#linuxcnc-devel Logs
Apr 18 2017
#linuxcnc-devel Calendar
03:12 AM KGB-linuxcnc: 03Chris Morley 05master 69497a9 06linuxcnc 10lib/python/hal_glib.py hal_glib -change some function names -add one * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=69497a9
03:13 AM KGB-linuxcnc: 03Chris Morley 05master ffe4ecc 06linuxcnc 10docs/src/Submakefile 03docs/src/gui/GStat.txt 10docs/src/index.tmpl docs -add info about GStat python module * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=ffe4ecc
03:22 AM KGB-linuxcnc: 03chris morley 05aries_master c6180f3 06linuxcnc 03share/gscreen/skins/aries/aries.glade 03share/gscreen/skins/aries/aries_handler.py GUI -add aries screen code * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=c6180f3
03:22 AM KGB-linuxcnc: 03chris morley 05aries_master 33e4af2 06linuxcnc 10(7 files) config -add a sample config for aries screen * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=33e4af2
03:22 AM KGB-linuxcnc: 03chris morley 05aries_master 73f140b 06linuxcnc 10(135 files in 20 dirs) aries -add theme files * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=73f140b
03:22 AM KGB-linuxcnc: 03Chris Morley 05aries_master b43002b 06linuxcnc 10(8 files in 3 dirs) aries -ongoing updates * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b43002b
03:25 AM linuxcnc-build: build #4889 of 1300.rip-precise-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1300.rip-precise-i386/builds/4889 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
03:26 AM linuxcnc-build: build #4892 of 1306.rip-precise-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1306.rip-precise-amd64/builds/4892 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
03:27 AM linuxcnc-build: build #3251 of 1404.rip-wheezy-rtpreempt-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1404.rip-wheezy-rtpreempt-amd64/builds/3251 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
03:27 AM linuxcnc-build: build #2566 of 1402.rip-wheezy-rtpreempt-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1402.rip-wheezy-rtpreempt-i386/builds/2566 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
03:28 AM linuxcnc-build: build #4100 of 1301.rip-precise-rtai-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1301.rip-precise-rtai-i386/builds/4100 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
03:28 AM linuxcnc-build: build #3050 of 1400.rip-wheezy-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1400.rip-wheezy-i386/builds/3050 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
03:30 AM linuxcnc-build: build #1518 of 1520.rip-jessie-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1520.rip-jessie-amd64/builds/1518 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
03:30 AM linuxcnc-build: build #3051 of 1403.rip-wheezy-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1403.rip-wheezy-amd64/builds/3051 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
03:30 AM linuxcnc-build: build #2717 of 1401.rip-wheezy-rtai-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1401.rip-wheezy-rtai-i386/builds/2717 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
03:30 AM linuxcnc-build: build #1517 of 1500.rip-jessie-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1500.rip-jessie-i386/builds/1517 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
03:31 AM linuxcnc-build: build #1517 of 1510.rip-jessie-rtpreempt-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1510.rip-jessie-rtpreempt-i386/builds/1517 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
03:31 AM linuxcnc-build: build #1519 of 1530.rip-jessie-rtpreempt-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1530.rip-jessie-rtpreempt-amd64/builds/1519 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
