[libcamera-devel] [RFC PATCH v2] libcamera: pipeline: ipu3: Apply the requested test pattern mode
Jean-Michel Hautbois
jeanmichel.hautbois at ideasonboard.com
Mon Oct 25 08:29:58 CEST 2021
Hi Hiro,
Thanks for the patch !
On 05/10/2021 07:44, Hirokazu Honda wrote:
> This enables ipu3 pipeline handler to apply the test pattern mode
> per frame.
>
> Signed-off-by: Hirokazu Honda <hiroh at chromium.org>
> ---
> src/libcamera/pipeline/ipu3/ipu3.cpp | 38 ++++++++++++++++++++++++++--
> 1 file changed, 36 insertions(+), 2 deletions(-)
>
> diff --git a/src/libcamera/pipeline/ipu3/ipu3.cpp b/src/libcamera/pipeline/ipu3/ipu3.cpp
> index 92e86925..07fda65f 100644
> --- a/src/libcamera/pipeline/ipu3/ipu3.cpp
> +++ b/src/libcamera/pipeline/ipu3/ipu3.cpp
> @@ -69,6 +69,7 @@ public:
> void statBufferReady(FrameBuffer *buffer);
> void queuePendingRequests();
> void cancelPendingRequests();
> + void frameStart(uint32_t sequence);
>
> CIO2Device cio2_;
> ImgUDevice *imgu_;
> @@ -88,6 +89,7 @@ public:
> std::unique_ptr<ipa::ipu3::IPAProxyIPU3> ipa_;
>
> std::queue<Request *> pendingRequests_;
> + std::queue<Request *> processingRequests_;
>
> ControlInfoMap ipaControls_;
>
> @@ -568,6 +570,8 @@ int PipelineHandlerIPU3::configure(Camera *camera, CameraConfiguration *c)
> cio2->sensor()->sensorInfo(&sensorInfo);
> data->cropRegion_ = sensorInfo.analogCrop;
>
> + cio2->sensor()->setTestPatternMode(controls::draft::TestPatternModeOff);
There is a dependency on "[PATCH] libcamera: camera_sensor: Enable to
set a test pattern mode" (20211005043723.568685-1-hiroh at chromium.org),
right ?
Could you make it a series please :-) ?
> +
> /*
> * Configure the H/V flip controls based on the combination of
> * the sensor and user transform.
> @@ -801,6 +805,9 @@ void PipelineHandlerIPU3::stop(Camera *camera)
>
> ret |= data->imgu_->stop();
> ret |= data->cio2_.stop();
> +
> + data->processingRequests_ = {};
Shouldn't we do that in cancelPendingRequests() as a mirror of the push
which is done in queuePendingRequests() ?
I am a bit worry that it is done after cio2_->stop()...
> +
> if (ret)
> LOG(IPU3, Warning) << "Failed to stop camera " << camera->id();
>
> @@ -851,6 +858,8 @@ void IPU3CameraData::queuePendingRequests()
>
> info->rawBuffer = rawBuffer;
>
> + processingRequests_.push(request);
> +
> ipa::ipu3::IPU3Event ev;
> ev.op = ipa::ipu3::EventProcessControls;
> ev.frame = info->id;
> @@ -1107,8 +1116,8 @@ int PipelineHandlerIPU3::registerCameras()
> data->delayedCtrls_ =
> std::make_unique<DelayedControls>(cio2->sensor()->device(),
> params);
> - data->cio2_.frameStart().connect(data->delayedCtrls_.get(),
> - &DelayedControls::applyControls);
> + data->cio2_.frameStart().connect(data.get(),
> + &IPU3CameraData::frameStart);
The IPA will receive the ipa::ipu3::EventProcessControls immediately, so
it could be notified to stop trying to apply the algorithm for instance
(to be discussed). I think this is good :-).
>
> /* Convert the sensor rotation to a transformation */
> int32_t rotation = 0;
> @@ -1399,6 +1408,31 @@ void IPU3CameraData::statBufferReady(FrameBuffer *buffer)
> ipa_->processEvent(ev);
> }
>
> +void IPU3CameraData::frameStart(uint32_t sequence)
> +{
> + if (!processingRequests_.empty()) {
Add a LOG(IPU3, Debug) here ?
> + /* Assumes applying the test pattern mode affects immediately. */
> + Request *request = processingRequests_.front();
> + processingRequests_.pop();
> +
> + if (request->controls().contains(controls::draft::TestPatternMode)) {
Same here, to see which test pattern is applied ?
> + const int32_t testPatternMode = request->controls().get(
> + controls::draft::TestPatternMode);
> +
> + int ret = cio2_.sensor()->setTestPatternMode(testPatternMode);
> + if (ret) {
> + LOG(IPU3, Error) << "Failed to set test pattern mode: "
> + << ret;
> + } else {
> + request->metadata().set(controls::draft::TestPatternMode,
> + testPatternMode);
> + }
> + }
> + }
> +
And a small comment for this line ?
> + delayedCtrls_->applyControls(sequence);
> +}
> +
> REGISTER_PIPELINE_HANDLER(PipelineHandlerIPU3)
>
> } /* namespace libcamera */
>
Now, this patch makes the per-frame control on the IPA point of view
more important... How will we handle those ? This is still an open
question...
Reviewed-by: Jean-Michel Hautbois <jeanmichel.hautbois at ideasonboard.com>
More information about the libcamera-devel
mailing list