platform/kernel/linux-rpi.git
18 months agomedia: venc: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:13 +0000 (16:31 +0200)]
media: venc: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: vdec: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:12 +0000 (16:31 +0200)]
media: vdec: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: venus: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:11 +0000 (16:31 +0200)]
media: venus: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: venus: Warn only once about problems in .remove()
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:10 +0000 (16:31 +0200)]
media: venus: Warn only once about problems in .remove()

The only effect of returning an error code in a remove callback is that
the driver core emits a warning. The device is unbound anyhow.

As the remove callback already emits a (quite verbose) warning when ret
is non-zero, return zero to suppress the additional warning.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: camss: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:09 +0000 (16:31 +0200)]
media: camss: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: mx2_emmaprp: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:08 +0000 (16:31 +0200)]
media: mx2_emmaprp: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: imx7-media-csi: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:07 +0000 (16:31 +0200)]
media: imx7-media-csi: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: imx-pxp: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:06 +0000 (16:31 +0200)]
media: imx-pxp: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: imx-mipi-csis: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:05 +0000 (16:31 +0200)]
media: imx-mipi-csis: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: mxc-jpeg: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:04 +0000 (16:31 +0200)]
media: mxc-jpeg: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: dw100: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:03 +0000 (16:31 +0200)]
media: dw100: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Xavier Roumegue <xavier.roumegue@oss.nxp.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: vde: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:02 +0000 (16:31 +0200)]
media: vde: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: microchip-sama7g5-isc: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:01 +0000 (16:31 +0200)]
media: microchip-sama7g5-isc: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: microchip-sama5d2-isc: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:00 +0000 (16:31 +0200)]
media: microchip-sama5d2-isc: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: microchip-csi2dc: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:59 +0000 (16:30 +0200)]
media: microchip-csi2dc: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: mtk_vpu: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:58 +0000 (16:30 +0200)]
media: mtk_vpu: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: mtk_vcodec_enc_drv: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:57 +0000 (16:30 +0200)]
media: mtk_vcodec_enc_drv: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: mtk_vcodec_dec_drv: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:56 +0000 (16:30 +0200)]
media: mtk_vcodec_dec_drv: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: mtk-mdp3-core: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:55 +0000 (16:30 +0200)]
media: mtk-mdp3-core: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: mtk_mdp_core: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:54 +0000 (16:30 +0200)]
media: mtk_mdp_core: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: mtk_jpeg_core: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:53 +0000 (16:30 +0200)]
media: mtk_jpeg_core: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: marvell: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:52 +0000 (16:30 +0200)]
media: marvell: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: marvell: Simplify remove callback
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:51 +0000 (16:30 +0200)]
media: marvell: Simplify remove callback

Fold mmpcam_remove() into its only caller mmpcam_platform_remove(). Note
that cam can never be zero, as the probe function calls
platform_set_drvdata() with a non-NULL argument (or returns an error
code in which case .remove() won't be called).

Also use mmpcam_remove() as the function name for the remove callback to
align to mmpcam_probe() for .probe().

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: m2m-deinterlace: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:50 +0000 (16:30 +0200)]
media: m2m-deinterlace: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: pxa_camera: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:49 +0000 (16:30 +0200)]
media: pxa_camera: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: coda-common: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:48 +0000 (16:30 +0200)]
media: coda-common: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: cdns-csi2tx: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:47 +0000 (16:30 +0200)]
media: cdns-csi2tx: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: cdns-csi2rx: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:46 +0000 (16:30 +0200)]
media: cdns-csi2rx: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: atmel-isi: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:45 +0000 (16:30 +0200)]
media: atmel-isi: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: aspeed-video: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:44 +0000 (16:30 +0200)]
media: aspeed-video: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: vpu_drv: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:43 +0000 (16:30 +0200)]
media: vpu_drv: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: vpu_core: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:42 +0000 (16:30 +0200)]
media: vpu_core: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: ge2d: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:41 +0000 (16:30 +0200)]
media: ge2d: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: allegro-core: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:40 +0000 (16:30 +0200)]
media: allegro-core: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: zd1301_demod: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:39 +0000 (16:30 +0200)]
media: zd1301_demod: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: rtl2832_sdr: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:38 +0000 (16:30 +0200)]
media: rtl2832_sdr: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: tegra_cec: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:37 +0000 (16:30 +0200)]
media: tegra_cec: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: stm32-cec: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:36 +0000 (16:30 +0200)]
media: stm32-cec: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: stih-cec: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:35 +0000 (16:30 +0200)]
media: stih-cec: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: seco-cec: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:34 +0000 (16:30 +0200)]
media: seco-cec: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: s5p_cec: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:33 +0000 (16:30 +0200)]
media: s5p_cec: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: ao-cec: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:32 +0000 (16:30 +0200)]
media: ao-cec: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: ao-cec-g12a: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:31 +0000 (16:30 +0200)]
media: ao-cec-g12a: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: cros-ec-cec: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:30 +0000 (16:30 +0200)]
media: cros-ec-cec: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Guenter Roeck <groeck@chromium.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: cros-ec-cec: Don't exit early in .remove() callback
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:29 +0000 (16:30 +0200)]
media: cros-ec-cec: Don't exit early in .remove() callback

