PINE64 Don't be evil devkit (pine64-dontbeevil)
There is a 3.3V uart on the 2x5pin connector near the modem. The pin in the corner toward VOL- is GND and the next pin is TX.
The baudrate in u-boot and in Linux is 115200n8
|Rear camera||OmniVision OV5640||Yes, ov5640|
|Front camera||GalaxyCore GC2035||Nope|
|Ambient light / Proximity||SensorTek STK3335||works with stk3310|
The touchscreen is controlled by the
edt_ft5x06 module on i2c-0. The current issue is that it takes a long while before the i2c-0 bus works after booting. For some reason it takes ~110 seconds before communication is possible. This causes an issue because the module tries to probe the touchscreen before the bus works.
The current workaround is to build
edt_ft5x06 as a module and rmmod/modprobe the module after the i2c-0 bus works so it gets probed again.
[ 17.942111] edt_ft5x06 0-0038: touchscreen probe failed ...after rmmod edt_ft5x06; modprobe edt_ft5x06... [ 3989.326377] input: generic ft5x06 (12) as /devices/platform/soc/1c2ac00.i2c/i2c-0/0-0038/input/input2
The communication to the touchscreen controller is only possible when power is supplied the the 4G modem using the hardware switch (the light should be off) since the power rail for the 4G modem is used as pull-up for the i2c bus that connects the touchscreen controller to the sopine.
The camera's both share the csi-0 bus. Currently only the rear camera (OV5640) has mainline linux support.
The device tree nodes for the camera and the csi bus have been added but so far it hasn't been possible to get images from the camera. It's not clear if this an issue with the OV5640 driver or with the allwinner csi kernel module.
It's also not clear how the camera switching would work on the software side, maybe Linux just deals with it if the device tree nodes for the second camera get added.
dontbeevil:~# ffmpeg -f v4l2 -i /dev/video0 [video4linux2,v4l2 @ 0xffff8042d140] ioctl(VIDIOC_G_PARM): Not a tty [video4linux2,v4l2 @ 0xffff8042d140] Time per frame unknown [video4linux2,v4l2 @ 0xffff8042d140] ioctl(VIDIOC_STREAMON): Invalid argument /dev/video0: Invalid argument
The issue is probably that the camera isn't recognized by the driver
[ 20.697863] ov5640 2-003c: Linked as a consumer to regulator.8 [ 20.697930] ov5640 2-003c: Linked as a consumer to regulator.17 [ 20.697984] ov5640 2-003c: Linked as a consumer to regulator.13 [ 20.726501] ov5640 2-003c: ov5640_check_chip_id: wrong chip identifier, expected 0x5640, got 0x0 [ 20.726703] ov5640 2-003c: Dropping the link to regulator.8 [ 20.726772] ov5640 2-003c: Dropping the link to regulator.17 [ 20.726818] ov5640 2-003c: Dropping the link to regulator.13
The wifi interface shows up and is able to scan for networks, it doesn't want to actually connect
[ 1962.148414] RTL8723BS: set bssid:e4:8d:8c:49:44:e2 [ 1962.148468] RTL8723BS: set ssid [BrixIT] fw_state = 0x00000088 [ 1962.159624] RTL8723BS: start auth [ 1962.161053] RTL8723BS: auth success, start assoc [ 1972.568295] RTL8723BS: set bssid:00:00:00:00:00:00 [ 1985.233165] RTL8723BS: set bssid:e4:8d:8c:49:44:e2 [ 1985.233228] RTL8723BS: set ssid [BrixIT] fw_state = 0x00000088 [ 1985.304510] RTL8723BS: start auth [ 1985.306184] RTL8723BS: auth success, start assoc [ 1995.657207] RTL8723BS: set bssid:00:00:00:00:00:00 [ 2006.185941] RTL8723BS: set bssid:e4:8d:8c:49:44:e2 [ 2006.185994] RTL8723BS: set ssid [BrixIT] fw_state = 0x00000088 [ 2006.193313] RTL8723BS: start auth [ 2006.194839] RTL8723BS: auth success, start assoc [ 2010.497801] RTL8723BS: set bssid:00:00:00:00:00:00
Haven't looked at yet
All sensors work and are accessible as Industrial I/O (iio) devices in linux. To query the data you can cat the files in `/sys/bus/iio/` to make the kernel module query the sensors
dontbeevil:~# cat /sys/bus/iio/devices/iio\:device0/name mpu6050 dontbeevil:~# cat /sys/bus/iio/devices/iio\:device0/in_accel_z_raw -17594 dontbeevil:~# echo "Verified gravity still exists"
stk3310 module doesn't load automatically, after running
modprobe stk3310 the third device in `/sys/bus/iio` should show up.
All the required gpio-led stuff is added to the device tree but the gpio-led module doesn't do anything. No led devices, No errors, No log messages.
The A64 SoC audio should work on mainline with the
CONFIG_SUN50I_CODEC_ANALOG option enabled. The required nodes are in the device tree. The kernel module won't load correctly and the mixer won't show up:
dontbeevil:~# alsamixer cannot open mixer: No such file or directory
USB gets loaded but doesn't work yet:
[ 0.186880] usbcore: registered new device driver usb [ 0.535763] sun4i-usb-phy 1c19400.phy: failed to get clock usb0_phy [ 0.572453] ehci-platform 1c1a000.usb: EHCI Host Controller [ 0.572477] ehci-platform 1c1a000.usb: new USB bus registered, assigned bus number 1 [ 0.573646] ehci-platform 1c1a000.usb: irq 17, io mem 0x01c1a000 [ 0.586033] ehci-platform 1c1a000.usb: USB 2.0 started, EHCI 1.00 [ 0.587976] ohci-platform 1c1a400.usb: Generic Platform OHCI controller [ 0.587997] ohci-platform 1c1a400.usb: new USB bus registered, assigned bus number 2 [ 0.588307] ohci-platform 1c1a400.usb: irq 18, io mem 0x01c1a400
pine-dontbeevil:~# lsusb Bus 005 Device 001: ID 1d6b:0002 Bus 003 Device 001: ID 1d6b:0002 Bus 001 Device 001: ID 1d6b:0002 Bus 004 Device 001: ID 1d6b:0001 Bus 002 Device 001: ID 1d6b:0001
- pmaports!308 Initial merge request