Adv7282 PAL to MIPI CSI Linux Driver Development

kakeh · December 1, 2021, 1:27pm

cd $HOME/nvidia/sources/Linux_for_Tegra/source/public/kernel/kernel-4.9/ &&
TEGRA_KERNEL_OUT=$HOME/nvidia/final_out/ &&
LOCALVERSION=-tegra &&
mkdir -p $TEGRA_KERNEL_OUT &&
make O=$TEGRA_KERNEL_OUT tegra_defconfig &&
make O=$TEGRA_KERNEL_OUT -j6 &&
sudo cp $TEGRA_KERNEL_OUT/arch/arm64/boot/Image /boot/;
sudo cp $TEGRA_KERNEL_OUT/arch/arm64/boot/dts/tegra194-p3668-all-p3509-0000.dtb /boot/dtb/

Just building the Image and dtb through this process, shall update results tomorrow.

I asusmed menuconfig is doing the issue, i tried with out it but did not work
or anything else i am missing in specific ? what log would identify this issue ?

ShaneCCC · December 1, 2021, 3:32pm

Should be the command “export LOCALVERSION=-tegra” for setting building configure.

kakeh · December 2, 2021, 5:54am

export prefix worked.

i have a doubt here in the i2c0 of device tree what is the register address to be used ? the existing imx219 device tree uses 0x10 , in the code i have adopted its 0x21 , what is the correct i2c address of cam0 of xavier nx development board

ShaneCCC · December 2, 2021, 6:24am

You need consult with device vendor to know the slave address.

kakeh · December 2, 2021, 6:35am

i think you have got me wrong, i am speaking about the i2c address which is present in device tree of file
tegra194-camera-rbpcv2-imx219.dtsi

like below, what is the i2c address mentioned below, isnt it the address of the i2c mux present on tegra ?

cam_i2cmux {
		i2c_0:i2c@0 {
			status = "okay";				
			imx219_cam0: rbpcv2_imx219_a@10 {
				compatible = "adi,adv7282-m";
				/* I2C device address */
				reg = <0x10>;

				/* V4L2 device node location */
				devnode = "video0";

				/* Physical dimensions of sensor */
                physical_w = "4.713";
                physical_h = "3.494";

				sensor_model = "adv7282-m";

				use_sensor_mode_id = "false";

kakeh · December 2, 2021, 6:41am

to breif more below is my schematic of connector and 7282

ShaneCCC · December 2, 2021, 6:45am

The i2c address here is the slave address of ADV7282 doesn’t matter with the mux.

kakeh · December 2, 2021, 9:06am

i was able to see video node in my /dev/ but the main problem observed from dmesg logs is there is no ack from the address 0x21(the address is correct for the device)
if you see my daughter board schematic i did not put any pullup on my board, i have seen 47k pulled up to 3.3v on both scl and sda lines of carrier board scheamtic, please confirm if i require pullups on my daughter board too ?

below is the snap of standard carrier board schematic

Trumany · December 2, 2021, 1:43pm

You can use an oscilloscope to observe the I2C waveform to confirm if it is correct. Basically no other pull-up is necessary.

kakeh · December 4, 2021, 6:56am

actual slave address as per datasheet is 0x43 but i get respose with address being 0x44 using below command
i2cget -y 2 0x44 0x11

is it like the i2c get includes r/w bit also in the address which is why read address is actually 0x44 ?

surprisingly the command i2cdetect over channel 2 doesnt detect chip what info the i2c detect will try to see ?

kakeh · December 7, 2021, 12:42pm

the device is responding to i2cget -y 2 0x21 , but not for i2c detect
i understood that the adv devices mostly wont respond to i2cdetect as they need write followed by read (assumption)
leaving i2cdetect issue aside

i went forward , i was able to see the appearance of video0 node in /dev
i went to see dmesg | grep 7282 the log is as follows

