AM625: V4L2 stream on can't get data

upload TI_AM625_ticket_packages.tgz from Seyond team as they can not upload that to this ticket. it includes dts, code change, log that collected by customer. 

https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/791/TI_5F00_AM625_5F00_ticket_5F00_packages.tgz

Hi Yong,

tom zhang said:

we use the comand:   v4l2-ctl -d /dev/video0 --stream-mmap --stream-count=10  to get stream data

but there is no data.

Can you share the log of running this command?

Also, please refer to this FAQ for debugging: What are the common reasons V4L2 based applications fail to capture images from a probed CSI sensor?

Regards,

Jianzhong

Dear Tom.

would you please provide the log as Jianzhong requested?

there is only dmesg log in the compressed file provided by you in the email, see my last post in this ticket. and it is not enough.

[ 0.491042] /bus@f0000/ticsi2rx@30102000/csi-bridge@30101000: Fixed dependency cycle(s) with /ads6311_dummy
[ 0.491132] /ads6311_dummy: Fixed dependency cycle(s) with /bus@f0000/ticsi2rx@30102000/csi-bridge@30101000

[ 1.089563] ads6311_dummy ads6311_dummy: ads6311_dummy_probe: start
[ 1.089606] ads6311_dummy ads6311_dummy: ads6311_dummy sensor registered successfully

2.176457] cdns-csi2rx 30101000.csi-bridge: Probed CSI2RX with 4/4 lanes, 4 streams, external D-PHY

4237.671233] ads6311_dummy ads6311_dummy: ads6311_dummy_s_stream: stream ON
[ 4240.084908] ads6311_dummy ads6311_dummy: ads6311_dummy_s_stream: stream OFF

thanks a lot!

yong

When I run the command :v4l2-ctl -d /dev/video0 --stream-mmap --stream-count=10  to get stream data

there is no any log return .

It seems to wait data

Hi Jianzhong,

I dumped the list of csi2rx registers configured during the stream on process. Could you help check if there are any configuration errors?

soft_reset -> 0x3

stream0_ctrl -> 0x10
stream1_ctrl -> 0x10
stream2_ctrl -> 0x10
stream3_ctrl -> 0x10

soft_reset -> 0x0

stream0_ctrl -> 0x0
stream1_ctrl -> 0x0
stream2_ctrl -> 0x0
stream3_ctrl -> 0x0

static_cfg -> 0x43210400

dphy_lane_control -> 0x1f01f

stream0_cfg -> 0x100
stream0_data_cfg -> 0x0
stream0_ctrl -> 0x1

stream1_cfg -> 0x100
stream1_data_cfg -> 0x0
stream1_ctrl -> 0x1

stream2_cfg -> 0x100
stream2_data_cfg -> 0x0
stream2_ctrl -> 0x1

stream3_cfg -> 0x100
stream3_data_cfg -> 0x0
stream3_ctrl -> 0x1

hello，the fmt type is RAW12, width=256, height=192

Dear Jianzhong.

Today we have a short discussion with SEYOND team.

1. sample drive code for new camera sensor as reference.

https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/791/0001_2D00_kernel_2D00_debug_2D00_ov_2D00_ox02c10.patch
2. check with SEYOND on SPAD output, virtual channel number, RAW12, height, width.
                 SPAD output: 2x channel, RAW8, ?(height) x ?(width), need SEYOND team double confirm.
3. register dump script for AM620-Q1 CSIRX I/F

am62x_csirx_reg_dump.gel

4. suggest to let SPAD to output one channel, then have a try firstly.

5. suggest to prepare environment, SK-AM62B-P1 and ALINX AN5641, then it can be reference.

