- 15 Apr, 2024 12 commits
-
-
Ricardo Ribalda authored
Fix smatch error: drivers/media/spi/cxd2880-spi.c:391 cxd2880_start_feed() warn: Please consider using kzalloc instead of kmalloc Signed-off-by:
Ricardo Ribalda <ribalda@chromium.org> Signed-off-by:
Hans Verkuil <hverkuil-cisco@xs4all.nl>
-
Ricardo Ribalda authored
Add missing release_firmware on the error paths. drivers/media/usb/dvb-usb/dib0700_devices.c:2415 stk9090m_frontend_attach() warn: 'state->frontend_firmware' from request_firmware() not released on lines: 2415. drivers/media/usb/dvb-usb/dib0700_devices.c:2497 nim9090md_frontend_attach() warn: 'state->frontend_firmware' from request_firmware() not released on lines: 2489,2497. Signed-off-by:
-
Ricardo Ribalda authored
Do not cache the file, instead load it on demand. This makes smatch a happy parser: drivers/media/dvb-frontends/drx39xyj/drxj.c:11908 drx_ctrl_u_code() warn: 'fw' from request_firmware() not released on lines: 11877,11886,11896. Signed-off-by:
-
Ricardo Ribalda authored
Make sure the firmware is released when we leave xc_load_fw_and_init_tuner() This change makes smatch happy: drivers/media/tuners/xc5000.c:1213 xc_load_fw_and_init_tuner() warn: 'fw' from request_firmware() not released on lines: 1213. Cc: Shuah Khan <shuah.kh@samsung.com> Signed-off-by:
-
Ricardo Ribalda authored
release_firmware() in the same function that it was requested. It is more clear and makes smatch happy. drivers/media/platform/st/sti/c8sectpfe/c8sectpfe-core.c:1146 load_c8sectpfe_fw() warn: 'fw' from request_firmware() not released on lines: 1125,1132. Signed-off-by:
-
Ricardo Ribalda authored
The handled version of request_irq let us remove the free_irq and makes smatch happier: drivers/media/pci/saa7134/saa7134-alsa.c:1186 alsa_card_saa7134_create() warn: 'dev->pci->irq' from request_irq() not released on lines: 1186 Signed-off-by:
-
Ricardo Ribalda authored
Let devm handle the life cycle of the irq request. Makes smatch happier: drivers/media/pci/solo6x10/solo6x10-core.c:631 solo_pci_probe() warn: 'pdev' from pci_request_regions() not released on lines: 631. Signed-off-by:
-
Ricardo Ribalda authored
Instead of handling manually the release of the memory regions let devm do that for us. Makes smatch happy: drivers/media/pci/solo6x10/solo6x10-core.c:631 solo_pci_probe() warn: 'pdev' from pci_request_regions() not released on lines: 631. Signed-off-by:
-
Ricardo Ribalda authored
The structure is packed, which requires that all its fields need to be also packed. ./include/uapi/linux/videodev2.h:1810:2: warning: field within 'struct v4l2_ext_control' is less aligned than 'union v4l2_ext_control::(anonymous at ./include/uapi/linux/videodev2.h:1810:2)' and is usually due to 'struct v4l2_ext_control' being packed, which can lead to unaligned accesses [-Wunaligned-access] Explicitly set the inner union as packed. Marking the inner union as 'packed' does not change the layout, since the whole struct is already packed, it just silences the clang warning. See also this llvm discussion: https://github.com/llvm/llvm-project/issues/55520Signed-off-by:
-
Ricardo Ribalda authored
The structure is packed, which requires that all its fields need to be also packed. ./include/uapi/linux/dvb/frontend.h:854:2: warning: field within 'struct dtv_stats' is less aligned than 'union dtv_stats::(anonymous at ./include/uapi/linux/dvb/frontend.h:854:2)' and is usually due to 'struct dtv_stats' being packed, which can lead to unaligned accesses [-Wunaligned-access] Explicitly set the inner union as packed. Marking the inner union as 'packed' does not change the layout, since the whole struct is already packed, it just silences the clang warning. See also this llvm discussion: https://github.com/llvm/llvm-project/issues/55520Signed-off-by:
