Re: [PATCH v3 0/3] media: pci: intel: ivsc: Add driver of Intel Visual Sensing Controller(IVSC)

[Sakari Ailus <sakari.ailus@linux.intel.com>]

Hi Wentong,

On Mon, Mar 27, 2023 at 08:13:29AM +0000, Wu, Wentong wrote:
> 
> 
> > -----Original Message-----
> > From: Sakari Ailus <sakari.ailus@linux.intel.com>
> > Sent: Monday, March 27, 2023 3:50 PM
> > 
> > Hi Wentong,
> > 
> > On Mon, Mar 27, 2023 at 07:33:48AM +0000, Wu, Wentong wrote:
> > >
> > >
> > > > -----Original Message-----
> > > > From: Sakari Ailus <sakari.ailus@linux.intel.com>
> > > > Sent: Monday, March 27, 2023 3:21 PM
> > > >
> > > > Hi Wentong,
> > > >
> > > > On Mon, Mar 27, 2023 at 02:23:05PM +0800, Wentong Wu wrote:
> > > > > Intel Visual Sensing Controller (IVSC), codenamed "Clover Falls",
> > > > > is a companion chip designed to provide secure and low power
> > > > > vision capability to IA platforms. IVSC is available in existing
> > > > > commercial platforms from multiple OEMs.
> > > > >
> > > > > The primary use case of IVSC is to bring in context awareness.
> > > > > IVSC interfaces directly with the platform main camera sensor via
> > > > > a CSI-2 link and processes the image data with the embedded AI
> > > > > engine. The detected events are sent over I2C to ISH (Intel Sensor
> > > > > Hub) for additional data fusion from multiple sensors. The fusion
> > > > > results are used to implement advanced use cases like:
> > > > >  - Face detection to unlock screen
> > > > >  - Detect user presence to manage backlight setting or waking up
> > > > > system
> > > > >
> > > > > Since the Image Processing Unit(IPU) used on the host processor
> > > > > needs to configure the CSI-2 link in normal camera usages, the
> > > > > CSI-2 link and camera sensor can only be used in
> > > > > mutually-exclusive ways by host IPU and IVSC. By default the IVSC
> > > > > owns the CSI-2 link and camera sensor. The IPU driver can take
> > > > > ownership of the CSI-2 link and camera sensor using interfaces exported
> > via v4l2 sub-device.
> > > > >
> > > > > Switching ownership requires an interface with two different
> > > > > hardware modules inside IVSC. The software interface to these
> > > > > modules is via Intel MEI (The Intel Management Engine) commands.
> > > > > These two hardware modules have two different MEI UUIDs to
> > > > > enumerate. These hardware
> > > > modules are:
> > > > >  - ACE (Algorithm Context Engine): This module is for algorithm
> > > > > computing when IVSC owns camera sensor. Also ACE module controls
> > > > > camera sensor's ownership. This hardware module is used to set
> > > > > ownership of
> > > > camera sensor.
> > > > >  - CSI (Camera Serial Interface): This module is used to route
> > > > > camera sensor data either to IVSC or to host for IPU driver and application.
> > > > >
> > > > > IVSC also provides a privacy mode. When privacy mode is turned on,
> > > > > camera sensor can't be used. This means that both ACE and host IPU
> > > > > can't get image data. And when this mode is turned on, users are
> > > > > informed via
> > > > > v4l2 control API.
> > > > >
> > > > > In summary, to acquire ownership of camera by IPU driver, first
> > > > > ACE module needs to be informed of ownership and then to setup
> > > > > MIPI CSI-2 link for the camera sensor and IPU.
> > > > >
> > > > > Implementation:
> > > > > There are two different drivers to handle ACE and CSI hardware
> > > > > modules inside IVSC.
> > > > >  - ivsc_csi: MEI client driver to send commands and receive
> > > > > notifications from CSI module.
> > > > >  - ivsc_ace: MEI client driver to send commands and get status
> > > > > from ACE module.
> > > > > Interface is exposed via v4l2 sub-devcie APIs to acquire and
> > > > > release camera sensor and CSI-2 link.
> > > >
> > > > Thanks for the update.
> > > >
> > > > Could you elaborate the decision of keeping the csi_bridge entirely
> > > > separate from the cio2_bridge (to be turned to ipu_bridge first)?
> > > > Both are doing essentially the same and using the same data structures,
> > aren't they?
> > >
> > > yes, they're doing same thing to bridge the software nodes needed by
> > > v4l2, but they have different type devices(pci and mei_client device)
> > > and dependency. And they have same SSDB definition in DSDT, so the
> > > structures are almost same.
> > 
> > If there are differences, what are they?
> > 
> The properties of swnode are same, because we need same info to configure
> CSI2. But csi bridge can get sensor connected to IVSC by dependency info
> instead of just polling the only supported sensors for ipu. Also IVSC

It's totally fine the method for finding the sensor is different, this
should be an extra argument for the bridge init function and a small number
of lines of additional code.

> doesn't need vcm. And the topology of dependency is different.

The VCM is optional for current cio2_bridge, too.

It also seems that this currently creates swnodes on mei csi side only for
the connection towards the sensor, not the IPU. At the moment you can't
reliably add software nodes to a device that has may be already probing so
both of these should be created at the same time.

> 
> > What comes to cio2_bridge, the fact that it's related to a PCI device doesn't
> > seem to matter after initialisation so it could as well work with struct device.
> > 
> > >
> > > I have no idea what the ipu bridge would be like, but IVSC csi bridge
> > > can be configured via kconfig to enable/disable.
> > 
> > Please work out the details with Bingbu.
> 
> @bingbu.cao@linux.intel.com @Sakari Ailus Please share your design idea here.
> What the ipu bridge would be like?  What's the responsibility of ipu bridge?

I'd expect Bingbu to have patches to turn the current cio2_bridge to an
ipu_bridge at some point. These should come on top of those patches.

-- 
Kind regards,

Sakari Ailus