3.2.2.2. CSI2RX — Processor SDK AM62x Documentation

6. More guidline for camera.

Use Camera — AM62x Academy

thanks a lot!

yong

Dear Tom.

one more question, is it final configuration?

thanks a lot!

yong

Hi Tom,

tom zhang said:

When I run the command :v4l2-ctl -d /dev/video0 --stream-mmap --stream-count=10  to get stream data

You'll need to provide the resolution and pixel format to this command. For example,

root@am62axx-evm:~/imx728# v4l2-ctl -d /dev/video-imx728-cam0 --set-fmt-video=width=3856,height=2176,pixelformat='RG12' --stream-mmap=6 --stream-count=100
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< 30.00 fps
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< 30.00 fps
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< 30.00 fps
<<<<<<<<

Can you share the output of "media-ctl -p"?

Thank you.

Jianzhong

One more thing. Since your SPAD driver is a dummy one, you should start sending from your SPAD sensor before you start capturing on AM625.

sorry,the format is SBGGR12, 792*4096

update log from Tom, Seyond team.

Here is the media-ctl -p log

Media controller API version 6.6.58

Media device information
------------------------
driver          j721e-csi2rx
model           TI-CSI2RX
serial          
bus info        platform:30102000.ticsi2rx
hw revision     0x1
driver version  6.6.58

Device topology
- entity 1: 30102000.ticsi2rx (5 pads, 5 links, 1 route)
            type V4L2 subdev subtype Unknown flags 0
            device node name /dev/v4l-subdev0
	routes:
		0/0 -> 1/0 [ACTIVE]
	pad0: Sink
		[stream:0 fmt:UYVY8_1X16/640x480 field:none colorspace:srgb xfer:srgb ycbcr:601 quantization:lim-range]
		<- "cdns_csi2rx.30101000.csi-bridge":1 [ENABLED,IMMUTABLE]
	pad1: Source
		[stream:0 fmt:UYVY8_1X16/640x480 field:none colorspace:srgb xfer:srgb ycbcr:601 quantization:lim-range]
		-> "30102000.ticsi2rx context 0":0 [ENABLED,IMMUTABLE]
	pad2: Source
		-> "30102000.ticsi2rx context 1":0 [ENABLED,IMMUTABLE]
	pad3: Source
		-> "30102000.ticsi2rx context 2":0 [ENABLED,IMMUTABLE]
	pad4: Source
		-> "30102000.ticsi2rx context 3":0 [ENABLED,IMMUTABLE]

- entity 7: cdns_csi2rx.30101000.csi-bridge (5 pads, 2 links, 1 route)
            type V4L2 subdev subtype Unknown flags 0
            device node name /dev/v4l-subdev1
	routes:
		0/0 -> 1/0 [ACTIVE]
	pad0: Sink
		[stream:0 fmt:UYVY8_1X16/640x480 field:none colorspace:srgb xfer:srgb ycbcr:601 quantization:lim-range]
		<- "ads6311_dummy":0 [ENABLED,IMMUTABLE]
	pad1: Source
		[stream:0 fmt:UYVY8_1X16/640x480 field:none colorspace:srgb xfer:srgb ycbcr:601 quantization:lim-range]
		-> "30102000.ticsi2rx":0 [ENABLED,IMMUTABLE]
	pad2: Source
	pad3: Source
	pad4: Source

- entity 13: ads6311_dummy (1 pad, 1 link, 0 routes)
             type V4L2 subdev subtype Sensor flags 0
             device node name /dev/v4l-subdev2
	pad0: Source
		-> "cdns_csi2rx.30101000.csi-bridge":0 [ENABLED,IMMUTABLE]

- entity 19: 30102000.ticsi2rx context 0 (1 pad, 1 link)
             type Node subtype V4L flags 0
             device node name /dev/video0
	pad0: Sink
		<- "30102000.ticsi2rx":1 [ENABLED,IMMUTABLE]

- entity 25: 30102000.ticsi2rx context 1 (1 pad, 1 link)
             type Node subtype V4L flags 0
             device node name /dev/video1
	pad0: Sink
		<- "30102000.ticsi2rx":2 [ENABLED,IMMUTABLE]