-
Ricardo Ribalda authored
This structure is embedded in multiple other structures that are packed, which conflicts with it being aligned. drivers/media/usb/as102/as10x_cmd.h:379:30: warning: field reg_addr within 'struct as10x_dump_memory::(unnamed at drivers/media/usb/as102/as10x_cmd.h:373:2)' is less aligned than 'struct as10x_register_addr' and is usually due to 'struct as10x_dump_memory::(unnamed at drivers/media/usb/as102/as10x_cmd.h:373:2)' being packed, which can lead to unaligned accesses [-Wunaligned-access] Mark it as being packed. Marking the inner struct as 'packed' does not change the layout, since the whole struct is already packed, it just silences the clang warning. See also this llvm discussion: https://github.com/llvm/llvm-project/issues/55520Signed-off-by:
-
Ricardo Ribalda authored
Make dependency on DEBUG_FS conditional, that way we are not forced to enable DEBUG_FS if we can to use this driver. Signed-off-by:
-
- 11 Apr, 2024 11 commits
-
-
Hans de Goede authored
The atomisp driver's Android heritage makes it weird in that even though it uses MC + subdev-s it is designed to primarily be controlled through its /dev/video# node. It implements s_input() on /dev/video# to select which sensor to use, while ignoring link_setup() calls to enable a link to another sensor. Add support for selecting the active sensor the MC way by adding link_setup() support. The implementation is a bit convoluted due to the atomisp driver's heritage. 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>
-
Hans de Goede authored
The link from the ISP's source pad to the /dev/video# output v4l2-dev is always enabled and immutable, mark it as such. Reviewed-by:
Kieran Bingham <kieran.bingham@ideasonboard.com> Reviewed-by:
-
Hans de Goede authored
Change the generic "ATOMISP_SUBDEV" name to "ATOM ISP" to make clear that this is the subdev for the ISP itself. Reviewed-by:
-
Hans de Goede authored
Add DMABUF support and while at it drop userptr support. Now that atomisp has been ported to videobuf2 this is trivial. Reviewed-by:
Kieran Bingham <kieran.bingham+renesas@ideasonboard.com> Reviewed-by:
-
Hans de Goede authored
Stop setting sd->devnode for the ATOMISP_SUBDEV v4l2-subdev, so that a proper /dev/v4l-subdev# gets created for it. Reviewed-by:
-
Hans de Goede authored
Drop all ISP controls and ioctls which just get passed through to the camera subdev. Instead these calls should be done directly at the sensor subdev. Reviewed-by:
-
Hans de Goede authored
Remove the custom VCM handling, instead the VCM should be controlled through its own v4l2-subdev and the new ipu-bridge.c code already supports instantiating an i2c_client for this and setting up the necessary endpoints in the fwnode graph. Reviewed-by:
-
Hans de Goede authored
Rename atomisp_set_crop_and_fmt() to atomisp_set_sensor_crop_and_fmt() to make clear that it operates on the sensor subdev. Reviewed-by:
-
Hans de Goede authored
isp_subdev_propagate() is a wrapper around atomisp_subdev_set_selection() which gets only used in a single place. Call atomisp_subdev_set_selection() directly in that single place instead. Reviewed-by:
-
Zhipeng Lu authored
The allocation failure of mycs->yuv_scaler_binary in load_video_binaries() is followed with a dereference of mycs->yuv_scaler_binary after the following call chain: sh_css_pipe_load_binaries() |-> load_video_binaries(mycs->yuv_scaler_binary == NULL) | |-> sh_css_pipe_unload_binaries() |-> unload_video_binaries() In unload_video_binaries(), it calls to ia_css_binary_unload with argument &pipe->pipe_settings.video.yuv_scaler_binary[i], which refers to the same memory slot as mycs->yuv_scaler_binary. Thus, a null-pointer dereference is triggered. Link: https://lore.kernel.org/r/20240118151303.3828292-1-alexious@zju.edu.cn Fixes: a49d2536 ("staging/atomisp: Add support for the Intel IPU v2") Signed-off-by:Zhipeng Lu <alexious@zju.edu.cn> Reviewed-by:
