[libcamera-devel] [PATCH 2/2] libcamera: pipeline-handler: Consider max in-flight requests constraint
Jacopo Mondi
jacopo at jmondi.org
Thu Dec 15 09:55:40 CET 2022
Hi Paul
On Thu, Dec 15, 2022 at 02:18:22PM +0900, Paul Elder wrote:
> Hi Jacopo,
>
> On Wed, Dec 14, 2022 at 02:41:30PM +0100, Jacopo Mondi wrote:
> > Hi Paul
> >
> > On Wed, Dec 14, 2022 at 09:13:06PM +0900, Paul Elder wrote:
> > > Hi,
> > >
> > > On Wed, Dec 14, 2022 at 10:19:54AM +0100, Jacopo Mondi via libcamera-devel wrote:
> > > > Hi Umang
> > > >
> > > > On Wed, Dec 14, 2022 at 02:11:12PM +0530, Umang Jain wrote:
> > > > > Hi Jacopo
> > > > >
> > > > > On 12/14/22 1:29 PM, Jacopo Mondi wrote:
> > > > > > Hi Umang
> > > > > >
> > > > > > On Wed, Dec 14, 2022 at 11:07:25AM +0530, Umang Jain wrote:
> > > > > > > Hi Jacopo,
> > > > > > >
> > > > > > > On 12/13/22 9:25 PM, Jacopo Mondi wrote:
> > > > > > > > Hi Umang,
> > > > > > > >
> > > > > > > > On Mon, Dec 12, 2022 at 08:32:56PM +0530, Umang Jain via libcamera-devel wrote:
> > > > > > > > > This patch introduces a new constraint in the pipeline handler base
> > > > > > > > > class which allows various derived pipeline handlers to set and
> > > > > > > > > restrict the number of maximum in-flight requests that can be queued to
> > > > > > > > > the underlying components (for e.g. the IPA).
> > > > > > > > It's my understanding that the max number of in-flight requests is
> > > > > > > > what the properties::PipelineDepth property expresses.
> > > > > > > Indeed, that's why understanding as well. The properties::PipelineDepth
> > > > > > > should be set by the pipeline-handlers.
> > >
> > > (Besides the fact that I can't find properties::PipelineDepth)
> > > controls::PipelineDepth says that it "Specifies the number of pipeline
> > > stages the frame went through from when it was exposed to when the final
> > > completed result was available to the framework.", which to me sounds
> > > like the *minimum* amount of in-flight requests.
> >
> > It's a draft control copied in to please Android.
> > My understanding was that it had to be made a non-mutable property
>
> Ah, indeed it says that :)
>
> >
> > >
> > > I thought that here we're defining the *maximum* allowed in-flight
> > > requests. Not that all of the requests have to actually be doing
> > > something in hardware, but that it's the amount of requests the the
> > > pipeline handler is capable of handling internally without overflowing.
> > >
> > > > > > >
> > > > > > > We need to identify and atleast agree upon the correct PipelineDepth for the
> > > > > > > two platforms that uses the FCQueue (IPU3 and RkISP1) as well. Some grey
> > > > > > > areas there as well
> > > > > > I think 16 is fine for now. At least doesn't change what we already
> > > > > > have today
> > > > > >
> > > > > > > > Can't we access that property at queue request time, or maybe better we
> > > > > > > > could cache it at camera registration time, since it's constant ?
> > > > > > > If we intent to use the properties::PipelineDepth, as the size of the
> > > > > > > FCQueue - it needs to exist / set before IPA is constructed.
> > > > > > >
> > > > > > > (See IPAIPU3::IPAIPU3() for context)
> > > > > > >
> > > > > > The property can still be defined in the IPA header as IPA can access
> > > > > > it safely.
> > > > >
> > > > > You mean via a global const variable like it's done in Patch 1/2 ?
> > > >
> > > > Yes, like in patch 1/2
> > > >
> > > >
> > > > > >
> > > > > > Pipeline Handlers will have to register the property -before- the
> > > > > > camera is registered, but looking at IPU3 that's already the case now
> > > > > > (data->properties_ gets populated before calling registerCamera())
> > > > > >
> > > > > >
> > > > > > > So I think queue request time is quite late there. Camera registeration time
> > > > > > > might be possible with some changes (looking at
> > > > > > > PipelineHandlerIPU3::registerCameras())
> > > > > > What I meant is either
> > > > > >
> > > > > > 1) At PipelineHandler::queueRequest() time, the PipelineHandler base
> > > > > > class accesses Camera::properties() and fetches
> > > > > > properties::PipelineDepth
> > > > >
> > > > > I think it's a over-kill here. queueRequest() is a recurring call and I
> > > > > don't want fetching of PipelineDepth here.