Exiting early in remove without releasing all acquired resources yields
leaks. Note that e.g. memory allocated with devm_zalloc() is freed after
.remove() returns, even if the return code was negative.

While blocking_notifier_chain_unregister() won't fail and so the
change is somewhat cosmetic, platform driver's .remove callbacks are
about to be converted to return void. To prepare that, keep the error
message but don't return early.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: cec-gpio: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:30:26 +0000 (16:30 +0200)]
media: cec-gpio: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: rkvdec: fix use after free bug in rkvdec_remove
Zheng Wang [Mon, 13 Mar 2023 16:42:20 +0000 (16:42 +0000)]
media: rkvdec: fix use after free bug in rkvdec_remove

In rkvdec_probe, rkvdec->watchdog_work is bound with
rkvdec_watchdog_func. Then rkvdec_vp9_run may
be called to start the work.

If we remove the module which will call rkvdec_remove
 to make cleanup, there may be a unfinished work.
 The possible sequence is as follows, which will
 cause a typical UAF bug.

Fix it by canceling the work before cleanup in rkvdec_remove.

CPU0                  CPU1

                    |rkvdec_watchdog_func
rkvdec_remove       |
 rkvdec_v4l2_cleanup|
  v4l2_m2m_release  |
    kfree(m2m_dev); |
                    |
                    | v4l2_m2m_get_curr_priv
                    |   m2m_dev->curr_ctx //use

Fixes: cd33c830448b ("media: rkvdec: Add the rkvdec driver")
Signed-off-by: Zheng Wang <zyytlz.wz@163.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: cedrus: fix use after free bug in cedrus_remove due to race condition
Zheng Wang [Mon, 13 Mar 2023 16:31:20 +0000 (16:31 +0000)]
media: cedrus: fix use after free bug in cedrus_remove due to race condition

In cedrus_probe, dev->watchdog_work is bound with cedrus_watchdog function.
In cedrus_device_run, it will started by schedule_delayed_work. If there is
an unfinished work in cedrus_remove, there may be a race condition and
trigger UAF bug.

CPU0                  CPU1

                    |cedrus_watchdog
cedrus_remove       |
  v4l2_m2m_release  |
  kfree(m2m_dev)    |
                    |
                    | v4l2_m2m_get_curr_priv
                    |   m2m_dev //use

Fix it by canceling the worker in cedrus_remove.

Fixes: 7c38a551bda1 ("media: cedrus: Add watchdog for job completion")
Signed-off-by: Zheng Wang <zyytlz.wz@163.com>
Acked-by: Jernej Skrabec <jernej.skrabec@gmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: verisilicon: VP9: Only propose 10 bits compatible pixels formats
Benjamin Gaignard [Mon, 20 Feb 2023 10:48:49 +0000 (10:48 +0000)]
media: verisilicon: VP9: Only propose 10 bits compatible pixels formats

When decoding a 10bits bitstreams VP9 driver should only expose
10bits pixel formats.
To fulfill this requirement it is needed to call hantro_reset_raw_fmt()
when bit depth change and to correctly set match_depth in pixel formats
enumeration.

Signed-off-by: Benjamin Gaignard <benjamin.gaignard@collabora.com>
Reviewed-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: verisilicon: HEVC: Only propose 10 bits compatible pixels formats
Benjamin Gaignard [Mon, 20 Feb 2023 10:48:48 +0000 (10:48 +0000)]
media: verisilicon: HEVC: Only propose 10 bits compatible pixels formats

When decoding a 10bits bitstreams HEVC driver should only expose
10bits pixel formats.
To fulfill this requirement it is needed to call hantro_reset_raw_fmt()
when bit depth change and to correctly set match_depth in pixel formats
enumeration.

Signed-off-by: Benjamin Gaignard <benjamin.gaignard@collabora.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Reviewed-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: verisilicon: Do not change context bit depth before validating the format
Benjamin Gaignard [Mon, 20 Feb 2023 10:48:47 +0000 (10:48 +0000)]
media: verisilicon: Do not change context bit depth before validating the format

It is needed to check if the proposed pixels format is valid before
updating context bit depth and other internal states.
Stop using ctx->bit_depth to check format depth match and return
result to the caller.