Line 594: [    2.158727] adv7282 9-0021: chip found @ 0x21 (i2c-2-mux (chan_id 0))
Line 607: [    2.176037] adv7282: probe of 9-0021 failed with error -121
Line 608: [    2.176116] adv7282 10-0021: chip found @ 0x21 (i2c-2-mux (chan_id 1))
Line 610: [    2.197643] adv7282 10-0021: adv7282-m probed!
Line 610: [    2.197643] adv7282 10-0021: adv7282-m probed!
Line 816: [    3.179292] tegra194-vi5 15c10000.vi: subdev adv7282 10-0021 bound

so i can clearly see the device is probed sucessfully and video node is created

i went forward to see whether i can open the cheese app,it crashed
the dmesg then read with latest error as below

Line 1057: [  126.993369] tegra194-vi5 15c10000.vi: corr_err: discarding frame 1, flags: 0, err_data 64

What could be the issues ?
I am attached the dmesg log also, kindly suggest where exactly i am doing wrong.
log.txt (69.7 KB)

ShaneCCC · December 7, 2021, 12:50pm

Using v4l2-ctl and get the trace log to check.

https://elinux.org/Jetson/l4t/Camera_BringUp

kakeh · December 7, 2021, 1:10pm

[  126.982514] [RCE] vi5_hwinit: firmware CL2018101701 protocol version 2.2
[  126.993369] tegra194-vi5 15c10000.vi: corr_err: discarding frame 1, flags: 0, err_data 64
[  127.094630] [RCE] WARNING: t194/isp5.c:901 [config_channel] "All error notifications not enabled: correctable=0x00 uncorrectable=0x00"
[ 1010.279683] tegradc 15200000.nvdisplay: blank - powerdown
[ 1010.340088] extcon-disp-state external-connection:disp-state: cable 47 state 0
[ 1010.340093] Extcon AUX1(HDMI) disable
[ 1010.362677] tegra_nvdisp_handle_pd_disable: Powergated Head1 pd
[ 1010.363213] tegra_nvdisp_handle_pd_disable: Powergated Head0 pd
[ 1250.786347] tegradc 15200000.nvdisplay: unblank
[ 1250.789496] tegra_nvdisp_handle_pd_enable: Unpowergated Head0 pd
[ 1250.790038] tegra_nvdisp_handle_pd_enable: Unpowergated Head1 pd
[ 1250.799203] Parent Clock set for DC plld2
[ 1250.803686] tegradc 15200000.nvdisplay: hdmi: tmds rate:148500K prod-setting:prod_c_hdmi_111m_223m
[ 1250.805107] tegradc 15200000.nvdisplay: hdmi: get YCC quant from EDID.
[ 1250.841632] extcon-disp-state external-connection:disp-state: cable 47 state 1
[ 1250.841637] Extcon AUX1(HDMI) enable
[ 1250.860333] tegradc 15200000.nvdisplay: unblank
[ 1250.860349] tegradc 15210000.nvdisplay: blank - powerdown
[ 1637.863589] tegra194-vi5 15c10000.vi: =================  START STATUS  =================
[ 1637.863597] tegra194-vi5 15c10000.vi: ==================  END STATUS  ==================
[ 1637.867148] adv7282 10-0021: adv7180_g_dv_timings():
[ 1637.874207] Unable to handle kernel NULL pointer dereference at virtual address 000001e8
[ 1637.874379] Mem abort info:
[ 1637.874435]   ESR = 0x96000005
[ 1637.874491]   Exception class = DABT (current EL), IL = 32 bits
[ 1637.874590]   SET = 0, FnV = 0
[ 1637.874645]   EA = 0, S1PTW = 0
[ 1637.874702] Data abort info:
[ 1637.874753]   ISV = 0, ISS = 0x00000005
[ 1637.874819]   CM = 0, WnR = 0
[ 1637.874879] user pgtable: 4k pages, 39-bit VAs, pgd = ffffffc1330a6000
[ 1637.875089] [00000000000001e8] *pgd=0000000000000000, *pud=0000000000000000
[ 1637.875235] Internal error: Oops: 96000005 [#1] PREEMPT SMP
[ 1637.875338] Modules linked in: fuse zram overlay bnep rtl8822ce rtk_btusb btusb btrtl btbcm btintel cfg80211 spidev userspace_alert nvgpu bluedroid_pm ip_tables x_tables
[ 1637.875756] CPU: 0 PID: 11880 Comm: v4l2-compliance Tainted: G        W       4.9.201-tegra #12
[ 1637.876404] Hardware name: NVIDIA Jetson Xavier NX Developer Kit (DT)
[ 1637.876897] task: ffffffc1908cc600 task.stack: ffffffc1516d8000
[ 1637.877364] PC is at tegra_vi5_g_volatile_ctrl+0x48/0x120
[ 1637.881784] LR is at tegra_vi5_g_volatile_ctrl+0x2c/0x120
[ 1637.887034] pc : [<ffffff8008b49268>] lr : [<ffffff8008b4924c>] pstate: 40400145
[ 1637.894644] sp : ffffffc1516dbac0
[ 1637.897623] x29: ffffffc1516dbac0 x28: ffffffc1e0fd2800 
[ 1637.903229] x27: 000000000f000000 x26: 0000000000000000 
[ 1637.908820] x25: ffffff800a01f750 x24: 0000000000000047 
[ 1637.914067] x23: 0000000000000001 x22: 0000000000000018 
[ 1637.919141] x21: ffffffc1516dbb60 x20: ffffffc1e2aaf820 
[ 1637.924827] x19: ffffffc1e0fd2800 x18: 0000000000000001 
[ 1637.930429] x17: 0000007f83021530 x16: ffffff8008272650 
[ 1637.935867] x15: ffffffffffffffff x14: ffffffc1516dbac0 
[ 1637.941555] x13: ffffffc1516db9c5 x12: 0000000000000000 
[ 1637.947491] x11: ffffffc1516db980 x10: ffffffc1516db980 
[ 1637.953181] x9 : 0000000000000002 x8 : 0000000000000002 
[ 1637.958954] x7 : ffffff8008fc7100 x6 : 0000000000000090 
[ 1637.964215] x5 : 000000000000008d x4 : 0000000000000001 
[ 1637.969804] x3 : 00000000009a2033 x2 : 0000000000000000 
[ 1637.975140] x1 : ffffffc1908cc600 x0 : 00000000009a2032

