[libcamera-devel] [PATCH v3 01/11] libcamera: Always check return value of MediaDevice::acquire()

[Laurent Pinchart]

Hi Niklas,

Thank you for the patch.

On Sat, May 11, 2019 at 11:18:57AM +0200, Niklas Söderlund wrote:
> In preparation for adding more responsibility to MediaDevice::acquire()
> remove unneeded calls to acquire() and release(), and make sure all
> needed calls to acquire() are checked and acted on.
> 
> Signed-off-by: Niklas Söderlund <niklas.soderlund at ragnatech.se>
> ---
>  src/libcamera/pipeline/ipu3/ipu3.cpp         | 10 ++++------
>  src/libcamera/pipeline/uvcvideo.cpp          |  3 ++-
>  src/libcamera/pipeline/vimc.cpp              |  3 ++-
>  test/media_device/media_device_link_test.cpp |  6 +++++-
>  test/v4l2_device/v4l2_device_test.cpp        |  4 ----
>  test/v4l2_subdevice/v4l2_subdevice_test.cpp  |  7 -------
>  6 files changed, 13 insertions(+), 20 deletions(-)
> 
> diff --git a/src/libcamera/pipeline/ipu3/ipu3.cpp b/src/libcamera/pipeline/ipu3/ipu3.cpp
> index 0041662514e1d7ca..75e878afad4e67a9 100644
> --- a/src/libcamera/pipeline/ipu3/ipu3.cpp
> +++ b/src/libcamera/pipeline/ipu3/ipu3.cpp
> @@ -614,21 +614,19 @@ bool PipelineHandlerIPU3::match(DeviceEnumerator *enumerator)
>  	imgu_dm.add("ipu3-imgu 1 viewfinder");
>  	imgu_dm.add("ipu3-imgu 1 3a stat");
>  
> -	/*
> -	 * It is safe to acquire both media devices at this point as
> -	 * DeviceEnumerator::search() skips the busy ones for us.
> -	 */
>  	cio2MediaDev_ = enumerator->search(cio2_dm);
>  	if (!cio2MediaDev_)
>  		return false;
>  
> -	cio2MediaDev_->acquire();
> +	if (!cio2MediaDev_->acquire())
> +		return false;
>  
>  	imguMediaDev_ = enumerator->search(imgu_dm);
>  	if (!imguMediaDev_)
>  		return false;
>  
> -	imguMediaDev_->acquire();
> +	if (!imguMediaDev_->acquire())
> +		return false;
>  
>  	/*
>  	 * Disable all links that are enabled by default on CIO2, as camera
> diff --git a/src/libcamera/pipeline/uvcvideo.cpp b/src/libcamera/pipeline/uvcvideo.cpp
> index 5d2f1c98fa36380c..e40b052f4d877d5d 100644
> --- a/src/libcamera/pipeline/uvcvideo.cpp
> +++ b/src/libcamera/pipeline/uvcvideo.cpp
> @@ -183,7 +183,8 @@ bool PipelineHandlerUVC::match(DeviceEnumerator *enumerator)
>  	if (!media_)
>  		return false;
>  
> -	media_->acquire();
> +	if (!media_->acquire())
> +		return false;
>  
>  	std::unique_ptr<UVCCameraData> data = utils::make_unique<UVCCameraData>(this);
>  
> diff --git a/src/libcamera/pipeline/vimc.cpp b/src/libcamera/pipeline/vimc.cpp
> index cdf43770a9bfc40f..7b6ebd4cc0a27e25 100644
> --- a/src/libcamera/pipeline/vimc.cpp
> +++ b/src/libcamera/pipeline/vimc.cpp
> @@ -193,7 +193,8 @@ bool PipelineHandlerVimc::match(DeviceEnumerator *enumerator)
>  	if (!media_)
>  		return false;
>  
> -	media_->acquire();
> +	if (!media_->acquire())
> +		return false;
>  
>  	std::unique_ptr<VimcCameraData> data = utils::make_unique<VimcCameraData>(this);
>  
> diff --git a/test/media_device/media_device_link_test.cpp b/test/media_device/media_device_link_test.cpp
> index 58a55cdfee634c0b..484d3691310f78a0 100644
> --- a/test/media_device/media_device_link_test.cpp
> +++ b/test/media_device/media_device_link_test.cpp
> @@ -51,7 +51,11 @@ class MediaDeviceLinkTest : public Test
>  			return TestSkip;
>  		}
>  
> -		dev_->acquire();
> +		if (!dev_->acquire()) {
> +			cerr << "Unable to acquire media device "
> +			     << dev_->deviceNode() << endl;
> +			return TestSkip;
> +		}
>  
>  		if (dev_->open()) {
>  			cerr << "Failed to open media device at "
> diff --git a/test/v4l2_device/v4l2_device_test.cpp b/test/v4l2_device/v4l2_device_test.cpp
> index 4225291bbb6e23f0..833038d56ea4631d 100644
> --- a/test/v4l2_device/v4l2_device_test.cpp
> +++ b/test/v4l2_device/v4l2_device_test.cpp
> @@ -46,8 +46,6 @@ int V4L2DeviceTest::init()
>  	if (!media_)
>  		return TestSkip;
>  
> -	media_->acquire();
> -

Why is this call and the ones below unneeded ? Isn't it a good practice
to always acquire the device before operating on it ? I suppose it's not
strictly mandatory here as we only inspect the media device to find
viden and subdev nodes, but is there any drawback in keeping these calls
?

>  	MediaEntity *entity = media_->getEntityByName("vivid-000-vid-cap");
>  	if (!entity)
>  		return TestSkip;
> @@ -61,8 +59,6 @@ int V4L2DeviceTest::init()
>  
>  void V4L2DeviceTest::cleanup()
>  {
> -	media_->release();
> -
>  	capture_->streamOff();
>  	capture_->releaseBuffers();
>  	capture_->close();
> diff --git a/test/v4l2_subdevice/v4l2_subdevice_test.cpp b/test/v4l2_subdevice/v4l2_subdevice_test.cpp
> index 72d5f72543d29378..5810ef3c962fab8e 100644
> --- a/test/v4l2_subdevice/v4l2_subdevice_test.cpp
> +++ b/test/v4l2_subdevice/v4l2_subdevice_test.cpp
> @@ -45,20 +45,16 @@ int V4L2SubdeviceTest::init()
>  		return TestSkip;
>  	}
>  
> -	media_->acquire();
> -
>  	int ret = media_->open();
>  	if (ret) {
>  		cerr << "Unable to open media device: " << media_->deviceNode()
>  		     << ": " << strerror(ret) << endl;
> -		media_->release();
>  		return TestSkip;
>  	}
>  
>  	MediaEntity *videoEntity = media_->getEntityByName("Scaler");
>  	if (!videoEntity) {
>  		cerr << "Unable to find media entity 'Scaler'" << endl;
> -		media_->release();
>  		return TestFail;
>  	}
>  
> @@ -67,7 +63,6 @@ int V4L2SubdeviceTest::init()
>  	if (ret) {
>  		cerr << "Unable to open video subdevice "
>  		     << scaler_->entity()->deviceNode() << endl;
> -		media_->release();
>  		return TestSkip;
>  	}
>  
> @@ -76,7 +71,5 @@ int V4L2SubdeviceTest::init()
>  
>  void V4L2SubdeviceTest::cleanup()
>  {
> -	media_->release();
> -
>  	delete scaler_;
>  }

-- 
Regards,

Laurent Pinchart