> > >
> > > Oh one hand, yeah imo getting it from Camera::properties() is more
> > > correct, but on the other hand it does indeed seem overkill.
> > >
> > > >
> > > > While I consider a lookup on an unordered_map<> of a few items not
> > > > that expensive, I agree it's in an hot path, and it would be better to
> > > > cache the value somewhere. Caching it inside Camera or Camera::Private
> > > > is a duplication, as it's already part of
> > > > Camera::Private::properties_, but I would be fine with the
> > > > PipelineHandler base class doing something like this (as proposed in
> > > > my previous 2) point below)
> > > >
> > > > void PipelineHandler::registerCamera(std::shared_ptr<Camera> camera)
> > > > {
> > > >
> > > > /* Cache the pipeline depth in the Camera private data. */
> > > > const auto &pipelineDepth = camera->properties().get(properties::PipelineDepth);
> > > >
> > > > Camera::Private *data = camera->_d();
> > > > data->pipelineDepth = pipelineDepth ? 0 : *pipelineDepth;
> > > >
> > > > ...
> > > > }
> > > >
> > > > Otherwise the PipelineHandler base class can be added with a class
> > > > member, and each registred camera will overwrite that value, it's less
> > > > nice, but all cameras from the same PH instance will have the same
> > > > pipeline depth, so...
> > >
> > > Perhaps caching would be a good middle ground.
> > >
> > > >
> > > >
> > > > > > 2) At PipelineHandler::registerCamera() time the property value is
> > > > > > cached, but this would require to cache per-camera data in
> > > > > > PipelineHandler, something we don't have, or to populate a class
> > > > > > member of Camera::Private for that purpose..
> > > > >
> > > > > That's interesting because I initially didn't think of multiple cameras
> > > > > registered with one instance of pipeline-handler.
> > > > > It also makes me wonder -
> > > > >
> > > > > - Is properties::PipelineDepth Camera specific or platform-specific ? Seems
> > > > > now it's Camera specific
> > > > > - Then defining it IPA headers won't be a good option, in cases where
> > > > > PipelineDepth for different cameras differ.
> > > >
> > > > Why camera specific ? Each camera will indeed register it as a
> > > > property for applications to retrieve it, but the
> > > > value depends on the platform afaict and it can be a compile-time constant
> > > > defined by the IPA headers. Even better, the same compile-time
> > > > constant will be used to size the FCQ.
> > >
> > > If it's platform-specific then wouldn't the pipeline handler have to
> > > report it based on what platform it's running on? So we can't really
> > >hardcode a constant. Maybe putting it in properties is indeed better.
> >
> > As far as I get it, this property applies to pipelines with an ISP (no
> > uvc, no simple) and those pipelines by definition runs on a single
> > platform, don't they ?
>
> Well, the rkisp1 pipeline runs on imx8mp too... (with a few on-the-list
> but out-of-tree kernel patches)
>
Correct, but it does so as it uses the same ISP, same IPA
implementation, same FCQ depth ?
> >
> > >
> > > >
> > > > >
> > > > > Need some thinking ...
> > > > >
> > > > > But none the less, this is a something that can be done on top with wider
> > > > > discussion / context on PipelineDepth. It shouldn't necessarily block the
> > > > > series.
> > > >
> > > > My whole point is that I think we should find a better mechanism for
> > > > registering pipeline depth compared to having pipeline handlers call
> > > > setMaxQueueRequests() as it requires an explicit action that can
> > > > easily be overlooked, while we want the queue overflow protection to
> > >
> > > I mean, we could initialize kMaxFrameContexts to an invalid value (-1?)
> > > and then force *all* pipeline handlers to explictly set it, even if to
> > > zero. Then it becomes really easy to catch.
> >
> > The property doesn't make much sense for UVC, right ?
> >
> > Not all pipelines will use libipa and have an FCQ (RPi in example at
> > the moment). I don't think forcing all of them is actually necessary.
>
> Hm, that's true. No initial invalid value then.
>
> >
> > But the pipelines that do need to throttle the number of requests in
> > flight with the help of the PipelineHandler base class should ideally
> > get it for free by simply registering the property ?
>
> Yeah that does sound more nice.
>
> >
> > >
> > > Or note it down as a compliancy thing and have lc-compliance test for
> > > it. I presume this will become a required property to be exposed anyway.
