On Tue, 22 Aug 2017 19:43:39 +0200
Maxime Ripard [off-list ref] wrote:

Hi Yong,

On Mon, Jul 31, 2017 at 11:16:40AM +0800, Yong wrote:

quoted

quoted

quoted

@@ -143,6 +143,7 @@ source "drivers/media/platform/am437x/Kconfig"
 source "drivers/media/platform/xilinx/Kconfig"
 source "drivers/media/platform/rcar-vin/Kconfig"
 source "drivers/media/platform/atmel/Kconfig"
+source "drivers/media/platform/sun6i-csi/Kconfig"

We're going to have several different drivers in v4l eventually, so I
guess it would make sense to move to a directory of our own.

Like this?
drivers/media/platform/sunxi/sun6i-csi

Yep.

quoted

quoted

quoted

+static int sun6i_graph_notify_complete(struct v4l2_async_notifier *notifier)
+{
+	struct sun6i_csi *csi =
+			container_of(notifier, struct sun6i_csi, notifier);
+	struct sun6i_graph_entity *entity;
+	int ret;
+
+	dev_dbg(csi->dev, "notify complete, all subdevs registered\n");
+
+	/* Create links for every entity. */
+	list_for_each_entry(entity, &csi->entities, list) {
+		ret = sun6i_graph_build_one(csi, entity);
+		if (ret < 0)
+			return ret;
+	}
+
+	/* Create links for video node. */
+	ret = sun6i_graph_build_video(csi);
+	if (ret < 0)
+		return ret;

Can you elaborate a bit on the difference between a node parsed with
_graph_build_one and _graph_build_video? Can't you just store the
remote sensor when you build the notifier, and reuse it here?

There maybe many usercases:
1. CSI->Sensor.
2. CSI->MIPI->Sensor.
3. CSI->FPGA->Sensor1
            ->Sensor2.
FPGA maybe some other video processor. FPGA, MIPI, Sensor can be
registered as v4l2 subdevs. We do not care about the driver code
of them. But they should be linked together here.

So, the _graph_build_one is used to link CSI port and subdevs. 
_graph_build_video is used to link CSI port and video node.

So the graph_build_one is for the two first cases, and the
_build_video for the latter case?

No. 
The _graph_build_one is used to link the subdevs found in the device 
tree. _build_video is used to link the closest subdev to video node.
Video node is created in the driver, so the method to get it's pad is
diffrent to the subdevs.

If so, you should take a look at the last iteration of the
subnotifiers rework by Nikas S?derlund (v4l2-async: add subnotifier
registration for subdevices).

It allows subdevs to register notifiers, and you don't have to build
the graph from the video device, each device and subdev can only care
about what's next in the pipeline, but not really what's behind it.

That would mean in your case that you can only deal with your single
CSI pad, and whatever subdev driver will use it care about its own.

Do you mean the subdevs create pad link in the notifier registered by
themself ?
If so, _graph_build_one is needless. But how to make sure the pipeline
has linked correctly when operateing the pipeline. 
I will lookt at this in more detail.

quoted

This part is also difficult to understand for me. The one CSI module
have only one DMA channel(single port). But thay can be linked to 
different physical port (Parallel or MIPI)(multiple ep) by IF select
register.

For now, the binding can have several ep, the driver will just pick
the first valid one.

Yeah, I'm not really sure how we could deal with that, but I guess we
can do it later on.

quoted

quoted

quoted

+struct sun6i_csi_ops {
+	int (*get_supported_pixformats)(struct sun6i_csi *csi,
+					const u32 **pixformats);
+	bool (*is_format_support)(struct sun6i_csi *csi, u32 pixformat,
+				  u32 mbus_code);
+	int (*s_power)(struct sun6i_csi *csi, bool enable);
+	int (*update_config)(struct sun6i_csi *csi,
+			     struct sun6i_csi_config *config);
+	int (*update_buf_addr)(struct sun6i_csi *csi, dma_addr_t addr);
+	int (*s_stream)(struct sun6i_csi *csi, bool enable);
+};

Didn't we agreed on removing those in the first iteration? It's not
really clear at this point whether they will be needed at all. Make
something simple first, without those ops. When we'll support other
SoCs we'll have a better chance at seeing what and how we should deal
with potential quirks.

OK. But without ops, it is inappropriate to sun6i_csi and sun6i_csi.
Maybe I should merge the two files.

I'm not sure what you meant here, but if you think that's appropriate,
please go ahead.

quoted

quoted

quoted

+		return IRQ_HANDLED;
+	}
+
+	if (status & CSI_CH_INT_STA_FD_PD) {
+		sun6i_video_frame_done(&sdev->csi.video);
+	}
+
+	regmap_write(regmap, CSI_CH_INT_STA_REG, status);

Isn't it redundant with the one you did in the condition a bit above?

You should also check that your device indeed generated an
interrupt. In the occurence of a spourious interrupt, your code will
return IRQ_HANDLED, which is the wrong thing to do.

I think you should reverse your logic a bit here to make this
easier. You should just check that your status flags are indeed set,
and if not just bail out and return IRQ_NONE.

And if they are, go on with treating your interrupt.

OK. I will add check for status flags.
BTW, how can a spurious interrupt occurred?

Usually it's either through some interference, or some poorly designed
controller. This is unlikely, but it's something you should take into
account.

Thanks!
Maxime

-- 
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com

Thanks,
Yong