Signed-off-by: Benjamin Gaignard <benjamin.gaignard@collabora.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Reviewed-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: verisilicon: Do not set ctx->bit_depth in hantro_try_ctrl()
Benjamin Gaignard [Mon, 20 Feb 2023 10:48:46 +0000 (10:48 +0000)]
media: verisilicon: Do not set ctx->bit_depth in hantro_try_ctrl()

In hantro_try_ctrl() we should only check the values inside
control parameters and not set ctx->bit_depth. That must
be done in controls set function.
Create a set control function for hevc where ctx->bit_depth is
set at the right time.

Signed-off-by: Benjamin Gaignard <benjamin.gaignard@collabora.com>
Reviewed-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: verisilicon: Do not use ctx fields as format storage when resetting
Benjamin Gaignard [Mon, 20 Feb 2023 10:48:45 +0000 (10:48 +0000)]
media: verisilicon: Do not use ctx fields as format storage when resetting

Source and destination pixel formats fields of context structure should
not be used as storage when resetting the format.
Use local variables instead and let hantro_set_fmt_out() and
hantro_set_fmt_cap() set them correctly later.

Signed-off-by: Benjamin Gaignard <benjamin.gaignard@collabora.com>
Reviewed-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: verisilicon: Do not set context src/dst formats in reset functions
Benjamin Gaignard [Mon, 20 Feb 2023 10:48:44 +0000 (10:48 +0000)]
media: verisilicon: Do not set context src/dst formats in reset functions

Setting context source and destination formats should only be done
in hantro_set_fmt_out() and hantro_set_fmt_cap() after check that
the targeted queue is not busy.
Remove these calls from hantro_reset_encoded_fmt() and
hantro_reset_raw_fmt() to clean the driver.

Signed-off-by: Benjamin Gaignard <benjamin.gaignard@collabora.com>
Reviewed-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mediatek: vcodec: change lat thread decode error condition
Yunfei Dong [Wed, 1 Feb 2023 07:33:16 +0000 (07:33 +0000)]
media: mediatek: vcodec: change lat thread decode error condition

If lat thread can't get lat buffer, it should be that current instance
don't be schedulded, the driver can't free the src buffer directly.

Fixes: 7b182b8d9c85 ("media: mediatek: vcodec: Refactor get and put capture buffer flow")
Signed-off-by: Yunfei Dong <yunfei.dong@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mediatek: vcodec: making sure queue_work successfully
Yunfei Dong [Wed, 1 Feb 2023 07:33:15 +0000 (07:33 +0000)]
media: mediatek: vcodec: making sure queue_work successfully

Putting core work to work queue using queue_work maybe fail, call
queue_work again when the count of core work in work queue is less
than core_list_cnt, making sure all the buffer in core list can be
scheduled.

Fixes: 365e4ba01df4 ("media: mtk-vcodec: Add work queue for core hardware decode")
Signed-off-by: Yunfei Dong <yunfei.dong@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mediatek: vcodec: remove unused lat_buf
Yunfei Dong [Wed, 1 Feb 2023 07:33:14 +0000 (07:33 +0000)]
media: mediatek: vcodec: remove unused lat_buf

Remove unused lat_buf from core list, or leading to core list access
NULL point.

Fixes: 365e4ba01df4 ("media: mtk-vcodec: Add work queue for core hardware decode")
Signed-off-by: Yunfei Dong <yunfei.dong@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mediatek: vcodec: add core decode done event
Yunfei Dong [Wed, 1 Feb 2023 07:33:13 +0000 (07:33 +0000)]
media: mediatek: vcodec: add core decode done event

Need to make sure core decode done before current instance is free.

Fixes: 365e4ba01df4 ("media: mtk-vcodec: Add work queue for core hardware decode")
Signed-off-by: Yunfei Dong <yunfei.dong@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mediatek: vcodec: move lat_buf to the top of core list
Yunfei Dong [Wed, 1 Feb 2023 07:33:12 +0000 (07:33 +0000)]
media: mediatek: vcodec: move lat_buf to the top of core list

Current instance will decode done when begin to wait lat buf full,
move the lat_buf of current instance to the top of core list to make
sure current instance's lat_buf will be used firstly.