> > >
> > > In any case, the goal is to report to the user the "maximum number of
> > > requests that are allowed to be queued in the camera", right? So it'll
> > > have to go through properties anyway. Perhaps then it's best to expose
> > > it via properties (as it's already the interface we have for exposing
> > > properties to the user), and then the base PipelineHandler can cache it
> > > for the queuing guard. That would get rid of the special explicit call,
> > > and would still achieve the goal.
> >
> > That would be my preference. There might be a discussions if
> > properties::PipelineDepth is the right property to report it or it
> > will conflict with Android's pipeline depth, which is different.
>
> +1
>
> Although to me this sounds separate from PipelineDepth, because my
> understanding is that that is the maximum amount of Requests that can be
> in-flight, while what we want here is that plus some extra that the
> application is allowed to "overqueue" that the pipeline handler will
> automagically handle internally.
I think you're right here.
Actually, with this request blocking protection in place it is not
just "some extra" but application can theoretically queue an infinite
number of requests, but only up to X of them will be processed
concurrently. Which makes me wonder what is the use for applications
of knowing that, if the requests queue length from their point of view is
infinite (aka they can queue as much requests as they like).
It is instead more relevant for applications to be able to identify the
number of processing steps a request has to go through before being
completed, as it represents the latency (in frames I presume) before
an image gets exposed to the time it is delivered to userspace.
The Android documentation for 'pipelineMaxDepth' reports:
-------------------------------------------------------------------------------
A typical minimum value for this is 2 (one stage to expose,one stage
to readout) from the sensor. The ISP then usually adds its own stages
to do custom HW processing. Further stages may be added by SW
processing.
Depending on what settings are used (e.g. YUV, JPEG) and what
processing is enabled (e.g. face detection), the actual pipeline depth
(specified by android.request.pipelineDepth) may be less than the max
pipeline depth.
A pipeline depth of X stages is equivalent to a pipeline latency of X
frame intervals.
This value will normally be 8 or less, however, for high speed capture
session,the max pipeline depth will be up to 8 x size of high speed
capture request list.
-------------------------------------------------------------------------------
and we probably want something very similar.
Ok, scratch my properties::PipelineDepth suggestion, but I still don't
like very much the explicit function call. Would it make sense to have
a protected PipelineHandler class member for pipelines to populate ?
>
> Am I misunderstanding something?
not at all, I probably was :)
>
>
> Paul
>
> >
> > >
> > > I suppose then we'd actually need a compliance test for the property,
> > > but I guess we have a few more of those anyway.
> > >
> > > My two cents.
> >
> > Thanks ;)
> > j
> >
> > >
> > >
> > > Paul
> > >
> > > > happen transparently if a pipeline handler register
> > > > properties::PipelineDepth ? Am I overthinking this ?
> > > >
> > > >
> > > > > >
> > > > > > Sorry for not having been clear
> > > > > >
> > > > > >
> > > > > > >
> > > > > > > > Also, can this feature be tested in test/ ? Probably not, as it
> > > > > > > > requires a compliant pipeline handler ? Maybe vimc can be instrumented
> > > > > > > I have been thinking about that as well. It's shouldn't be too difficult.
> > > > > > > > ?
> > > > > > > >
> > > > > > > > > The pipeline handler is now equipped with the responsibility of not to
> > > > > > > > > over queue the requests to the underlying layers. The derived
> > > > > > > > > pipeline handler (or even IPA) can also have various kind of requests
> > > > > > > > > queue(arrays, queues or ring-buffer) hence, it might be an issue where
> > > > > > > > > the application queues the requests at a rate where these kind of
> > > > > > > > > queues can over-flow. The patch ensures that the base PipelineHandler
> > > > > > > > > will never let the requests overflow to underlying layers, once the
> > > > > > > > > derived pipeline handler sets its maximum capacity of handling
> > > > > > > > > in-flight requests using PipelineHandler::setMaxQueueRequests().
> > > > > > > > >
> > > > > > > > > The queue request management introduced in the pipeline handler base
> > > > > > > > > class is now used by the IPU3 and RkISP1 pipeline handlers. This will
> > > > > > > > > prevent over-writing of frame contexts (i.e. FCQueue) in these two
> > > > > > > > > pipeline handlers.
