Progress! I went back to -ti-r9, and I get pruss_uio!

TL;DR: pruss_uio works on -ti-r9, not on -ti-r15 or -ti-r16.

I re-read this thread carefully:

  https://groups.google.com/forum/#!msg/beagleboard/0bG2lt9Xy_U/2FgquA60IQAJ

Decided the only thing I hadn't tried was being on the kernel jalodi used. So, I installed -ti-r9, set uEnv.txt to:

$ cat /boot/uEnv.txt
uname_r=4.1.4-ti-r9
dtb=am335x-boneblack-emmc-overlay.dtb
cmdline=coherent_pool=1M quiet

##Example v4.1.x
#cape_disable=bone_capemgr.disable_partno=BB-BONELT-HDMI,BB-BONELT-HDMIN
#cape_enable=bone_capemgr.enable_partno=BB-ENABLE-PRU

uuid=69f859b1-c308-41c9-8ca7-a3d0042031e8

Booted, then:

root@beaglebone:/home/debian# echo BB-ENABLE-PRU > /sys/devices/platform/bone_capemgr/slots
root@beaglebone:/home/debian# dmesg | grep pru
[ 114.058915] pruss_uio 4a300000.pruss: pins are not configured from the driver

Then tried running my code. It was able to load the PRU, and execute the code, and light up the ws281x LEDs I have connected.

Side note: it promptly locked up because I had the current limit set too low on the power supply to support the LEDs, so voltage sagged. Increased the current limit, and it's working much better now.

YOu should be save up to: 4.1.5-ti-r10

as in 4.1.6-ti-r11 ti merged a bunch of remoteproc changes:

Here's the ti diff:

https://github.com/RobertCNelson/ti-linux-kernel/compare/cc8cb4e2deb0fdb26e1bd884abab3f32451f0f35...54cf3481dbece37ef7d4d825a4dbfccb146cc931

Regards,

Thanks, Robert. That thread mentioned stability issues in r9 that you said TI was working on. Do you know if they got addressed in r10? And how/where do I report the UIO issues since r11?

Thanks!

Thanks, Robert. That thread mentioned stability issues in r9 that you said TI was working on. Do you know if they got addressed in r10?

both r9/r10 contain the musb revert which had caused the rebooting in
earlier v4.1.x kernels.

And how/where do I report the UIO issues since r11?

I'd ping: Suman Anna <s-anna@ti.com> he's usually pretty busy, but
he's the one at ti working on remoteproc/etc..

Regards,

Thanks, Robert. That thread mentioned stability issues in r9 that you said TI was working on. Do you know if they got addressed in r10?

both r9/r10 contain the musb revert which had caused the rebooting in
earlier v4.1.x kernels.

Pity. If I'm not using USB, do you know if the rebooting still happens?

And how/where do I report the UIO issues since r11?

I'd ping: Suman Anna <s-anna@ti.com> he's usually pretty busy, but
he's the one at ti working on remoteproc/etc..

Thanks, done.

Pity. If I’m not using USB, do you know if the rebooting still happens?

The only way I’ve seen this problem manifest its self is only when USB power is not used. Our fix was ground VUSB to USB ground via a simple USB PCB connector. e.g. we folded the support mount tabs into the pins on each side, solder them, and then ran a mini usb to USB-A cable from the beaglebone, into the connector.

The only way I've seen this problem manifest its self is only when USB power is not used. Our fix was ground VUSB to USB ground via a simple USB PCB connector. e.g. we folded the support mount tabs into the pins on each side, solder them, and then ran a mini usb to USB-A cable from the beaglebone, into the connector.

A bit hacky, but until uio is back in the kernel, I suppose that will do. Thanks!

No it’s fixed in r9… (r8 and earlier have the issue)

Regards,

Oh really? In this post, the OP says he had a spontaneous reboot, and you say "it should be better, but TI's looking at it now." I took that to mean it was having problems.

https://groups.google.com/d/msg/beagleboard/0bG2lt9Xy_U/4ADanqm8IQAJ

Hopefully what they found is that it's okay. Anyway, I'll try -r10, and -r11 to verify that's when the problem appeared, and hopefully won't see any crashes.