Hey, cross posting from seeed studio forums.
tldr: clocking in data on spi0 looks cattywompus, spi1 ok
Was noticing an issue with some custom hardware/software that utilizes both spi channels, was working on beaglebone black prior. Wanted to make it easily reproducible so flashed a bbgw and used adafruit bbio library as an example.
steps:
1.) flash BBGW with latest production image 4.4.15-ti-r37 (other kernel versions and imaged used as well)
2.) prevent sound drivers from loading, as they conflict with spi1:
cd /lib/modules/4.4.15-ti-r37/kernel/sound/core
mv snd.ko snd.ko.bkp
3.) install adafruit library from source (need latest greatest, pip upgrade didn’t work)
cd ~
git clone GitHub - adafruit/adafruit-beaglebone-io-python: Adafruit's BeagleBone IO Python Library
cd adafruit-beaglebone-io-python
python setup.py install
cd …
4.) reboot
5.) use wires to loopback MOSI to MISO, both channels
short P9.18 and P9.21 with a piece of wire (SPI0)
short P9.29 and P9.30 with a piece of wire (SPI1)
6.) put the following into a file and run it
from Adafruit_BBIO.SPI import SPI
def run(x,y):print ‘running spi {} {}’.format(x,y)
spi = SPI(x,y)
spi.mode=2
spi.msh=4000000
bytes2send = range(20)
print ‘sending {}’.format(bytes2send)
print spi.xfer2(bytes2send)
spi.close()run(0,0)
run(1,0)
result:
sending [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19]
[128, 7, 28, 31, 112, 231, 124, 127, 192, 135, 156, 159, 240, 231, 252, 255, 0, 7, 28, 31]
running spi 1 0
sending [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19]
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19]
so, spi1 works in loopback. spi0…wtf? there are places where it looks like it’s double clocking MISO (like a 0x00000001 becomes a 0x00000011), some of it just looks random. Works fine on beaglebone black.
Thoughts?
If anyone has spi0 working on a bbgw would love to hear about the setup.
Thanks!
