linux.git/drivers/staging/media, branch v6.6

staging: media: tegra-video: fix infinite recursion regression

2023-10-02T08:02:53+00:00

Since commit 9bf19fbf0c8b ("media: v4l: async: Rework internal lists"), aka
v6.6-rc1~97^2~198, probing the tegra-video VI driver causes infinite
recursion due tegra_vi_graph_parse_one() calling itself until:

[    1.571168] Insufficient stack space to handle exception!
...
[    1.591416] Internal error: kernel stack overflow: 0 [#1] PREEMPT SMP ARM
...
[    3.861013]  of_phandle_iterator_init from __of_parse_phandle_with_args+0x40/0xf0
[    3.868497]  __of_parse_phandle_with_args from of_fwnode_graph_get_remote_endpoint+0x68/0xa8
[    3.876938]  of_fwnode_graph_get_remote_endpoint from fwnode_graph_get_remote_port_parent+0x30/0x7c
[    3.885984]  fwnode_graph_get_remote_port_parent from tegra_vi_graph_parse_one+0x7c/0x224
[    3.894158]  tegra_vi_graph_parse_one from tegra_vi_graph_parse_one+0x144/0x224
[    3.901459]  tegra_vi_graph_parse_one from tegra_vi_graph_parse_one+0x144/0x224
[    3.908760]  tegra_vi_graph_parse_one from tegra_vi_graph_parse_one+0x144/0x224
[    3.916061]  tegra_vi_graph_parse_one from tegra_vi_graph_parse_one+0x144/0x224
...
[    4.857892]  tegra_vi_graph_parse_one from tegra_vi_graph_parse_one+0x144/0x224
[    4.865193]  tegra_vi_graph_parse_one from tegra_vi_graph_parse_one+0x144/0x224
[    4.872494]  tegra_vi_graph_parse_one from tegra_vi_init+0x574/0x6d4
[    4.878842]  tegra_vi_init from host1x_device_init+0x84/0x15c
[    4.884594]  host1x_device_init from host1x_video_probe+0xa0/0x114
[    4.890770]  host1x_video_probe from really_probe+0xe0/0x400

The reason is the mentioned commit changed tegra_vi_graph_find_entity() to
search for an entity in the done notifier list:

> @@ -1464,7 +1464,7 @@ tegra_vi_graph_find_entity(struct tegra_vi_channel *chan,
>  	struct tegra_vi_graph_entity *entity;
>  	struct v4l2_async_connection *asd;
>
> -	list_for_each_entry(asd, &chan->notifier.asc_list, asc_entry) {
> +	list_for_each_entry(asd, &chan->notifier.done_list, asc_entry) {
>  		entity = to_tegra_vi_graph_entity(asd);
>  		if (entity->asd.match.fwnode == fwnode)
>  			return entity;

This is not always correct, being tegra_vi_graph_find_entity() called in
three locations, in this order:

 1. tegra_vi_graph_parse_one()    -- called while probing
 2. tegra_vi_graph_notify_bound() -- the .bound notifier op
 3. tegra_vi_graph_build()        -- called in the .complete notifier op

Locations 1 and 2 are called before moving the entity from waiting_list to
done_list, thus they won't find what they are looking for in
done_list. Location 3 happens afterwards and thus it is not broken, however
it means tegra_vi_graph_find_entity() should not search in the same list
every time.

The error appears at step 1: tegra_vi_graph_parse_one() iterates
recursively until it finds the entity already notified, which now never
happens.

Fix by passing the specific notifier list pointer to
tegra_vi_graph_find_entity() instead of the channel, so each caller can
search in whatever list is correct.

Also improve the tegra_vi_graph_find_entity() comment.

Fixes: 9bf19fbf0c8b ("media: v4l: async: Rework internal lists")
Cc: Thierry Reding 
Cc: Jonathan Hunter 
Cc: Sowjanya Komatineni 
Signed-off-by: Luca Ceresoli 
[Sakari Ailus: Wrapped some long lines.]
Signed-off-by: Sakari Ailus 
Signed-off-by: Hans Verkuil

media: ipu-bridge: Fix Kconfig dependencies

2023-10-02T08:02:53+00:00

The current Kconfig symbol dependencies allow having the following Kconfig
symbol values

	CONFIG_IPU_BRIDGE=m
	CONFIG_VIDEO_IPU3_CIO2=y
	CONFIG_CIO2_BRIDGE=n

This does not work as the IPU bridge API is conditional to
IS_ENABLED(CONFIG_IPU_BRIDGE). Fix this by changing the dependencies so
that CONFIG_IPU_BRIDGE can be separately selected.