03:42 AM KGB-linuxcnc: 03Chris Morley 05master 92a7aa7 06linuxcnc 10docs/src/index.tmpl docs -fix link * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=92a7aa7
03:49 AM linuxcnc-build: build #3059 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/3059 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
03:49 AM linuxcnc-build: build #4907 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/4907 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
04:01 AM linuxcnc-build: build #4893 of 1306.rip-precise-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1306.rip-precise-amd64/builds/4893 blamelist: dummy, Joe Hildreth <joeh@threerivershospital.com>, andypugh <andy@bodgesoc.org>, Norbert Schechner <nieson@web.de>, Chris Radek <chris@timeguy.com>, chris morley
04:01 AM linuxcnc-build: <chrisinnanaimo@hotmail.com>, Phillip Carter <phillcarter54@gmail.com>, Jeff Epler <jepler@unpythonic.net>, joseph calderon <calderon.joe@gmail.com>, Alexander Brock <alexander@lunar-orbit.de>, Sebastian Kuzminsky <seb@highlab.com>, John Thornton <bjt128@gmail.com>, Jon Elson <elson@pico-systems.com>, Chris Morley <chrisinnanaimo@hotmail.com>, Dewey
04:01 AM linuxcnc-build: Garrett <dgarrett@panix.com>
04:02 AM linuxcnc-build: build #4890 of 1300.rip-precise-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1300.rip-precise-i386/builds/4890 blamelist: dummy, Joe Hildreth <joeh@threerivershospital.com>, andypugh <andy@bodgesoc.org>, Norbert Schechner <nieson@web.de>, Chris Radek <chris@timeguy.com>, chris morley
04:02 AM linuxcnc-build: <chrisinnanaimo@hotmail.com>, Phillip Carter <phillcarter54@gmail.com>, Jeff Epler <jepler@unpythonic.net>, joseph calderon <calderon.joe@gmail.com>, Alexander Brock <alexander@lunar-orbit.de>, Sebastian Kuzminsky <seb@highlab.com>, John Thornton <bjt128@gmail.com>, Jon Elson <elson@pico-systems.com>, Chris Morley <chrisinnanaimo@hotmail.com>, Dewey
04:02 AM linuxcnc-build: Garrett <dgarrett@panix.com>
04:03 AM linuxcnc-build: build #4101 of 1301.rip-precise-rtai-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1301.rip-precise-rtai-i386/builds/4101 blamelist: dummy, Joe Hildreth <joeh@threerivershospital.com>, andypugh <andy@bodgesoc.org>, Norbert Schechner <nieson@web.de>, Chris Radek <chris@timeguy.com>,
04:03 AM linuxcnc-build: chris morley <chrisinnanaimo@hotmail.com>, Phillip Carter <phillcarter54@gmail.com>, Jeff Epler <jepler@unpythonic.net>, joseph calderon <calderon.joe@gmail.com>, Alexander Brock <alexander@lunar-orbit.de>, Sebastian Kuzminsky <seb@highlab.com>, John Thornton <bjt128@gmail.com>, Jon Elson <elson@pico-systems.com>, Chris Morley
04:03 AM linuxcnc-build: <chrisinnanaimo@hotmail.com>, Dewey Garrett <dgarrett@panix.com>
04:05 AM linuxcnc-build: build #3051 of 1400.rip-wheezy-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1400.rip-wheezy-i386/builds/3051 blamelist: dummy, Joe Hildreth <joeh@threerivershospital.com>, andypugh <andy@bodgesoc.org>, Norbert Schechner <nieson@web.de>, Chris Radek <chris@timeguy.com>, chris morley
04:05 AM linuxcnc-build: <chrisinnanaimo@hotmail.com>, Phillip Carter <phillcarter54@gmail.com>, Jeff Epler <jepler@unpythonic.net>, joseph calderon <calderon.joe@gmail.com>, Alexander Brock <alexander@lunar-orbit.de>, Sebastian Kuzminsky <seb@highlab.com>, John Thornton <bjt128@gmail.com>, Jon Elson <elson@pico-systems.com>, Chris Morley <chrisinnanaimo@hotmail.com>, Dewey
04:05 AM linuxcnc-build: Garrett <dgarrett@panix.com>
04:06 AM linuxcnc-build: build #2567 of 1402.rip-wheezy-rtpreempt-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1402.rip-wheezy-rtpreempt-i386/builds/2567 blamelist: dummy, Joe Hildreth <joeh@threerivershospital.com>, andypugh <andy@bodgesoc.org>, Norbert Schechner <nieson@web.de>, Chris Radek