- entity 31: 30102000.ticsi2rx context 2 (1 pad, 1 link)
             type Node subtype V4L flags 0
             device node name /dev/video2
	pad0: Sink
		<- "30102000.ticsi2rx":3 [ENABLED,IMMUTABLE]

- entity 37: 30102000.ticsi2rx context 3 (1 pad, 1 link)
             type Node subtype V4L flags 0
             device node name /dev/video3
	pad0: Sink
		<- "30102000.ticsi2rx":4 [ENABLED,IMMUTABLE]

thanks a lot!

yong

- entity 13: ov5640 4-003c (1 pad, 1 link, 0 routes)
type V4L2 subdev subtype Sensor flags 0
device node name /dev/v4l-subdev2
pad0: Source
[stream:0 fmt:UYVY8_1X16/640x480@1/30 field:none colorspace:srgb xfer:srgb ycbcr:601 quantization:full-range
crop.bounds:(0,0)/2624x1964
crop:(16,14)/2592x1944]
-> "cdns_csi2rx.30101000.csi-bridge":0 [ENABLED,IMMUTABLE]

This is the new media-ctl -p log

We have set the format

entity 13: ov5640 4-003c (1 pad, 1 link, 0 routes)

0 routes is right?

Dear Tom,

sorry, please expect delay as weekend.

and would you please copy full log to this ticket to easy analysis for TI expert?

thanks a lot!

yong

Dear Jianzhong.

customer feedback that current status is that there is only below difference in new log for "media-ctl -p".

please help check and give suggestion.

- entity 13: ov5640 4-003c (1 pad, 1 link, 0 routes)
type V4L2 subdev subtype Sensor flags 0
device node name /dev/v4l-subdev2
pad0: Source
[stream:0 fmt:UYVY8_1X16/640x480@1/30 field:none colorspace:srgb xfer:srgb ycbcr:601 quantization:full-range
crop.bounds:(0,0)/2624x1964
crop:(16,14)/2592x1944]
-> "cdns_csi2rx.30101000.csi-bridge":0 [ENABLED,IMMUTABLE]

thanks a lot!

yong

Hi Yong,

Sorry, but I don't understand you or Tom. What you shared is from my log.

Thanks,

Jianzhong

Hi Jianzhong：

We have get the data from video device.So don't care about the previous log of media-ctl -p

Hi:

we have some problems about these registers:stream0_cfg,stream0_data_cfg.

Could you please tell us more details about the two registers to help us understand.

Hi Tom,

What specific questions do you have regarding those two registers? Please refer to the Cadence CSI-Rx driver regarding how those registers are configured.

Regards,

Jianzhong

When we use the command to get frame data, it seems drop some frames.

Is the fps too fast so that it drops some frames.

v4l2-ctl --verbose -d /dev/video0 --set-fmt-video=width=792,height=129,pixelformat='BG12' --stream-mmap=6 --stream-skip=1 --stream-to=/tmp/792x129_SBGGR12_1X12_VC0.raw12 --stream-count=10 --stream-poll