The CONFIG_CIO2_BRIDGE symbol becomes redundant as a result and is
removed.

Fixes: 2545a2c02ba1 ("media: ipu3-cio2: allow ipu_bridge to be a module again")
Suggested-by: Arnd Bergmann 
Signed-off-by: Sakari Ailus 
Signed-off-by: Hans Verkuil

media: dvb: symbol fixup for dvb_attach()

2023-09-09T07:15:11+00:00

In commit 9011e49d54dc ("modules: only allow symbol_get of
EXPORT_SYMBOL_GPL modules") the use of symbol_get is properly restricted
to GPL-only marked symbols.  This interacts oddly with the DVB logic
which only uses dvb_attach() to load the dvb driver which then uses
symbol_get().

Fix this up by properly marking all of the dvb_attach attach symbols as
EXPORT_SYMBOL_GPL().

Fixes: 9011e49d54dc ("modules: only allow symbol_get of EXPORT_SYMBOL_GPL modules")
Cc: stable 
Reported-by: Stefan Lippers-Hollmann 
Cc: Mauro Carvalho Chehab 
Cc: Christoph Hellwig 
Cc: linux-media@vger.kernel.org
Cc: linux-modules@vger.kernel.org
Acked-by: Luis Chamberlain 
Acked-by: Hans Verkuil 
Link: https://lore.kernel.org/r/20230908092035.3815268-2-gregkh@linuxfoundation.org
Signed-off-by: Greg Kroah-Hartman

Merge tag 'media/v6.6-1' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media

2023-09-01T19:21:32+00:00

Pull media updates from Mauro Carvalho Chehab:

 - new i2c drivers: ds90ub913, ds90ub953, ds90ub960, dw9719, ds90ub913

 - new Intel IVSC MEI drivers

 - some Mediatek platform drivers were moved to a common location

 - Intel atomisp2 driver is now working with the main ov2680 driver. Due
   to that, the atomisp2 ov2680 staging one was removed

 - the bttv driver was finally converted to videobuf2 framework. This
   was the last one upstream using videobuf version 1 core. We'll likely
   remove the old videobuf framework on 6.7

 - lots of improvements at atomisp driver: it now works with normal I2C
   sensors. Several compile-mode dependecies to select between ISP2400
   and ISP2401 are now solved in runtime

 - a new ipu-bridge logic was added to work with IVSC MEI drivers

 - venus driver gained better support for new VPU versions

 - the v4l core async framework has gained lots of improvements and
   cleanups

 - lots of other cleanups, improvements and driver fixes

* tag 'media/v6.6-1' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media: (358 commits)
  media: ivsc: Add ACPI dependency
  media: bttv: convert to vb2
  media: bttv: use audio defaults for winfast2000
  media: bttv: refactor bttv_set_dma()
  media: bttv: move vbi_skip/vbi_count out of buffer
  media: bttv: remove crop info from bttv_buffer
  media: bttv: remove tvnorm field from bttv_buffer
  media: bttv: remove format field from bttv_buffer
  media: bttv: move do_crop flag out of bttv_fh
  media: bttv: copy vbi_fmt from bttv_fh
  media: bttv: copy vid fmt/width/height from fh
  media: bttv: radio use v4l2_fh instead of bttv_fh
  media: bttv: replace BUG with WARN_ON
  media: bttv: use video_drvdata to get bttv
  media: i2c: rdacm21: Fix uninitialized value
  media: coda: Remove duplicated include
  media: vivid: fix the racy dev->radio_tx_rds_owner
  media: i2c: ccs: Check rules is non-NULL
  media: i2c: ds90ub960: Fix PLL config for 1200 MHz CSI rate
  media: i2c: ds90ub953: Fix use of uninitialized variables
  ...

media: atomisp: Drop atomisp-ov2680 sensor driver

2023-08-14T18:27:58+00:00

After recent improvements to atomisp and the standard ov2680
sensor driver, the atomisp driver now works fine with
the standard ov2680 driver.