04:06 AM linuxcnc-build: <chris@timeguy.com>, chris morley <chrisinnanaimo@hotmail.com>, Phillip Carter <phillcarter54@gmail.com>, Jeff Epler <jepler@unpythonic.net>, joseph calderon <calderon.joe@gmail.com>, Alexander Brock <alexander@lunar-orbit.de>, Sebastian Kuzminsky <seb@highlab.com>, John Thornton <bjt128@gmail.com>, Jon Elson <elson@pico-systems.com>, Chris Morley
04:06 AM linuxcnc-build: <chrisinnanaimo@hotmail.com>, Dewey Garrett <dgarrett@panix.com>
04:06 AM linuxcnc-build: build #1518 of 1500.rip-jessie-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1500.rip-jessie-i386/builds/1518 blamelist: dummy, Joe Hildreth <joeh@threerivershospital.com>, andypugh <andy@bodgesoc.org>, Norbert Schechner <nieson@web.de>, Chris Radek <chris@timeguy.com>, chris morley
04:06 AM linuxcnc-build: <chrisinnanaimo@hotmail.com>, Phillip Carter <phillcarter54@gmail.com>, Jeff Epler <jepler@unpythonic.net>, joseph calderon <calderon.joe@gmail.com>, Alexander Brock <alexander@lunar-orbit.de>, Sebastian Kuzminsky <seb@highlab.com>, John Thornton <bjt128@gmail.com>, Jon Elson <elson@pico-systems.com>, Chris Morley <chrisinnanaimo@hotmail.com>, Dewey
04:06 AM linuxcnc-build: Garrett <dgarrett@panix.com>
04:07 AM linuxcnc-build: build #3052 of 1403.rip-wheezy-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1403.rip-wheezy-amd64/builds/3052 blamelist: dummy, Joe Hildreth <joeh@threerivershospital.com>, andypugh <andy@bodgesoc.org>, Norbert Schechner <nieson@web.de>, Chris Radek <chris@timeguy.com>, chris morley
04:07 AM linuxcnc-build: <chrisinnanaimo@hotmail.com>, Phillip Carter <phillcarter54@gmail.com>, Jeff Epler <jepler@unpythonic.net>, joseph calderon <calderon.joe@gmail.com>, Alexander Brock <alexander@lunar-orbit.de>, Sebastian Kuzminsky <seb@highlab.com>, John Thornton <bjt128@gmail.com>, Jon Elson <elson@pico-systems.com>, Chris Morley <chrisinnanaimo@hotmail.com>, Dewey
04:07 AM linuxcnc-build: Garrett <dgarrett@panix.com>
04:07 AM linuxcnc-build: build #1518 of 1510.rip-jessie-rtpreempt-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1510.rip-jessie-rtpreempt-i386/builds/1518 blamelist: dummy, Joe Hildreth <joeh@threerivershospital.com>, andypugh <andy@bodgesoc.org>, Norbert Schechner <nieson@web.de>, Chris Radek
04:07 AM linuxcnc-build: <chris@timeguy.com>, chris morley <chrisinnanaimo@hotmail.com>, Phillip Carter <phillcarter54@gmail.com>, Jeff Epler <jepler@unpythonic.net>, joseph calderon <calderon.joe@gmail.com>, Alexander Brock <alexander@lunar-orbit.de>, Sebastian Kuzminsky <seb@highlab.com>, John Thornton <bjt128@gmail.com>, Jon Elson <elson@pico-systems.com>, Chris Morley
04:07 AM linuxcnc-build: <chrisinnanaimo@hotmail.com>, Dewey Garrett <dgarrett@panix.com>
04:07 AM linuxcnc-build: build #1519 of 1520.rip-jessie-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1520.rip-jessie-amd64/builds/1519 blamelist: dummy, Joe Hildreth <joeh@threerivershospital.com>, andypugh <andy@bodgesoc.org>, Norbert Schechner <nieson@web.de>, Chris Radek <chris@timeguy.com>, chris morley
04:07 AM linuxcnc-build: <chrisinnanaimo@hotmail.com>, Phillip Carter <phillcarter54@gmail.com>, Jeff Epler <jepler@unpythonic.net>, joseph calderon <calderon.joe@gmail.com>, Alexander Brock <alexander@lunar-orbit.de>, Sebastian Kuzminsky <seb@highlab.com>, John Thornton <bjt128@gmail.com>, Jon Elson <elson@pico-systems.com>, Chris Morley <chrisinnanaimo@hotmail.com>, Dewey
04:07 AM linuxcnc-build: Garrett <dgarrett@panix.com>
04:07 AM linuxcnc-build: build #3252 of 1404.rip-wheezy-rtpreempt-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1404.rip-wheezy-rtpreempt-amd64/builds/3252 blamelist: dummy, Joe Hildreth <joeh@threerivershospital.com>, andypugh <andy@bodgesoc.org>, Norbert Schechner <nieson@web.de>, Chris Radek