cap dqbuf: 0 seq: 0 bytesused: 204336 ts: 70.256985 field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 1 seq: 1 bytesused: 204336 ts: 70.258692 delta: 1.707 ms field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 2 seq: 2 bytesused: 204336 ts: 70.262145 delta: 3.453 ms field: Any (ts-monotonic, ts-src-eof)
cap dqbuf: 3 seq: 3 bytesused: 204336 ts: 70.263879 delta: 1.734 ms field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 4 seq: 4 bytesused: 204336 ts: 70.265613 delta: 1.734 ms field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 5 seq: 5 bytesused: 204336 ts: 70.267323 delta: 1.710 ms field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 0 seq: 6 bytesused: 204336 ts: 70.270791 delta: 3.468 ms dropped: 3 field: Any (ts-monotonic, ts-src-eof)
cap dqbuf: 1 seq: 7 bytesused: 204336 ts: 70.272529 delta: 1.738 ms field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 2 seq: 8 bytesused: 204336 ts: 70.274242 delta: 1.713 ms field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 3 seq: 9 bytesused: 204336 ts: 70.275975 delta: 1.733 ms field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 4 seq: 10 bytesused: 204336 ts: 70.279432 delta: 3.457 ms dropped: 3 field: Any (ts-monotonic, ts-src-eof)
cap dqbuf: 5 seq: 11 bytesused: 204336 ts: 70.282891 delta: 3.459 ms field: Any (ts-monotonic, ts-src-eof)
cap dqbuf: 0 seq: 12 bytesused: 204336 ts: 70.284625 delta: 1.734 ms field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 1 seq: 13 bytesused: 204336 ts: 70.286327 delta: 1.702 ms field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 2 seq: 14 bytesused: 204336 ts: 70.289796 delta: 3.469 ms fps: 433.98 dropped: 2 field: Any (ts-monotonic, ts-src-eof)
cap dqbuf: 3 seq: 15 bytesused: 204336 ts: 70.291524 delta: 1.728 ms fps: 433.98 field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 4 seq: 16 bytesused: 204336 ts: 70.294984 delta: 3.460 ms fps: 426.32 dropped: 1 field: Any (ts-monotonic, ts-src-eof)
cap dqbuf: 5 seq: 17 bytesused: 204336 ts: 70.300177 delta: 5.193 ms fps: 394.40 field: Any (ts-monotonic, ts-src-eof)
cap dqbuf: 0 seq: 18 bytesused: 204336 ts: 70.301909 delta: 1.732 ms fps: 394.40 field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 1 seq: 19 bytesused: 204336 ts: 70.303648 delta: 1.739 ms fps: 394.40 field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 2 seq: 20 bytesused: 204336 ts: 70.305374 delta: 1.726 ms fps: 394.40 field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 3 seq: 21 bytesused: 204336 ts: 70.307088 delta: 1.714 ms fps: 394.40 field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 4 seq: 22 bytesused: 204336 ts: 70.308833 delta: 1.745 ms fps: 394.40 field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 5 seq: 23 bytesused: 204336 ts: 70.312276 delta: 3.443 ms fps: 418.90 dropped: 5 field: Any (ts-monotonic, ts-src-eof)
cap dqbuf: 0 seq: 24 bytesused: 204336 ts: 70.315730 delta: 3.454 ms fps: 410.56 field: Any (ts-monotonic, ts-src-eof)
cap dqbuf: 1 seq: 25 bytesused: 204336 ts: 70.317476 delta: 1.746 ms fps: 410.56 field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 2 seq: 26 bytesused: 204336 ts: 70.319192 delta: 1.716 ms fps: 410.56 field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 3 seq: 27 bytesused: 204336 ts: 70.322664 delta: 3.472 ms fps: 413.09 dropped: 2 field: Any (ts-monotonic, ts-src-eof)
cap dqbuf: 4 seq: 28 bytesused: 204336 ts: 70.324393 delta: 1.729 ms fps: 413.09 field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 5 seq: 29 bytesused: 204336 ts: 70.326122 delta: 1.729 ms fps: 413.09 field: Any (error, ts-monotonic, ts-src-eof)
cap dqbuf: 0 seq: 30 bytesused: 204336 ts: 70.329578 delta: 3.456 ms fps: 415.23 dropped: 2 field: Any (ts-monotonic, ts-src-eof)

So your sensor is 792x129, 12-bit. What's the frame rate of the sensor output?

Even at 500fps, the total data rate is about 735Mbps including overhead. The SoC should have no problem receiving this data rate.

 The frame rate of the sensor output should be 440 fps.

Can you try not to save data to a file? File I/O may slow down the captre.

For example, just do: v4l2-ctl --verbose -d /dev/video0 --set-fmt-video=width=792,height=129,pixelformat='BG12' --stream-mmap=6

I have tested, it stiil drop frames.