Drop the no longer necessary atomisp specific atomisp-ov2680
sensor driver.

Reviewed-by: Andy Shevchenko 
Signed-off-by: Hans de Goede 
Signed-off-by: Sakari Ailus 
Signed-off-by: Mauro Carvalho Chehab

media: atomisp: csi2-bridge: Add support for VCM I2C-client instantiation

2023-08-10T05:58:41+00:00

Fill sensor->vcm_type and call intel_cio2_bridge_instantiate_vcm() from
the v4l2-async bound op so that an I2C-client will be instatiated for
the VCM.

Note unfortunately on atomisp the _DSM to get the VCM type sometimes
returns a VCM even though there is none. Since VCMs are typically only
used together with certain sensors, work around this by adding a vcm
field to atomisp_sensor_config and only check for a VCM when that is set.

Signed-off-by: Hans de Goede 
Signed-off-by: Sakari Ailus 
Signed-off-by: Mauro Carvalho Chehab

media: atomisp: csi2-bridge: Add dev_name() to acpi_handle_info() logging

2023-08-10T05:58:41+00:00

acpi_handle_info() uses the ACPI path to the handle as prefix for messages
e.g. : "\_SB_.I2C2.CAM8".

This makes it hard for users to figure out which csi2-bridge messages
belong to which sensor since the actual sensor drivers uses the ACPI
device name (typically "HID:00") for logging.

Extend the acpi_handle_info() (and err and warn) logging to also log
the device name to make it easier to match csi2-bridge messages with
sensor driver log messages.

Suggested-by: Sakari Ailus 
Signed-off-by: Hans de Goede 
Reviewed-by: Andy Shevchenko 
Signed-off-by: Sakari Ailus 
Signed-off-by: Mauro Carvalho Chehab

media: atomisp: csi2-bridge: Switch to new common ipu_bridge_init()

2023-08-10T05:58:40+00:00

Remove the duplicate IPU ACPI bridge code and use the new
shared ipu_bridge_init() functionality.

Note this will also use / assume v4l2-async device instantiation for
ov5693 sensors on atomisp devices since ipu_supported_sensors[]
already contains a match for this.

This is fine since recent atomisp improvements allow the atomisp code
to work with generic v4l2 sensor drivers and using an unmodified
drivers/media/i2c/ov5693.c has been successfully tested on
an Acer Iconia W4 820 tablet with an ISP2400 + OV5693 sensor.

Reviewed-by: Andy Shevchenko 
Reviewed-by: Daniel Scally 
Signed-off-by: Hans de Goede 
Signed-off-by: Sakari Ailus 
Signed-off-by: Mauro Carvalho Chehab

media: atomisp: ov2680: Convert to new CCI register access helpers

2023-08-10T05:58:39+00:00

Use the new comon CCI register access helpers to replace the private
register access helpers in the ov2680 driver.

While at it also switch to using the same register address defines
as the standard drivers/media/i2c/ov2680.c driver to make merging
the 2 drivers simpler.

Reviewed-by: Laurent Pinchart 
Signed-off-by: Hans de Goede 
Signed-off-by: Sakari Ailus 
Signed-off-by: Mauro Carvalho Chehab

staging: media: imx: Merge VIDEO_IMX_CSI into VIDEO_IMX_MEDIA

2023-08-10T05:58:37+00:00

The VIDEO_IMX_MEDIA Kconfig symbol used to select helpers shared between
independent drivers for different i.MX SoCs, and VIDEO_IMX_MEDIA then
selects drivers specific to the i.MX5 and i.MX6. Now that i.MX7 and
i.MX8 support has moved to drivers/media/ and doesn't depend on
VIDEO_IMX_CSI, there's no need to have separate Kconfig options. Merge
VIDEO_IMX_CSI into VIDEO_IMX_MEDIA.

Signed-off-by: Laurent Pinchart 
Signed-off-by: Hans Verkuil