-
Hans de Goede authored
s/is not unsupported/is not supported/ in the unsupported revision error. Suggested-by:
-
- 10 Apr, 2024 17 commits
-
-
Laurent Pinchart authored
When support for streams was added to the V4L2 subdev API, the v4l2_subdev_crop structure was extended with a stream field, but the field was not handled in the core code that translates the VIDIOC_SUBDEV_[GS]_CROP ioctls to the selection API. Fix it. Fixes: 2f91e10e ("media: subdev: add stream based configuration") Signed-off-by:
Laurent Pinchart <laurent.pinchart@ideasonboard.com> Reviewed-by:
Tomi Valkeinen <tomi.valkeinen@ideasonboard.com> Signed-off-by:
Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by:
-
Tomi Valkeinen authored
The graph walk tries to follow all links, even if they are not between pads. This causes a crash with, e.g. a MEDIA_LNK_FL_ANCILLARY_LINK link. Fix this by allowing the walk to proceed only for MEDIA_LNK_FL_DATA_LINK links. Signed-off-by:
-
Sakari Ailus authored
Remove e-mail addresses no longer in use. Signed-off-by:
-
Sakari Ailus authored
Remove inactive e-mail addresses. Signed-off-by:
-
Sakari Ailus authored
Remove e-mail addresses that are no longer valid. Signed-off-by:
-
Uwe Kleine-König authored
Using __exit for the remove function results in the remove callback being discarded with CONFIG_VIDEO_ET8EK8=y. When such a device gets unbound (e.g. using sysfs or hotplug), the driver is just removed without the cleanup being performed. This results in resource leaks. Fix it by compiling in the remove callback unconditionally. This also fixes a W=1 modpost warning: WARNING: modpost: drivers/media/i2c/et8ek8/et8ek8: section mismatch in reference: et8ek8_i2c_driver+0x10 (section: .data) -> et8ek8_remove (section: .exit.text) Fixes: c5254e72 ("[media] media: Driver for Toshiba et8ek8 5MP sensor") Signed-off-by:
Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Signed-off-by:
-
Bjorn Helgaas authored
ipu-cio2 uses generic power management, and pci_pm_runtime_suspend() and pci_pm_runtime_resume() already take care of setting the PCI device power state, so the driver doesn't need to do it explicitly. Remove explicit setting to D3hot or D0 during runtime suspend and resume. Remove #defines that are no longer used. Signed-off-by:
Bjorn Helgaas <bhelgaas@google.com> Cc: Hyungwoo Yang <hyungwoo.yang@intel.com> Cc: Rajmohan Mani <rajmohan.mani@intel.com> Cc: Vijaykumar Ramya <ramya.vijaykumar@intel.com> Cc: Samu Onkalo <samu.onkalo@intel.com> Cc: Jouni Högander <jouni.hogander@intel.com> Cc: Jouni Ukkonen <jouni.ukkonen@intel.com> Cc: Antti Laakso <antti.laakso@intel.com> [Sakari Ailus: Remove now-redundant local variables.] Signed-off-by:
-
Rajeshwar R Shinde authored
The kernel configuration VIDEO_DEV is defined twice in Kconfig. Thus, the duplicate Kconfig dependency is removed. Signed-off-by:
Rajeshwar R Shinde <coolrrsh@gmail.com> Signed-off-by:
-
Colin Ian King authored
Pointer css_pipe is being assigned a value that is never read afterwards. The assignment is redundant and can be removed. Cleans up clang scan build warning: drivers/staging/media/ipu3/ipu3-css.c:2058:3: warning: Value stored to 'css_pipe' is never read [deadcode.DeadStores] Signed-off-by:
Colin Ian King <colin.i.king@gmail.com> Signed-off-by:
-
Fabio Estevam authored
Document the link-frequencies property as recommended by the following document: https://www.kernel.org/doc/html/v6.9-rc1/driver-api/media/camera-sensor.html#handling-clocksSigned-off-by:
