed-off-by:
-
Ladislav Michl authored
Gpiolib API is preferred way to access gpios. Signed-off-by:
Ladislav Michl <ladis@linux-mips.org> Signed-off-by:
-
Ladislav Michl authored
gpio_ir_recv_platform_data are not used anywhere in kernel tree, so remove it. Signed-off-by:
-
Ladislav Michl authored
There already is standard macro providing driver name, use it. Signed-off-by:
-
Ladislav Michl authored
Use of devm_request_irq simplifies error unwinding and as free_irq was the last user of driver remove function, remove it too. Signed-off-by:
-
Ladislav Michl authored
Requesting any context irq is not actually great idea since threaded interrupt handler is run at too unpredictable time which turns timing information wrong. Fix it by requesting regular interrupt. Signed-off-by:
-
Ladislav Michl authored
Use of devm_rc_register_device simplifies error unwinding. Signed-off-by:
-
Ladislav Michl authored
Use of devm_gpio_request_one simplifies error unwinding. Signed-off-by:
-
Ladislav Michl authored
Use of devm_rc_allocate_device simplifies error unwinding. Signed-off-by:
-
Ladislav Michl authored
Use of devm_kzalloc simplifies error unwinding. Signed-off-by:
-
Ladislav Michl authored
Using explicit struct device variable makes code a bit more readable. Signed-off-by:
-
Colin Ian King authored
Don't populate the const arrays vfd_packet6 and fp_packet on the stack, instead make them static. Makes the object code smaller by over 600 bytes: Before: text data bss dec hex filename 43794 17920 1024 62738 f512 drivers/media/rc/imon.o After: text data bss dec hex filename 42994 18080 1024 62098 f292 drivers/media/rc/imon.o Signed-off-by:
-
Sean Young authored
Only the nec protocol is understood, but then it doesn't pass on the full scancode and it ignores the nec repeats its own remote sends, so holding buttons does not work. Signed-off-by:
-
Sean Young authored
This name is also stored in the input's device name, but that is not available in TX only hardware (no input device). Signed-off-by:
-