When I test the format:792*4128. The FPS maybe 14 FPS.when I test with the commad: v4l2-ctl --verbose -d /dev/video0 --set-fmt-video=width=792,height=4128,pixelformat='BG12' --stream-mmap=6 --stream-skip=1 --stream-to=/tmp/test_frame.raw --stream-count=10 --stream-pollI

It still drop frames.

But when I set the FPS lower，then use the same command to get frames, it won't drop frames.

Can you use memdev2 to dump the following registers when you had frames drops and when you didn't have frame drops?

CSIRX_MONITOR_IRQS                                            (30101018h)

CSIRX_MONITOR_IRQS_MASK_CFG                      (3010101Ch)

CSIRX_ERROR_IRQS                                                 (30101028h)

CSIRX_ERROR_IRQS_MASK_CFG                            (3010102Ch)

CSIRX_DPHY_STATUS                                              (30101048h)

CSIRX_DPHY_ERR_STATUS_IRQ                             (3010104Ch)

CSIRX_DPHY_ERR_IRQ_MASK_CFG                        (30101050h)

frames drop:

30101018h = 0x00000000

3010101C =  0x00000000

30101028 = 0x00020910

3010102C =  0x00000000

30101048 =  0x00222206

3010104C =  0x00000000

30101050 =  0x00000000

frames doesn't drop:

30101018h = 0x00000000

3010101C =  0x00000000

30101028 =  0x00020100

3010102C =  0x00000000

30101048 =  0x00333306

3010104C =  0x00000000

30101050 =  0x00000000

792*4128 is the test format.

The format we want to use is 792*4608. The FPS will be higher.

So if the frame loss issue is related to the frame rate，792*4608 will also drop frames.

Please help us resolve this issue

Dear Tom.

as we aligned yesterday, please help clarify the real requirement on your side. here is my record, please help correct that.

1. 1. 792x4608, 10FPS
   2. 792x9, 512FPS

and we will start from first resolution/FPS. would you please help provide register dump again for this?

thanks a lot!

yong

b,should be 792x9 5120 FPS

tom zhang said:

30101028 = 0x00020910

This indicates CRC error:

tom zhang said:

30101048 =  0x00222206

This indicates some issues with D-PHY data lane:

tom zhang said:

When I test the format:792*4128. The FPS maybe 14 FPS.

With this frame size and frame rate, the data rate is about 0.66Gbps (including protocol overhead). Your device tree overlay enables 4 data lanes, so data rate per lane is about 165Mbps. 

Do you know the link frequency used by the sensor? I don't see it in the device tree overlay.

Regards,

Jianzhong

the link frequency is 1G. We use 2 channels to test.

The data rate shoule be 1.396Gbps, each lane should be 349Mbps

If we set the link frequency as 1.5G，it can't get any frames.

when I run the command v4l2-ctl -d /dev/video0 --set-fmt-video=width=792,height=4128,pixelformat='BG12' --stream-mmap=6 --stream-skip=1 --stream-to=/tmp/test_frame.raw --stream-count=10 --stream-pollI

it returns select timeout.

I tried to set stream-mmap=32 or not dump to file, it also returns select timeout.

why?

tom zhang said:

If we set the link frequency as 1.5G，it can't get any frames.

Please see my response on the other thread here: e2e.ti.com/.../am625-can-not-receive-any-data-from-mipi

Can you tell me the reason about frames drops (792x4128)?

Hi Tom

In our visit today, we found frame drops with customer board but none with TI EVM at 1.2/1.4Gbps. You may check with your colleagues and also make similar test with TI EVM. This frame drop may have something todo with board design.

Regards,

Adam

Dear Tom.

would you please help share the status on your side? what do you need support from TI?

thanks a lot!

yong

Dear Tom.

please let us know if you still need support on this ticket. Or we can close this ticket in this week, right?

thanks a lot!

yong

Dear Tom.

we will close this ticket. and please notice HW review is necessary for CSIRX as low speed only is ok.

thanks a lot!

yong

Hi Yong:

Ok, thank you for your help.