> > > > > > > > >
> > > > > > > > > Signed-off-by: Umang Jain <umang.jain at ideasonboard.com>
> > > > > > > > > ---
> > > > > > > > > include/libcamera/internal/pipeline_handler.h | 4 ++
> > > > > > > > > src/libcamera/pipeline/ipu3/ipu3.cpp | 1 +
> > > > > > > > > src/libcamera/pipeline/rkisp1/rkisp1.cpp | 1 +
> > > > > > > > > src/libcamera/pipeline_handler.cpp | 51 ++++++++++++++++++-
> > > > > > > > > 4 files changed, 56 insertions(+), 1 deletion(-)
> > > > > > > > >
> > > > > > > > > diff --git a/include/libcamera/internal/pipeline_handler.h b/include/libcamera/internal/pipeline_handler.h
> > > > > > > > > index ec4f662d..83f8bd9f 100644
> > > > > > > > > --- a/include/libcamera/internal/pipeline_handler.h
> > > > > > > > > +++ b/include/libcamera/internal/pipeline_handler.h
> > > > > > > > > @@ -70,6 +70,7 @@ public:
> > > > > > > > > protected:
> > > > > > > > > void registerCamera(std::shared_ptr<Camera> camera);
> > > > > > > > > void hotplugMediaDevice(MediaDevice *media);
> > > > > > > > > + void setMaxQueueRequests(uint32_t maxRequests);
> > > > > > > > >
> > > > > > > > > virtual int queueRequestDevice(Camera *camera, Request *request) = 0;
> > > > > > > > > virtual void stopDevice(Camera *camera) = 0;
> > > > > > > > > @@ -97,6 +98,9 @@ private:
> > > > > > > > > Mutex lock_;
> > > > > > > > > unsigned int useCount_ LIBCAMERA_TSA_GUARDED_BY(lock_);
> > > > > > > > >
> > > > > > > > > + uint32_t maxQueueRequests_;
> > > > > > > > > + uint32_t requestsQueueCounter_;
> > > > > > > > > +
> > > > > > > > > friend class PipelineHandlerFactoryBase;
> > > > > > > > > };
> > > > > > > > >
> > > > > > > > > diff --git a/src/libcamera/pipeline/ipu3/ipu3.cpp b/src/libcamera/pipeline/ipu3/ipu3.cpp
> > > > > > > > > index e4d79ea4..d1d42f78 100644
> > > > > > > > > --- a/src/libcamera/pipeline/ipu3/ipu3.cpp
> > > > > > > > > +++ b/src/libcamera/pipeline/ipu3/ipu3.cpp
> > > > > > > > > @@ -422,6 +422,7 @@ CameraConfiguration::Status IPU3CameraConfiguration::validate()
> > > > > > > > > PipelineHandlerIPU3::PipelineHandlerIPU3(CameraManager *manager)
> > > > > > > > > : PipelineHandler(manager), cio2MediaDev_(nullptr), imguMediaDev_(nullptr)
> > > > > > > > > {
> > > > > > > > > + setMaxQueueRequests(ipa::ipu3::kMaxFrameContexts);
> > > > > > > > > }
> > > > > > > > >
> > > > > > > > > std::unique_ptr<CameraConfiguration>
> > > > > > > > > diff --git a/src/libcamera/pipeline/rkisp1/rkisp1.cpp b/src/libcamera/pipeline/rkisp1/rkisp1.cpp
> > > > > > > > > index eb9ad65c..a48adba9 100644
> > > > > > > > > --- a/src/libcamera/pipeline/rkisp1/rkisp1.cpp
> > > > > > > > > +++ b/src/libcamera/pipeline/rkisp1/rkisp1.cpp
> > > > > > > > > @@ -599,6 +599,7 @@ CameraConfiguration::Status RkISP1CameraConfiguration::validate()
> > > > > > > > > PipelineHandlerRkISP1::PipelineHandlerRkISP1(CameraManager *manager)
> > > > > > > > > : PipelineHandler(manager), hasSelfPath_(true)
> > > > > > > > > {
> > > > > > > > > + setMaxQueueRequests(ipa::rkisp1::kMaxFrameContexts);
> > > > > > > > > }
> > > > > > > > >
> > > > > > > > > std::unique_ptr<CameraConfiguration>
> > > > > > > > > diff --git a/src/libcamera/pipeline_handler.cpp b/src/libcamera/pipeline_handler.cpp
> > > > > > > > > index cfade490..103f9db0 100644
> > > > > > > > > --- a/src/libcamera/pipeline_handler.cpp
> > > > > > > > > +++ b/src/libcamera/pipeline_handler.cpp
> > > > > > > > > @@ -67,7 +67,8 @@ LOG_DEFINE_CATEGORY(Pipeline)
> > > > > > > > > * through the PipelineHandlerFactoryBase::create() function.