Fabio Estevam <festevam@denx.de> Reviewed-by:
Rob Herring <robh@kernel.org> Signed-off-by:
-
Fabio Estevam authored
The original .txt bindings had the OV2680 power supply names correct, but the transition from .txt to yaml spelled them incorrectly. Fix the OV2680 power supply names as the original .txt bindings as these are the names used by the OV2680 driver and in devicetree. Fixes: 57226cd8 ("media: dt-bindings: ov2680: convert bindings to yaml") Signed-off-by:
-
Fabio Estevam authored
Since commit 63b0cd30 ("media: ov2680: Add bus-cfg / endpoint property verification") the ov2680 no longer probes on a imx7s-warp7: ov2680 1-0036: error -EINVAL: supported link freq 330000000 not found ov2680 1-0036: probe with driver ov2680 failed with error -22 As the 'link-frequencies' property is not mandatory, allow the probe to succeed by skipping the link-frequency verification when the property is absent. Cc: stable@vger.kernel.org Fixes: 63b0cd30 ("media: ov2680: Add bus-cfg / endpoint property verification") Signed-off-by:
-
Fabio Estevam authored
Since commit 63b0cd30 ("media: ov2680: Add bus-cfg / endpoint property verification") even when the correct 'link-frequencies' property is passed in the devicetree, the driver fails to probe: ov2680 1-0036: probe with driver ov2680 failed with error -22 The reason is that the variable 'ret' may contain the -EINVAL value from a previous assignment: ret = fwnode_property_read_u32(dev_fwnode(dev), "clock-frequency", &rate); Fix the problem by clearing 'ret' on the successful path. Tested on imx7s-warp board with the following devicetree: port { ov2680_to_mipi: endpoint { remote-endpoint = <&mipi_from_sensor>; clock-lanes = <0>; data-lanes = <1>; link-frequencies = /bits/ 64 <330000000>; }; }; Cc: stable@vger.kernel.org Fixes: 63b0cd30 ("media: ov2680: Add bus-cfg / endpoint property verification") Suggested-by:
-
Sakari Ailus authored
The driver dug the supported link frequency up from the V4L2 fwnode endpoint and used it internally, but failed to report this in the LINK_FREQ and PIXEL_RATE controls. Fix this. Fixes: 0677a2d9 ("media: ov2740: Add support for 180 MHz link frequency") Cc: stable@vger.kernel.org # for v6.8 and later Signed-off-by:
Bingbu Cao <bingbu.cao@intel.com> Signed-off-by:
-
Sakari Ailus authored
The notifier_entry of a notifier is not re-initialised after unregistering the notifier. This leads to dangling pointers being left there so use list_del_init() to return the notifier_entry an empty list. Fixes: b8ec754a ("media: v4l: async: Set v4l2_device and subdev in async notifier init") Cc: <stable@vger.kernel.org> # for 6.6 and later Signed-off-by:
-
Sakari Ailus authored
The V4L2 device used to be set when the notifier was registered but this has been moved to the notifier initialisation. Don't touch the V4L2 device if registration fails. Fixes: b8ec754a ("media: v4l: async: Set v4l2_device and subdev in async notifier init") Cc: <stable@vger.kernel.org> # for 6.6 and later Signed-off-by:
-
Alexander Stein authored
struct v4l2_async_notifier has several list_head members, but only waiting_list and done_list are initialized. notifier_entry was kept 'zeroed' leading to an uninitialized list_head. This results in a NULL-pointer dereference if csi2_async_register() fails, e.g. node for remote endpoint is disabled, and returns -ENOTCONN. The following calls to v4l2_async_nf_unregister() results in a NULL pointer dereference. Add the missing list head initializer. Fixes: b8ec754a ("media: v4l: async: Set v4l2_device and subdev in async notifier init") Cc: <stable@vger.kernel.org> # for 6.6 and later Signed-off-by:
Alexander Stein <alexander.stein@ew.tq-group.com> Signed-off-by:
-