04:07 AM linuxcnc-build: <chris@timeguy.com>, chris morley <chrisinnanaimo@hotmail.com>, Phillip Carter <phillcarter54@gmail.com>, Jeff Epler <jepler@unpythonic.net>, joseph calderon <calderon.joe@gmail.com>, Alexander Brock <alexander@lunar-orbit.de>, Sebastian Kuzminsky <seb@highlab.com>, John Thornton <bjt128@gmail.com>, Jon Elson <elson@pico-systems.com>, Chris Morley
04:07 AM linuxcnc-build: <chrisinnanaimo@hotmail.com>, Dewey Garrett <dgarrett@panix.com>
04:07 AM linuxcnc-build: build #2718 of 1401.rip-wheezy-rtai-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1401.rip-wheezy-rtai-i386/builds/2718 blamelist: dummy, Joe Hildreth <joeh@threerivershospital.com>, andypugh <andy@bodgesoc.org>, Norbert Schechner <nieson@web.de>, Chris Radek <chris@timeguy.com>, chris
04:07 AM linuxcnc-build: morley <chrisinnanaimo@hotmail.com>, Phillip Carter <phillcarter54@gmail.com>, Jeff Epler <jepler@unpythonic.net>, joseph calderon <calderon.joe@gmail.com>, Alexander Brock <alexander@lunar-orbit.de>, Sebastian Kuzminsky <seb@highlab.com>, John Thornton <bjt128@gmail.com>, Jon Elson <elson@pico-systems.com>, Chris Morley <chrisinnanaimo@hotmail.com>,
04:07 AM linuxcnc-build: Dewey Garrett <dgarrett@panix.com>
04:08 AM linuxcnc-build: build #1520 of 1530.rip-jessie-rtpreempt-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1530.rip-jessie-rtpreempt-amd64/builds/1520 blamelist: dummy, Joe Hildreth <joeh@threerivershospital.com>, andypugh <andy@bodgesoc.org>, Norbert Schechner <nieson@web.de>, Chris Radek
04:08 AM linuxcnc-build: <chris@timeguy.com>, chris morley <chrisinnanaimo@hotmail.com>, Phillip Carter <phillcarter54@gmail.com>, Jeff Epler <jepler@unpythonic.net>, joseph calderon <calderon.joe@gmail.com>, Alexander Brock <alexander@lunar-orbit.de>, Sebastian Kuzminsky <seb@highlab.com>, John Thornton <bjt128@gmail.com>, Jon Elson <elson@pico-systems.com>, Chris Morley
04:08 AM linuxcnc-build: <chrisinnanaimo@hotmail.com>, Dewey Garrett <dgarrett@panix.com>
04:27 AM linuxcnc-build: build #3060 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/3060 blamelist: dummy, Joe Hildreth <joeh@threerivershospital.com>, andypugh <andy@bodgesoc.org>, Norbert Schechner <nieson@web.de>, Chris Radek <chris@timeguy.com>, chris morley
04:27 AM linuxcnc-build: <chrisinnanaimo@hotmail.com>, Phillip Carter <phillcarter54@gmail.com>, Jeff Epler <jepler@unpythonic.net>, joseph calderon <calderon.joe@gmail.com>, Alexander Brock <alexander@lunar-orbit.de>, Sebastian Kuzminsky <seb@highlab.com>, John Thornton <bjt128@gmail.com>, Jon Elson <elson@pico-systems.com>, Chris Morley <chrisinnanaimo@hotmail.com>, Dewey
04:27 AM linuxcnc-build: Garrett <dgarrett@panix.com>
04:27 AM linuxcnc-build: build #4908 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/4908 blamelist: dummy, Joe Hildreth <joeh@threerivershospital.com>, andypugh <andy@bodgesoc.org>, Norbert Schechner <nieson@web.de>, Chris Radek <chris@timeguy.com>, chris morley
04:27 AM linuxcnc-build: <chrisinnanaimo@hotmail.com>, Phillip Carter <phillcarter54@gmail.com>, Jeff Epler <jepler@unpythonic.net>, joseph calderon <calderon.joe@gmail.com>, Alexander Brock <alexander@lunar-orbit.de>, Sebastian Kuzminsky <seb@highlab.com>, John Thornton <bjt128@gmail.com>, Jon Elson <elson@pico-systems.com>, Chris Morley <chrisinnanaimo@hotmail.com>, Dewey
04:27 AM linuxcnc-build: Garrett <dgarrett@panix.com>
05:08 AM jthornton: seb_kuzminsky: I was building master uspace
08:14 AM skunkworks: 50us over night
10:01 AM skunkworks: pcw_home, if you have some time - could you make a bit file for the 7i92 that has this pinout?
10:01 AM skunkworks: http://tom-itx.no-ip.biz:81/~webpage/cnc/skunkworks/PIN_7I76_skunktest.vhd
10:02 AM skunkworks: (just ab encoder counters from 2 to 9
10:02 AM skunkworks: )
10:29 AM pcw_home: http://freeby.mesanet.com/7i92_skunktest.bit
11:02 AM skunkworks: Oh - thank you very much!