> > > > > > > > > */
> > > > > > > > > PipelineHandler::PipelineHandler(CameraManager *manager)
> > > > > > > > > - : manager_(manager), useCount_(0)
> > > > > > > > > + : manager_(manager), useCount_(0), maxQueueRequests_(0),
> > > > > > > > > + requestsQueueCounter_(0)
> > > > > > > > > {
> > > > > > > > > }
> > > > > > > > >
> > > > > > > > > @@ -428,6 +429,9 @@ void PipelineHandler::doQueueRequest(Request *request)
> > > > > > > > > Camera::Private *data = camera->_d();
> > > > > > > > > data->queuedRequests_.push_back(request);
> > > > > > > > >
> > > > > > > > > + if (maxQueueRequests_)
> > > > > > > > > + requestsQueueCounter_++;
> > > > > > > > > +
> > > > > > > > > request->_d()->sequence_ = data->requestSequence_++;
> > > > > > > > >
> > > > > > > > > if (request->_d()->cancelled_) {
> > > > > > > > > @@ -455,6 +459,10 @@ void PipelineHandler::doQueueRequests()
> > > > > > > > > if (!request->_d()->prepared_)
> > > > > > > > > break;
> > > > > > > > >
> > > > > > > > > + if (maxQueueRequests_ &&
> > > > > > > > > + requestsQueueCounter_ >= maxQueueRequests_)
> > > > > > > > > + break;
> > > > > > > > > +
> > > > > > > > > doQueueRequest(request);
> > > > > > > > > waitingRequests_.pop();
> > > > > > > > > }
> > > > > > > > > @@ -531,6 +539,9 @@ void PipelineHandler::completeRequest(Request *request)
> > > > > > > > > ASSERT(!req->hasPendingBuffers());
> > > > > > > > > data->queuedRequests_.pop_front();
> > > > > > > > > camera->requestComplete(req);
> > > > > > > > > +
> > > > > > > > > + if (maxQueueRequests_)
> > > > > > > > > + requestsQueueCounter_--;
> > > > > > > > > }
> > > > > > > > > }
> > > > > > > > >
> > > > > > > > > @@ -647,6 +658,44 @@ void PipelineHandler::disconnect()
> > > > > > > > > * constant for the whole lifetime of the pipeline handler.
> > > > > > > > > */
> > > > > > > > >
> > > > > > > > > +/**
> > > > > > > > > + * \var PipelineHandler::maxQueueRequests_
> > > > > > > > > + * \brief Maximum number of in-flight requests that can be queued
> > > > > > > > > + *
> > > > > > > > > + * A hardware can handle a certain number of maximum requests at a given
> > > > > > > > > + * point. If such a constraint exists, set maxQueueRequests_ via
> > > > > > > > > + * \a setMaxQueueRequests() in the derived pipeline handler.
> > > > > > > > > + *
> > > > > > > > > + * The derived pipeline handler can choose not to define such constraint as
> > > > > > > > > + * well. In that case, the derived pipeline handler can avoid setting
> > > > > > > > > + * \a setMaxQueueReqeuests(), hence \a maxQueueRequests_ and
> > > > > > > > > + * \a requestsQueueCounter_ will be 0.
> > > > > > > > > + */
> > > > > > > > > +
> > > > > > > > > +/**
> > > > > > > > > + * \var PipelineHandler::requestsQueueCounter_
> > > > > > > > > + * \brief Number of requests queued to the underlying hardware
> > > > > > > > > + *
> > > > > > > > > + * If \a setMaxQueueRequests() is set by the derived pipeline handler,
> > > > > > > > > + * requestsQueueCounter_ reflects the number of requests queued
> > > > > > > > > + * to the underlying hardware by the pipeline handler.
> > > > > > > > > + */
> > > > > > > > > +
> > > > > > > > > +/**
> > > > > > > > > + * \brief Sets the maximum number of requests that can be queued
> > > > > > > > > + * \param[in] maxRequests Maximum number of in-flight requests
> > > > > > > > > + *
> > > > > > > > > + * A hardware can handle a certain number of requests at a given point.
> > > > > > > > > + * This function sets the maximum number of in-flight requests that can
> > > > > > > > > + * be queued to the hardware by the pipeline handler. Each derived pipeline
> > > > > > > > > + * handler should set the maximum number of in-flight requests it can handle
> > > > > > > > > + * at a given point using this function, if at all such a constraint exists.
> > > > > > > > > + */
> > > > > > > > > +void PipelineHandler::setMaxQueueRequests(uint32_t maxRequests)
> > > > > > > > > +{
> > > > > > > > > + maxQueueRequests_ = maxRequests;
> > > > > > > > > +}
> > > > > > > > > +
> > > > > > > > > /**
> > > > > > > > > * \fn PipelineHandler::name()
> > > > > > > > > * \brief Retrieve the pipeline handler name
> > > > > > > > > --
> > > > > > > > > 2.38.1
> > > > > > > > >
> > > > >
More information about the libcamera-devel
mailing list