Fixes: 365e4ba01df4 ("media: mtk-vcodec: Add work queue for core hardware decode")
Signed-off-by: Yunfei Dong <yunfei.dong@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mediatek: vcodec: using each instance lat_buf count replace core ready list
Yunfei Dong [Wed, 1 Feb 2023 07:33:11 +0000 (07:33 +0000)]
media: mediatek: vcodec: using each instance lat_buf count replace core ready list

Core Hardware decoder depends on each instance lat_buf count,
calling queue_work decode again when the lat_buf count of each instance
isn't zero.

Fixes: 365e4ba01df4 ("media: mtk-vcodec: Add work queue for core hardware decode")
Signed-off-by: Yunfei Dong <yunfei.dong@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mediatek: vcodec: add params to record lat and core lat_buf count
Yunfei Dong [Wed, 1 Feb 2023 07:33:10 +0000 (07:33 +0000)]
media: mediatek: vcodec: add params to record lat and core lat_buf count

Using lat_buf to share decoder information between lat and core work
queue, adding params to record the buf count.

Fixes: 365e4ba01df4 ("media: mtk-vcodec: Add work queue for core hardware decode")
Signed-off-by: Yunfei Dong <yunfei.dong@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: amphion: support to decode RealVideo video
Ming Qian [Tue, 17 Jan 2023 02:31:55 +0000 (02:31 +0000)]
media: amphion: support to decode RealVideo video

amphion decoder add support for the RealVideo video.
RV30 and RV40 are support by amphion decoder.

Signed-off-by: Ming Qian <ming.qian@nxp.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: add RealVideo format RV30 and RV40
Ming Qian [Tue, 17 Jan 2023 02:31:54 +0000 (02:31 +0000)]
media: add RealVideo format RV30 and RV40

RealVideo, or also spelled as Real Video, is a suite of proprietary
video compression formats developed by RealNetworks -
the specific format changes with the version.
RealVideo codecs are identified by four-character codes.
RV30 and RV40 are RealNetworks' proprietary H.264-based codecs.

Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Ming Qian <ming.qian@nxp.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: amphion: support to decode sorenson spark video
Ming Qian [Thu, 12 Jan 2023 09:04:48 +0000 (09:04 +0000)]
media: amphion: support to decode sorenson spark video

Sorenson Spark is an implementation of H.263 for use
in Flash Video and Adobe Flash files.
amphion decoder can support it by insert some startcode
before sequence and picture.

Signed-off-by: Ming Qian <ming.qian@nxp.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: add Sorenson Spark video format
Ming Qian [Thu, 12 Jan 2023 09:04:47 +0000 (09:04 +0000)]
media: add Sorenson Spark video format

Sorenson Spark is an implementation of H.263 for use
in Flash Video and Adobe Flash files.
Sorenson Spark is an incomplete implementation of H.263.
It differs mostly in header structure and ranges of the coefficients.

Signed-off-by: Ming Qian <ming.qian@nxp.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mediatek: vcodec: Force capture queue format to MM21
Yunfei Dong [Sat, 18 Mar 2023 03:56:51 +0000 (03:56 +0000)]
media: mediatek: vcodec: Force capture queue format to MM21

While the decoder can produce frames in both MM21 and MT21C formats, only
MM21 is currently supported by userspace tools (like gstreamer and libyuv).
In order to ensure userspace keeps working after the SCP firmware is
updated to support both MM21 and MT21C formats, force the MM21 format for
the capture queue.

This is meant as a stopgap solution while dynamic format switching between
MM21 and MT21C isn't implemented in the driver.

Fixes: 7501edef6b1f ("media: mediatek: vcodec: Different codec using different capture format")
Signed-off-by: Yunfei Dong <yunfei.dong@mediatek.com>
Reviewed-by: Nicolas F. R. A. Prado <nfraprado@collabora.com>
Tested-by: Nicolas F. R. A. Prado <nfraprado@collabora.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mediatek: vcodec: Make MM21 the default capture format
Yunfei Dong [Sat, 18 Mar 2023 03:56:50 +0000 (03:56 +0000)]
media: mediatek: vcodec: Make MM21 the default capture format

Given that only the MM21 capture format is supported by userspace tools
(like gstreamer and libyuv), make it the default capture format.