11:04 AM seb_kuzminsky: jthornton: i get the crash too! wheezy rtai VM, master, configure --with-realtime=uspace, hard lockup halfway through the second runtests run
11:04 AM seb_kuzminsky: well i'll be a monkey's uncle
11:05 AM skunkworks: Jt isn't as crazy as we thought...
11:05 AM seb_kuzminsky: heh
11:07 AM seb_kuzminsky: looks like 3 out of the VM's 4 CPUs are 100% busy
11:12 AM seb_kuzminsky: bisecting from 2.7 to master is no fun, because a bunch of the ja commits in master dont build :-/
11:14 AM seb_kuzminsky: hmm, the realtime scriptstill loads the rtai modules, even though it's built for uspace
11:27 AM seb_kuzminsky: older versions of master don't load them, that seems suspicious
11:30 AM seb_kuzminsky: oh man, if loading and then immediately unloading the rtai modules makes the machine lock up... my days of not taking rtai seriously are certainly comming to a middle
12:21 PM JT-Shop: heh
01:59 PM jepler: in master branch, uspace includes rtai support
02:00 PM jepler: and yes it does need to load those modules in the rtai case
02:00 PM jepler: this went in circa july 2016
02:04 PM jepler: I thought I could successfully runtests more than once though
02:05 PM jepler: this would have been with vmlinuz-3.18.0-1-rtai-amd64
02:06 PM jepler: seb_kuzminsky, JT-Shop: ^^
02:07 PM seb_kuzminsky: yeah, it's not the merge of uspace-plus that introduced the problem
02:07 PM seb_kuzminsky: still bisecting here, it's slow
02:08 PM jepler: oh it's since then?
02:08 PM seb_kuzminsky: yeah
02:08 PM seb_kuzminsky: well, either that or my test missed it
02:11 PM jepler: are you running your bisect in a vm or on metal?
02:51 PM CaptHindsight: memleak said he's testing RTAI on the latest 3.16 from kernel.org this week and then he' start working on 4.1
02:53 PM CaptHindsight: seb_kuzminsky: memleak is wondering are you building RTAI math library with glibc or newlib?
02:56 PM jepler: we depend on rtai providing a working math library.
02:56 PM jepler: from rtai_math.ko
02:57 PM jepler: though for historic reasons we also supply inline versions of some math functions for 32-bit x86 platforms only (rtapi_math_i386.h)
02:58 PM jepler: tests/realtime-math/rtmath.comp is intended to check for the presence of the functions we depend on and which we declare in the header rtapi_math.h.
03:49 PM seb_kuzminsky: CaptHindsight, jepler : in the linux-image-3.4-9-rtai-686 we shipped on wheezy we used rtai's math, but in 4.1 i've been using newlib
03:50 PM jepler: seb_kuzminsky: thanks for the correction
03:50 PM jepler: what debian package ends up providing the math .ko ?
03:50 PM seb_kuzminsky: rtai-modules
03:51 PM seb_kuzminsky: https://github.com/SebKuzminsky/rtai/blob/vulcano-debs/debian/rules.in#L53
03:51 PM jepler: eek: /tmp/newlib/build/libm/libm.a
03:52 PM jepler: besides that it makes sense, I see the dependency on newlib-source
03:59 PM seb_kuzminsky: i'm helping on a project that needs coordinated motion of a couple of big stepper motors, and i'm thinking of using a beagle bone for the controller
04:00 PM seb_kuzminsky: running stepgens in the PRU
04:00 PM seb_kuzminsky: good thing cradek isn't here to talk me out of it
04:04 PM cradek: :-P
04:04 PM seb_kuzminsky: :-)
04:04 PM cradek: that sounds cool actually
04:04 PM cradek: did jmk port the PRU stepgen, or just IO?
04:04 PM seb_kuzminsky: just IO, i think
04:06 PM jepler: yes, just IO
04:07 PM jepler: something that could be done without building code and loading it into the PRU, which you don't seem to be able to do with just packages in Debian stable last time I looked.
