Okay, I have switched to a 3.8 kernel and now it boots Debian squeeze fine. Has anyone figured out the magic to access PWM's outside of the kernel on the BeagleBone black? I can see that the PWM support is being built into my kernel but I don't see any way to access/enable the functionality. I suspect it's something to do with the device tree configuration but I'm new to devicetree and I'm finding it has a pretty steep learning curve.
same her with ARCH Linux ARM.
So there's a difference in the pwm driver (kernel module?) being used by the BBB distro and Arch or Debian ? I'm not really (no, really not at all) very smart when it comes to this. I'm lost when Google does't know the answer to my problems (or the question which suits my problems ;-) )
Anyway the BBB Angstrom Distro should be fine for testing. I tried ARCH since BBB Angstrom comes with so much stuff I don't want (Cloud9 etc etc)
Just in case somebody is curious: I followed these (very simple) instructions to set up ARCH Linux ARM on a SD Card for the Beaglebone Black.
Am Mittwoch, 8. Mai 2013 17:11:13 UTC+2 schrieb Markus:Hi,my BBB is running on Debian 7 (netinstall 3.8.8-bone14 #1 SMP Wed Apr 24 08:55:00 UTC 2013 armv7l GNU/Linux)and, following Micahel's instructions, I've got the same problems as Neils reported.I noticed that there's only one capemgr device (/sys/devices/bone_capemgr.9/).When I echo bone_pwm_PX_XX > /sys/devices/bone_capemgr.9/slots I end up with pwm_test folders that look like this:pwm_test_P9_14.15/powerpwm_test_P9_14.15/power/controlpwm_test_P9_14.15/power/runtime_active_timepwm_test_P9_14.15/power/autosuspend_delay_mspwm_test_P9_14.15/power/runtime_statuspwm_test_P9_14.15/power/runtime_suspended_timepwm_test_P9_14.15/modaliaspwm_test_P9_14.15/subsystempwm_test_P9_14.15/ueventAny ideas what's wrong on my system?Regards,Markus
I found the source of the problem. The Ubuntu image did not have the EHRPWM_TEST set as opposed to the Angstorm image. I enabled it and reinstalled the kernel using the script here: http://www.eewiki.net/display/linuxonarm/BeagleBone+Black#BeagleBoneBlack-LinuxBuildScript:
This brought up the period and duty files that were missing.
It should also be noted that when compiling the module, I ran into compilation error (caused by a missing "}"). Changes I made can be found here: http://pastebin.com/s1CeYguF
On Thursday, May 23, 2013 7:52:04 PM UTC-4, derek...@gmail.com wrote:I'm also running into the same problems. I'm running the 12.10 image from http://elinux.org/BeagleBoardUbuntu and I don't have any files named duty or period show up.
echo am33xx_pwm > /sys/devices/bone_capemgr.#/slots
On Saturday, June 1, 2013 9:38:29 PM UTC+2, Thomas Carter wrote:ready for a noob question? I have been trying to not ask you this for the last hour, or two, but just how exactly do I "add the am33xx_pwm module" to the bone_capemgr. Are you using a cape? Do I have to ? Ready for it to get really dumb? I cannot see anywhere on the board where it describes a cape manager number... I am lost, sorry.
You received this message because you are subscribed to a topic in the Google Groups "BeagleBoard" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/beagleboard/wjbOVE6ItNg/unsubscribe?hl=en.
To unsubscribe from this group and all its topics, send an email to beagleboard...@googlegroups.com.
Thanks a lot. I was able to get an LED to blink using Angstrom with these steps. Notably the Debian Squeeze install I made with the instructions here: http://eewiki.net/display/linuxonarm/BeagleBone+Black only partially worked. The link for the PWM showed up but there were no "duty" and "period" files in the folder.
I was going to investigate reverse engineering bonescript but it wasn't clear to me that PWM was supported there in the first place. I think ideally there ought to be an official C/C++ library to complement the bonescript API. When I first booted the board I really expected there to be a libbeaglebone or some such thing for me to link to. I see a number of such libraries cropped up for the last revision of beaglebone but they didn't seem really seem to be promoted by the beagleboard folks. I would really like to encourage them to make some effort to make C/C++ a first class citizen when it comes to IO. At the very least I think it's important that reverse engineering bonescript is not the best option for getting going in C/C++. A simple cheat sheet of instructions like Michael presented linked on the beaglebone.org website would be enough to feel like I was being supported.