This allows us to force the MM21 format even when a MM21 and MT21C capable
firmware is available (which is needed while dynamic format switching isn't
implemented in the driver), without causing the following regressions on
v4l2-compliance:

        fail: v4l2-test-formats.cpp(478): pixelformat 3132544d (MT21) for buftype 9 not reported by ENUM_FMT
    test VIDIOC_G_FMT: FAIL
        fail: v4l2-test-formats.cpp(478): pixelformat 3132544d (MT21) for buftype 9 not reported by ENUM_FMT
    test VIDIOC_TRY_FMT: FAIL
        fail: v4l2-test-formats.cpp(478): pixelformat 3132544d (MT21) for buftype 9 not reported by ENUM_FMT
    test VIDIOC_S_FMT: FAIL

Fixes: 7501edef6b1f ("media: mediatek: vcodec: Different codec using different capture format")
Signed-off-by: Yunfei Dong <yunfei.dong@mediatek.com>
Reviewed-by: Nicolas F. R. A. Prado <nfraprado@collabora.com>
Tested-by: Nicolas F. R. A. Prado <nfraprado@collabora.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mediatek: vcodec: Use 4K frame size when supported by stateful decoder
Pin-yen Lin [Tue, 14 Mar 2023 10:22:41 +0000 (10:22 +0000)]
media: mediatek: vcodec: Use 4K frame size when supported by stateful decoder

After commit b018be06f3c7 ("media: mediatek: vcodec: Read max resolution
from dec_capability"), the stateful video decoder driver never really
sets its output frame size to 4K.

Parse the decoder capability reported by the firmware, and update the
output frame size in mtk_init_vdec_params to enable 4K frame size when
available.

Fixes: b018be06f3c7 ("media: mediatek: vcodec: Read max resolution from dec_capability")
Signed-off-by: Pin-yen Lin <treapking@chromium.org>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: dt-bindings: media: mediatek: vcodec: Change the max reg value to 2
Yunfei Dong [Fri, 3 Mar 2023 01:38:41 +0000 (01:38 +0000)]
media: dt-bindings: media: mediatek: vcodec: Change the max reg value to 2

Need to add racing control register base in device node for mt8195 support
inner racing mode. Changing the max reg value from 1 to 2.

Adding the description for VDEC_SYS and VDEC_MISC.

Signed-off-by: Yunfei Dong <yunfei.dong@mediatek.com>
Signed-off-by: Allen-KH Cheng <allen-kh.cheng@mediatek.com>
Acked-by: Rob Herring <robh@kernel.org>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Chen-Yu Tsai <wenst@chromium.org>
Tested-by: Chen-Yu Tsai <wenst@chromium.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: dt-bindings: media: mediatek: vcodec: adapt to the 'clock-names' of different...
Yunfei Dong [Fri, 3 Mar 2023 01:38:40 +0000 (01:38 +0000)]
media: dt-bindings: media: mediatek: vcodec: adapt to the 'clock-names' of different platforms

mt8195 and mt8192 have different clock numbers, separate 'clock-names'
according to compatible name.

Signed-off-by: Yunfei Dong <yunfei.dong@mediatek.com>
Signed-off-by: Allen-KH Cheng <allen-kh.cheng@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Reviewed-by: Chen-Yu Tsai <wenst@chromium.org>
Tested-by: Chen-Yu Tsai <wenst@chromium.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: dt-bindings: media: mediatek: Remove "dma-ranges" property for decoder
Allen-KH Cheng [Fri, 3 Mar 2023 01:38:38 +0000 (01:38 +0000)]
media: dt-bindings: media: mediatek: Remove "dma-ranges" property for decoder

Because the decoder nodes already make use of the iommus property to
configure the IOMMU for address translations, having a dma-ranges
property makes no sense.

In fact, after commit f1ad5338a4d5 ("of: Fix "dma-ranges" handling for
bus controllers"), having a dma-ranges property causes IOMMU faults.

Remove the dma-ranges property and update the example.

Signed-off-by: Allen-KH Cheng <allen-kh.cheng@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Acked-by: Rob Herring <robh@kernel.org>
Reviewed-by: Nícolas F. R. A. Prado <nfraprado@collabora.com>
Reviewed-by: Chen-Yu Tsai <wenst@chromium.org>
Tested-by: Chen-Yu Tsai <wenst@chromium.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: dt-bindings: media: mediatek: Rename child node names for decoder
Allen-KH Cheng [Fri, 3 Mar 2023 01:38:37 +0000 (01:38 +0000)]
media: dt-bindings: media: mediatek: Rename child node names for decoder

In order to make the names of the child nodes more generic, we rename
"vcodec-lat" and "vcodec-core" to "video-codec" for decoder in
patternProperties and example.

Signed-off-by: Allen-KH Cheng <allen-kh.cheng@mediatek.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Chen-Yu Tsai <wenst@chromium.org>
Tested-by: Chen-Yu Tsai <wenst@chromium.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mtk-jpeg: refactor hw dev initializaiton
kyrie wu [Fri, 10 Mar 2023 06:23:55 +0000 (06:23 +0000)]
media: mtk-jpeg: refactor hw dev initializaiton

Add a atomic_t variable for initialization of hw dev.