04:07 PM seb_kuzminsky: i recently dusted off my old hm2-bb branch, it still builds, so i think i'll scale the ambition back and just to stepgen as the next step
04:08 PM seb_kuzminsky: jepler: yeah, building pru code using gcc needs tools i could only get from github, last time i checked
04:08 PM seb_kuzminsky: that part is not awesome
04:09 PM seb_kuzminsky: https://github.com/dinuxbg/gnupru
04:14 PM jepler: at least there's a debian source package for gcc-pru http://repos.rcn-ee.com/debian/pool/main/g/gcc-pru/
04:16 PM jepler: "in debian" or not at least we could build it and serve it from wlo
06:28 PM memfrob: Hi seb_kuzminsky, so you build LinuxCNC using newlib, which means you download the newlib source tarball and build it that way? Wouldn't you need to make a deb package for newlib if you want to properly build a deb package for RTAI?
06:28 PM memfrob: Otherwise you're packaging RTAI without dpkg handling the dependencies to build it (newlib)
06:30 PM memfrob: I'm assuming you haven't had much luck linking the math build with ld-linux-x86-64.so.2
06:31 PM seb_kuzminsky: memfrob: my rtai debs build-depend on newlib-source: https://github.com/SebKuzminsky/rtai/blob/vulcano-debs/debian/control.in
06:31 PM memfrob: (or in your case ld-linux.so.2)
06:31 PM memfrob: debian packages newlib?!
06:31 PM seb_kuzminsky: and they build newlib as part of building rtai: https://github.com/SebKuzminsky/rtai/blob/vulcano-debs/debian/rules.in#L68
06:32 PM seb_kuzminsky: it builds and links fine, and linuxcnc (master branch) builds against it, and the runtests all pass, including the realtime-math test i wrote
06:33 PM seb_kuzminsky: yep, it's in debian: https://packages.debian.org/newlib-source
06:34 PM memfrob: wait if it's in debian why do you need to build newlib in your RTAI build scripts?
06:35 PM memfrob: won't newlib's libm.a already be installed somewhere?
06:36 PM seb_kuzminsky: the newlib-source package just contains the source, not the compiled library
06:36 PM memfrob: ah
06:37 PM memfrob: did you go with newlib due to symbol inconsistencies / errors with glibc?
06:37 PM seb_kuzminsky: because (i think) newlib is intended to be an embedded math library, not an alternative to libc's math library, so debian ship it in source format so the user (like us) can build it in whatever way makes sense for us
06:38 PM seb_kuzminsky: i felt uncomfortable copying glibc's libm.a into kernel space. i preferred to build the math code from source for its intented environment.
06:38 PM seb_kuzminsky: i dont think i even tried to get glibc to work in rtai's kernel modules
06:39 PM seb_kuzminsky: because that would be crazy
06:39 PM seb_kuzminsky: and i'm not crazy
06:39 PM jepler: https://goo.gl/photos/rALLs9SqPus1axgx5
06:40 PM memfrob: I don't see the problem aside from that it doesn't build for me, but nothing from rtai.org does so that's expected from my end.
06:40 PM seb_kuzminsky: hmm,
06:40 PM memfrob: jepler, that's the new calibration tree.
06:40 PM seb_kuzminsky: rtai.org's code generally builds for me
06:41 PM seb_kuzminsky: jepler: what kernel is that?
06:41 PM jepler: seb_kuzminsky: vmlinuz-3.18.0-1-rtai-amd64
06:41 PM seb_kuzminsky: one of the ones i built? or one you built?
06:41 PM jepler: not sure
06:41 PM memfrob: RTAI 5 has a very fragile self tuning calibration mechanism that has drastically changed from since the beginning.
06:42 PM memfrob: ^ in regards to jepler's photo
06:43 PM memfrob: what makes using glibc crazy?
06:43 PM seb_kuzminsky: heh, i think i might not ever have run any rtai5 kernel on bare metal, it's all been VMs for me
06:43 PM jepler: seb_kuzminsky: looks like it's probably one from the internet because I only find the deb in /var/cache/apt/archives and not elsewhere on my local disk
06:43 PM jepler: let me rebuild rtapi in kernel model and see if it locks so readily....
06:44 PM seb_kuzminsky: dpkg -s?
06:44 PM seb_kuzminsky: memfrob: my understanding is that userspace code and kernelspace code are built differently, and it's only a happy little accident if you can transplant it from one to another
06:45 PM memfrob: are you saying glibc is too big a library to be used in kernel space and newlib is more portable therefor more suitable?
06:45 PM seb_kuzminsky: that's not what i was saying, but that's maybe another good reason
06:46 PM memfrob: I'm trying to follow.
