#linuxcnc-devel Logs

Jul 12 2018

08:48 AM skunkworks: cradek: recycling cast iron for a small mill riser block
11:03 AM jepler: andypugh: support for LXRT and Xenomai in uspace are totally experimetal, and not enabled in builds from linuxcnc.org. If a user tries them, they get to keep the parts.
11:04 AM jepler: pretty sure that you have to specifically opt in by specfiying --with-realtime=uspace if a RTAI kernel is detected; if an RTAI kernel isn't detected, then you wouldn't get LXRT.
11:04 AM jepler: maybe I got the detection wrong, who knows.
11:05 AM jepler: anyway, it could be that the uspace LXRT RTAI doesn't work well (I think I had some evidence that it didn't get thread priorities right, so that the servo thread would interrupt the base thread, which would be awful and unusable!) but I couldn't get a stable kernel for testing on any of my machines so I just put it aside.
11:05 AM jepler: and if you told me that was 3 years ago now I'd believe you
02:34 PM linuxcnc-build: build #3675 of 1405.rip-wheezy-armhf is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1405.rip-wheezy-armhf/builds/3675 blamelist: Andy Pugh <andypugh@localhost.localdomain>
03:03 PM linuxcnc-build: build #5523 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/5523 blamelist: Andy Pugh <andypugh@localhost.localdomain>
03:32 PM seb_kuzminsky: i'm going to shut those armhf buildslaves off, this is ridiculous
04:47 PM pcw_mesa: any idea why the hm2_rpspi driver would complain that the platform is unsupported? (running on a RPI3)
04:52 PM micges: pcw_mesa: looking at code, it seems like some new rpi3 revision with new other cpu
04:54 PM andypugh: Hmm, the Forum isn’t even mentioned in the “Getting Help” section of the docs: http://linuxcnc.org/docs/2.7/html/getting-started/about-linuxcnc.html
04:54 PM rmu: pcw_mesa: what does the line "Hardware" in /proc/cpuinfo say?
04:55 PM rmu: the raspberry pi 3 b+ has a "BCM2837"
04:56 PM rmu: the pi 3 b (without plus) a "BCM2835"
04:56 PM pcw_mesa: no such line
04:56 PM pcw_mesa: it says "generic DT based system"
04:57 PM hazzy-lab: andypugh: The IRC link is also dead ...
04:57 PM andypugh: https://webchat.freenode.net
04:58 PM hazzy-lab: This is where it takes me: http://www.linuxcnc.org/index.php/english/community
04:59 PM rmu: pcw_mesa: "grep device_tree /boot/config.txt" should be something like device_tree=dtbs_rt2/bcm2710-rpi-3-b.dtb, but i'm no expert on that matter
05:00 PM rmu: generic seems odd for a rpi
05:03 PM hazzy-lab: andypugh: Would it be helpful if I fixed things like the bad IRC link, or does too many cooks spoil the pudding?
05:04 PM hazzy-lab: Its bad in both 2.7 and devel
05:04 PM pcw_mesa: yeah nothing about processor type in config.txt
05:04 PM andypugh: Fix what you want.. But I am just about to commit a fix for that doc. I am just dithering about Master or 2.7
05:04 PM hazzy-lab: Ok, that is what I was asking, if you already have it I won't complicate things :)
05:05 PM rmu: pcw_mesa: there should be a file bcm2710-rpi-3-b.dtb as part of a kernel image deb file somewhere in /boot or /boot/dtbs_rt, perhaps my kernel debs got that wrong
05:07 PM rmu: forget what i said about pi 3 b and BCM2835, i was confused
05:08 PM pcw_mesa: looks like the are various processor specific dtb files in boot but not mentioned by config.txt
05:08 PM hazzy-lab: Thanks for adding the forum section andypugh
05:09 PM rmu: booting with the wrong dt file works for me somewhat, but produces strange power and stability issues, so wouldn't recommend that ;)
05:09 PM rmu: in theory those are built as part of the kernel, but it seems they are mostly identical (between different kernel versions)
05:10 PM hazzy-lab: andypugh: might that be worth putting in 2.6 too? I guess so few people use it its not ..
05:11 PM andypugh: ANyone using 2.6 probably knows all that they need to know by now :-)
05:12 PM hazzy-lab: good point :D
05:18 PM pcw_mesa: rmu: coul you post your config.txt somewhere, even with config_tree=bcm2710-rpi-3-b.dtb I stlll get generic DT
05:19 PM rmu: device_tree
05:19 PM rmu: not config_tree
05:19 PM rmu: and it is a path relative to /boot AFAIK
05:19 PM rmu: (has to be on /boot partition)
05:19 PM rmu: mom
05:26 PM pcw_mesa: yeah we had device_tree but no path, does it need a path?
05:26 PM rmu: not if the file resides in /boot
05:27 PM rmu: i have it in dtbs_rt2 for some reason therefore dtbs_rt2/bcm...
05:31 PM rmu: https://www.raspberrypi.org/documentation/configuration/device-tree.md suggests checking for warnings or errors from the loader with sudo vcdbg log msg
05:33 PM rmu: dtc -I fs /proc/device-tree
05:34 PM rmu: should output currently active DT in human readable form
05:37 PM pcw_mesa: ok so replacing the RPIs default bcm2710-rpi-3-b.dtb with yours fixed things, no idea why
05:39 PM rmu: strange
05:40 PM rmu: that should also fix the detection issue
05:42 PM rmu: it's past midnight here, i'm off till morning, ttyl
06:19 PM pcw_mesa: OK progress, it finds the 7C80 card now
06:22 PM skunkworks: you know this solution would rival the acorn... with the acorn it is running some version of BBB - plus you still need a computer.
06:23 PM skunkworks: beable bone green
06:24 PM skunkworks: and a lot more i/o - expantion...
06:26 PM pcw_mesa: Yeah the graphics seem OK if you dont try to do too much
06:28 PM pcw_mesa: ( Youtube videos are ok but more than a couple glxgears bring it to its knees )
06:28 PM skunkworks: all it needs is to be able to do a preview...
06:29 PM pcw_mesa: Yeah complex previews are slow so I hear
06:30 PM pcw_mesa: all in all it seems about like the old atom D525s speed wise
06:34 PM pcw_mesa: hmm latency got better with a heatsink...
06:35 PM hazzy-lab: pcw_mesa: if it's as good as the D525s that's great!
06:46 PM skunkworks: neat
07:37 PM aircooled is now known as hazzy-lab