Signed-off-by: kyrie wu <kyrie.wu@mediatek.com>
Signed-off-by: irui wang <irui.wang@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mtk-jpeg: refactor some variables
kyrie wu [Fri, 10 Mar 2023 06:23:54 +0000 (06:23 +0000)]
media: mtk-jpeg: refactor some variables

Refactor some variables to reduce redundancy
and optimize the initialization process.

Signed-off-by: kyrie wu <kyrie.wu@mediatek.com>
Signed-off-by: irui wang <irui.wang@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mtk-jpeg: Remove some unnecessary variables
kyrie wu [Fri, 10 Mar 2023 06:23:53 +0000 (06:23 +0000)]
media: mtk-jpeg: Remove some unnecessary variables

remove is_jpgenc_multihw, is_jpgdec_multihw,
and unnecessary initialization.

Signed-off-by: kyrie wu <kyrie.wu@mediatek.com>
Signed-off-by: irui wang <irui.wang@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mtk-jpeg: reconstructs the initialization mode of worker
kyrie wu [Fri, 10 Mar 2023 06:23:52 +0000 (06:23 +0000)]
media: mtk-jpeg: reconstructs the initialization mode of worker

exclude is_jpgenc_multihw or is_jpgdec_multihw for judgment,
and re-initialize jpeg worker.

Signed-off-by: kyrie wu <kyrie.wu@mediatek.com>
Signed-off-by: irui wang <irui.wang@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mtk-jpeg: Fixes jpeg enc&dec worker sw flow
kyrie wu [Fri, 10 Mar 2023 06:23:51 +0000 (06:23 +0000)]
media: mtk-jpeg: Fixes jpeg enc&dec worker sw flow

1. Move removing buffer after sw setting and before hw setting
in enc&dec worker to prevents the operation of removing
the buffer twice if the sw setting fails.
2. Remove the redundant operation of queue work in the
jpegenc irq handler because the jpegenc worker has called
v4l2_m2m_job_finish to do it.

Fixes: 5fb1c2361e56 ("mtk-jpegenc: add jpeg encode worker interface")
Fixes: dedc21500334 ("media: mtk-jpegdec: add jpeg decode worker interface")
Signed-off-by: kyrie wu <kyrie.wu@mediatek.com>
Signed-off-by: irui wang <irui.wang@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mtk-jpeg: add jpeg single core initial function
kyrie wu [Fri, 10 Mar 2023 06:23:50 +0000 (06:23 +0000)]
media: mtk-jpeg: add jpeg single core initial function

add single core initial function to init reg_base,
irq and clk.

Signed-off-by: kyrie wu <kyrie.wu@mediatek.com>
Signed-off-by: irui wang <irui.wang@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
18 months agomedia: mtk-jpeg: Fixes jpeghw multi-core judgement
kyrie wu [Fri, 10 Mar 2023 06:23:49 +0000 (06:23 +0000)]
media: mtk-jpeg: Fixes jpeghw multi-core judgement

some chips have multi-hw, but others have only one,
modify the condition of multi-hw judgement

Fixes: 934e8bccac95 ("mtk-jpegenc: support jpegenc multi-hardware")
Signed-off-by: kyrie wu <kyrie.wu@mediatek.com>
Signed-off-by: irui wang <irui.wang@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: ov2685: Select VIDEO_V4L2_SUBDEV_API
Sakari Ailus [Thu, 16 Mar 2023 10:21:49 +0000 (11:21 +0100)]
media: ov2685: Select VIDEO_V4L2_SUBDEV_API

VIDEO_V4L2_SUBDEV_API is required for v4l2_subdev_get_pad_*() functions.
Select it.

Fixes: ("media: i2c: ov2685: Add .get_selection() support")
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: dt-bindings: media: convert meson-ir.txt to dt-schema
Neil Armstrong [Wed, 8 Mar 2023 13:27:31 +0000 (14:27 +0100)]
media: dt-bindings: media: convert meson-ir.txt to dt-schema

Convert the Amlogic Meson IR remote control receiver bindings to
dt-schema.

Take in account the used variant with amlogic,meson-gx-ir.

Reviewed-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: dt-bindings: media: i2c: imx334 add new link_freq
Shravan Chippa [Wed, 1 Mar 2023 07:34:11 +0000 (08:34 +0100)]
media: dt-bindings: media: i2c: imx334 add new link_freq

Add new supported link frequency in dt example.

Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Suggested-by: Sakari Ailus <sakari.ailus@iki.fi>
Signed-off-by: Shravan Chippa <shravan.chippa@microchip.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: i2c: imx334: support lower bandwidth mode
Shravan Chippa [Wed, 1 Mar 2023 07:34:10 +0000 (08:34 +0100)]
media: i2c: imx334: support lower bandwidth mode

Some platforms may not be capable of supporting the bandwidth
required for 12 bit or 3840x2160@60 resolutions.

Add support for dynamically selecting 10 bit and 1920x1080@30
resolutions while leaving the existing configuration as default.

Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
CC: Conor Dooley <conor.dooley@microchip.com>
Signed-off-by: Prakash Battu <Prakash.Battu@microchip.com>
Signed-off-by: Shravan Chippa <shravan.chippa@microchip.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: i2c: imx334: add missing reset values for mode 3840x2160_regs[]
Shravan Chippa [Wed, 1 Mar 2023 07:34:09 +0000 (08:34 +0100)]
media: i2c: imx334: add missing reset values for mode 3840x2160_regs[]

There are some missing reset reg_mode values for the 3840x2160@60
resolution. The camera sensor still works in 3840x2160@60 resolution mode
because of the register reset values. This is an issue when we change the
modes dynamically. As an example, when we change the mode from 1920x1080@30
 resolution to 3840x2160@60 resoultion then the mode values will be written
to the registers from the array mode_3840x2160_regs[] which gives the wrong
output which is incorrect resolution.

So add the missing reset values to the mode_3840x2160_regs[].

Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Shravan Chippa <shravan.chippa@microchip.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: i2c: imx334: replace __v4l2_ctrl_s_ctrl to __v4l2_ctrl_modify_range
Shravan Chippa [Wed, 1 Mar 2023 07:34:08 +0000 (08:34 +0100)]
media: i2c: imx334: replace __v4l2_ctrl_s_ctrl to __v4l2_ctrl_modify_range

For every mode we will get new set of values for hbalnk so use
__v4l2_ctrl_modify_range() to support multi modes for hblank.

The hblank value is readonly in the driver. because of this the function
returns error if we try to change. so added dumy return case in
imx334_set_ctrl function.

Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Suggested-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Shravan Chippa <shravan.chippa@microchip.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: v4l2-ctrls: Fix doc for v4l2_ctrl_request_hdl_find
Miaoqian Lin [Thu, 22 Dec 2022 11:56:37 +0000 (12:56 +0100)]
media: v4l2-ctrls: Fix doc for v4l2_ctrl_request_hdl_find

We should call v4l2_ctrl_request_hdl_put() instead of
v4l2_ctrl_request_put_hdl(). Fix the typo.

Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: dt-bindings: qcom,venus: document firmware-name
Krzysztof Kozlowski [Tue, 27 Dec 2022 14:41:02 +0000 (15:41 +0100)]
media: dt-bindings: qcom,venus: document firmware-name

Venus Linux driver loads firmware based on firmware-name property and
some DTS already have it:

  msm8996-oneplus3.dtb: video-codec@c00000: Unevaluated properties are not allowed ('firmware-name', 'interconnect-names', 'interconnects' were unexpected)

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: dt-bindings: qcom,sm8250-venus: document OPP table
Krzysztof Kozlowski [Tue, 27 Dec 2022 14:41:01 +0000 (15:41 +0100)]
media: dt-bindings: qcom,sm8250-venus: document OPP table

Venus on Qualcomm SM8250 uses Operating Performance Points (both in DTS
and driver):

  sm8250-hdk.dtb: video-codec@aa00000: 'operating-points-v2', 'opp-table' do not match any of the regexes: 'pinctrl-[0-9]+'

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: dt-bindings: qcom,sdm845-venus-v2: document OPP table
Krzysztof Kozlowski [Tue, 27 Dec 2022 14:41:00 +0000 (15:41 +0100)]
media: dt-bindings: qcom,sdm845-venus-v2: document OPP table

Venus on Qualcomm SDM845 uses Operating Performance Points (in Linux driver).