06:46 PM seb_kuzminsky: but probably you shouldn't listen to me, and if the glibc math library works in kernel space that's great
06:46 PM seb_kuzminsky: i believe there's no guarantee that the processor environment in userland and in the kernel are the same
06:46 PM seb_kuzminsky: maybe they are, but i don't think they're required to be
06:47 PM seb_kuzminsky: so building from source, specifically for the kernel, seemed safer to me
06:47 PM seb_kuzminsky: and it was easier to build newlib (rather than glibc)
06:47 PM memfrob: seb_kuzminsky, you're the only person I know who's able to build rtai.org vanilla with the new math rework, so why should you be the one I not listen to?
06:47 PM jepler: 256 iterations OK with kernel model realtime, same kernel
06:47 PM seb_kuzminsky: jepler: oh, was that crash uspace linuxcnc on rtai?
06:47 PM jepler: seb_kuzminsky: right
06:47 PM seb_kuzminsky: (you didn't tell us that part)
06:47 PM jepler: oh sorry
06:48 PM jepler: I didn't
06:48 PM seb_kuzminsky: my bisect is nearly done, and it looks like this:
06:48 PM seb_kuzminsky: bad: 9799eced1e
06:48 PM seb_kuzminsky: good: ce52ea1294
06:48 PM memfrob: I thought mcmodel=kernel makes sure that your concern doesn't happen?
06:48 PM seb_kuzminsky: i'm not super confident in my test's ability to trigger the crash, so the bad is probably reliable but the good might not be
06:49 PM seb_kuzminsky: memfrob: yes, i hope so!
06:49 PM seb_kuzminsky: memfrob: is glibc's math.a build with that too?
06:50 PM seb_kuzminsky: well, i'm off! see you folks later tonight here maybe
06:50 PM memfrob: cd glibc-git && git grep mcmodel :P
06:50 PM memfrob: take care!
06:54 PM jepler: seb_kuzminsky: that's an interesting finding from bisect
06:59 PM jepler: confirmed to the extent that I could run the test for 256 iterations at v2.7.7-1254-gce52ea1294
07:01 PM jepler: OK at v2.7.7-1255-gf059548899
07:03 PM jepler: and fail at v2.7.7-1256-g9799eced1e
07:03 PM jepler: so yeah your bisect is totally right
07:03 PM jepler: but .. how !?
07:05 PM CaptHindsight: seb_kuzminsky: I'm not sure we can help with RTAI after looking over the problems
07:06 PM CaptHindsight: memleak doesn't have the problems with RTAI since he just started fixing everything himself
07:08 PM CaptHindsight: memleak forked RTAI to keep it stable with LCNC
07:08 PM CaptHindsight: RTAI.org makes changes based to their tree on their agenda
07:10 PM CaptHindsight: on/for
07:22 PM jepler: the crash happens in WithRoot (or during the stretch of time WITH_ROOT) and the "triggering" commit causes a lot more of WITH_ROOT to happen
07:22 PM jepler: adding a printf + delay to WithRoot makes it trigger a lot less, too
07:23 PM jepler: but it still triggered before 256 runs were up
07:27 PM seb_kuzminsky: CaptHindsight: there weren't that many changes needed to linuxcnc to support rtai5
07:28 PM seb_kuzminsky: https://github.com/LinuxCNC/linuxcnc/commit/aa397543878d325a7117271a0637bf40c7646a2e
07:28 PM seb_kuzminsky: https://github.com/LinuxCNC/linuxcnc/commit/d6da65f3fd432b3fc105dcf58804f399ffc050dd
07:28 PM seb_kuzminsky: and maybe a dozen lines of build system stuff
07:30 PM jepler: and if I move 'loadrt threads' until after 'loadrt streamer' in and-or-not-mux.0/test.hal, it also runs to 256 iterations
07:31 PM jepler: sooo working hypothesis: race condition between setresuid() and RTAI LXRT, probably because setresuid() has to do SOMETHING to all the threads in the process, but it can't really do it to LXRT threads
07:31 PM jepler: in which case, what can we do in linuxcnc? probably not much :(
07:32 PM memfrob: The changes required aren't to be made in linuxcnc, you already have it compiling against it.
07:32 PM CaptHindsight: seb_kuzminsky: but, you still have a broken RTAI kernel
07:35 PM seb_kuzminsky: CaptHindsight: there's a race on module unload, true
07:35 PM seb_kuzminsky: i wonder if that's fixed in memleak's fork?