I assume the problem is at timings ?

ShaneCCC · December 7, 2021, 2:34pm

It should be the driver cause the kernel crash. You may need to debug to fix it first.

kakeh · December 8, 2021, 4:10am

i hardcoded the resolution in side the driver code which did not cause the previous error.
and ran below command

v4l2-ctl --device /dev/video0 --stream-mmap --stream-to=frame.raw --stream-count=1

i was able to avoid crash , now i am seeing below error

tegra194-vi5 15c10000.vi: corr_err: discarding frame 2, flags: 0, err_data 64
tegra194-vi5 15c10000.vi: corr_err: discarding frame 1, flags: 0, err_data 131072

ShaneCCC · December 8, 2021, 4:15am

What’s the trace log shows.

kakeh · December 8, 2021, 4:32am

i followed the guide you provided to drag trace log
i personally was not able to pin point the issue here
attached is the log.
trace_log.txt (619.4 KB)

ShaneCCC · December 8, 2021, 4:34am

Still see the short frame in the trace log. You may need to check more output size.

kakeh · December 8, 2021, 4:57am

i assume you are asking to increase the stream count and trace log file buffer size
made count to 5

v4l2-ctl --device /dev/video0 --stream-mmap --stream-to=frame.raw --stream-count=5

i was not able to see the process end at all, may be due to the errors caused, but i waited for atleast 2mins and then killed the v4l2 process(Ctrl+C)

attached are the log with increased buffer size, use gzip to extract.
trace_log3.txt.gz (600.0 KB)

ShaneCCC · December 8, 2021, 5:00am

I mean the sensor driver report correct resolution(output size) to fix the short frame issue.