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: dt-bindings: qcom,sc7280-venus: document OPP table
Krzysztof Kozlowski [Tue, 27 Dec 2022 14:40:59 +0000 (15:40 +0100)]
media: dt-bindings: qcom,sc7280-venus: document OPP table

Venus on Qualcomm SC7280 uses Operating Performance Points (both in DTS
and driver):

  sc7280-crd-r3.dtb: video-codec@aa00000: Unevaluated properties are not allowed ('operating-points-v2', 'opp-table' were unexpected)

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: dt-bindings: qcom,sc7180-venus: document OPP table
Krzysztof Kozlowski [Tue, 27 Dec 2022 14:40:58 +0000 (15:40 +0100)]
media: dt-bindings: qcom,sc7180-venus: document OPP table

Venus on Qualcomm SC7180 uses Operating Performance Points (both in DTS
and driver):

  sc7180-idp.dtb: video-codec@aa00000: Unevaluated properties are not allowed ('operating-points-v2', 'opp-table' were unexpected)

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: dt-bindings: qcom,msm8996-venus: document interconnects
Krzysztof Kozlowski [Tue, 27 Dec 2022 14:40:57 +0000 (15:40 +0100)]
media: dt-bindings: qcom,msm8996-venus: document interconnects

Venus on Qualcomm MSM8996 defines interconnects:

  apq8096-db820c.dtb: video-codec@c00000: Unevaluated properties are not allowed ('interconnect-names', 'interconnects' were unexpected)

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: dt-bindings: qcom,venus: split common properties
Krzysztof Kozlowski [Tue, 27 Dec 2022 14:40:56 +0000 (15:40 +0100)]
media: dt-bindings: qcom,venus: split common properties

All Qualcomm SoC Venus bindings share a lot of properties, so split
common part to re-usable schema to reduce the duplication and promote
unified style.

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: dt-bindings: qcom,venus: cleanup
Krzysztof Kozlowski [Tue, 27 Dec 2022 14:40:55 +0000 (15:40 +0100)]
media: dt-bindings: qcom,venus: cleanup

Cleanup the Qualcomm SoC Venus bindings:
 - Drop unneeded blank lines and quotes,
 - Fix indentation in example to 4-space (to match DT schema bindings
   style),
 - Add SoC name in each title.

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: max9286: Free control handler
Laurent Pinchart [Sat, 14 Jan 2023 21:46:50 +0000 (22:46 +0100)]
media: max9286: Free control handler

The control handler is leaked in some probe-time error paths, as well as
in the remove path. Fix it.

Fixes: 66d8c9d2422d ("media: i2c: Add MAX9286 driver")
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: i2c: imx258: Parse and register properties
Robert Mader [Wed, 4 Jan 2023 12:23:37 +0000 (13:23 +0100)]
media: i2c: imx258: Parse and register properties

Analogous to e.g. the imx219. This enables propagating
V4L2_CID_CAMERA_ORIENTATION and V4L2_CID_CAMERA_SENSOR_ROTATION values.
The motivation is to allow libcamera detect these values from the device
tree and propagate them further to e.g. Pipewire.

While at it, reserve space for 3 additional controls even if
v4l2_ctrl_new_fwnode_properties() can only register 2 of them, to fix the
existing implementation which reserve space for 8 controls but actually
registers 9.

[Sakari Ailus: Rewrapped the commit message and removed changelog]

Signed-off-by: Robert Mader <robert.mader@collabora.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: ov13b10: remove streaming mode set from reg_list
Bingbu Cao [Thu, 5 Jan 2023 11:02:49 +0000 (12:02 +0100)]
media: ov13b10: remove streaming mode set from reg_list

It is not expected that trying to set the sensor to
streaming mode when trying to set initial sensor configuration.

Signed-off-by: Bingbu Cao <bingbu.cao@intel.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: ov13b10: Support device probe in non-zero ACPI D state
Arec Kao [Tue, 14 Mar 2023 12:14:36 +0000 (13:14 +0100)]
media: ov13b10: Support device probe in non-zero ACPI D state

Tell ACPI device PM code that the driver supports the device being in
non-zero ACPI D state when the driver's probe function is entered.

Also do identification on the first access of the device, whether in
probe or when starting streaming.

Signed-off-by: Arec Kao <arec.kao@intel.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: i2c: ov2685: Add .get_selection() support
Luca Weiss [Wed, 1 Feb 2023 17:07:01 +0000 (18:07 +0100)]
media: i2c: ov2685: Add .get_selection() support

Add support for the .get_selection() pad operation to the ov2685 sensor
driver.

Report the native sensor size (pixel array), the crop bounds (readable
pixel array area) and the current and default analog crop rectangles.

Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Luca Weiss <luca@z3ntu.xyz>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
19 months agomedia: i2c: ov2685: Add controls from fwnode
Luca Weiss [Wed, 1 Feb 2023 17:07:00 +0000 (18:07 +0100)]
media: i2c: ov2685: Add controls from fwnode

Add V4L2_CID_CAMERA_ORIENTATION and V4L2_CID_CAMERA_SENSOR_ROTATION
controls to the ov2685 driver by attempting to parse them from firmware.

Signed-off-by: Luca Weiss <luca@z3ntu.xyz>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>