07:35 PM CaptHindsight: we went through the effort here to fix RTAI and have a working tool chain since we had to have LCNC working on our hardware
07:35 PM memfrob: well currently my tree is in the midst of many fixes and is broken right at this second.
07:36 PM CaptHindsight: we can't depend on RTAI.org or Ubuntu or Debian
07:36 PM jepler: I've had my fill of rtai for the afternoon. afk. seb_kuzminsky just reiterating that your bisect was right on, thanks
07:36 PM jepler: I'm sure based on this info somebody could write a standalone reproducer but then what
07:36 PM CaptHindsight: passes Advil to Jepler
07:37 PM memfrob: git reset --hard HEAD~4 will work :P
07:40 PM seb_kuzminsky: fwiw, the kernel that's giving jepler trouble is from the shabby repo from way back when, not rtai.org
07:41 PM memfrob: You can't just pair some kernel I worked on X many years ago with the latest stuff from rtai.org
07:41 PM memfrob: I can tell by the kernel output that he has calibration code not included in the tree I worked on.
07:42 PM seb_kuzminsky: jepler's using a newer kernel, the repro that i did was on a wheezy/linux-3.4/rtai-3.9 CM
07:42 PM seb_kuzminsky: *VM
07:42 PM seb_kuzminsky: so i guess the problem affects both kernels & rtai versions
07:42 PM memfrob: jepler's screenshot: https://goo.gl/photos/rALLs9SqPus1axgx5 is not stuff I've worked on.
07:42 PM seb_kuzminsky: i bet jepler was using an old 3.18 kernel and rtai-5.0-test1 modules i built a year or so ago from rtai.org
07:45 PM CaptHindsight: 3.18 is passed EOL anyway
10:10 PM Joe_Hildreth: anyone have a suggestion on where in the wiki to place my best wiring practice document?
10:16 PM seb_kuzminsky: hi Joe_Hildreth
10:16 PM Joe_Hildreth: Hi Seb.
10:17 PM seb_kuzminsky: the wiki is kind of a wild west, free-for-all kind of place, your guess is as good as mine there
10:17 PM seb_kuzminsky: in the long term, it might make sense to add a section to the Integrator's Manual with that kind of info?
10:18 PM Joe_Hildreth: I am thinking of adding my Best wiring practices document under the Misc.
10:18 PM Joe_Hildreth: I think it is good documentation, but talking to John, I don't know that he feels it belongs in the LCNC docs since it is a little out of scope.
10:19 PM seb_kuzminsky: well i can see that point too
10:20 PM Joe_Hildreth: Me too after thinking on it a while, but the other side of the coin is, as a newbie, without sound wiring practice, you may have issues with an install and setup that may lead them to think that the problem is the software and not their hardware wiring.
10:21 PM Joe_Hildreth: It is a bit of a gray area, and I suppose the line should be drawn somewhere.
10:23 PM Joe_Hildreth: There is an area in the wiki for tutorials, but they are links. I think for now, I will add it under misc. and then maybe if it belongs elsewhere it can be moved or later added to the docs if folks think it is worthy.
10:23 PM seb_kuzminsky: sounds good
11:06 PM Joe_Hildreth: Well it isn't very pretty and I will need to go in and clean it up some but it is there to see if you wish. http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Best_Wiring_Practices
11:07 PM seb_kuzminsky: looks good! the markup and links and things all look right
11:08 PM seb_kuzminsky: Joe_Hildreth: is this a doubled word? "Foil shielded wire has a thin metal aluminum foil that is ..."
11:09 PM Joe_Hildreth: I am not following? What do you mean double word?
11:11 PM seb_kuzminsky: oh, the wording there seems surprising, "metal aluminum foil". seems redundant
11:12 PM Joe_Hildreth: I can remove of them. Perhaps metal foil?
11:12 PM seb_kuzminsky: that sounds better to me
11:12 PM seb_kuzminsky: thanks
11:13 PM Joe_Hildreth: OK, I have fixed it. Gotta run, but if you find anything else, let me know and I will look in the IRC log or you can email me.
11:14 PM seb_kuzminsky: awesome!
11:14 PM seb_kuzminsky: thanks for taking the time to write that down for the next guy
11:15 PM Joe_Hildreth: Goodnight Seb. Hey, I am still a newbie myself and am trying to add things people like me can use to get off on the right foot. Thanks for the encouragment and support you have given me.
11:17 PM Joe_Hildreth: quit