#linuxcnc-devel | Logs for 2016-05-03

[07:24:08] <dgarr> skunkworks: dgarr/sync_home_final branch sync final move to joint homes: http://www.panix.com/~dgarrett/stuff/sync_home_final.jpg
[07:25:03] <skunkworks> dgarr!
[07:25:15] <skunkworks> Cool - Nice work!
[07:31:40] <skunkworks> cradek_, ^ would that be mergable?
[07:33:45] <jepler> dgarr/sync_home_final is relative to JA so it couldn't be merged without merging JA.
[07:33:59] <skunkworks> jepler, I am fine with that ;)
[07:34:22] <skunkworks> (no I have not been testing ja much yet. Stop yelling at me!)
[07:35:13] <skunkworks> jepler, your ethernet branch ran all night no issues.
[07:36:24] <jepler> skunkworks: that's good info, thanks
[07:37:28] <jepler> huh I guess there is no abs() in kernel space
[07:37:29] <jepler> linux-4.1.3$ find include/ | grep -w abs || echo no match
[07:37:29] <jepler> no match
[07:40:47] <skunkworks> what do you do then?
[07:49:47] <JT-Shop> PCW: thanks
[07:54:54] <skunkworks> cradek_, on jr - do you have machine zero at the upper right of the table or the lower left?
[08:03:25] <skunkworks> hmm - andy did a remap for lathe that used txxyy.. I wonder where that is.
[08:26:02] <cradek_> probably lower left?
[08:27:15] <cradek_> ooh sync_home_final is what I want synchronized (gantryish) homing to do in each step. I never got it working though...
[08:28:24] <jepler> one question that is not answered by the commit message is what happens with a configuration like 1,-1,-1
[08:30:24] <jepler> .. that commit message would be good fodder for the docs when someone writes them
[09:15:26] <cradek> jepler: I think the first block in the code makes them all negative in that case
[09:22:45] <jepler> that seems dandy
[10:06:50] <als> wondering about the patch for radiobutton widget and what the likely hood of someone adding a horizontal feature ?
[10:17:43] <cradek> als: I did see your patch on the mailing list, thank you for sharing it
[10:18:54] <cradek> I'm not sure I'm the right one to review a pyvcp change, but I have a couple requests. The signed-off-by is perfect, but the log message should be a summary of the change and why you made it, not the name of the file you changed. also there should be a corresponding doc change, I think maybe it should be in docs/src/gui/pyvcp.txt around line 500
[10:20:13] <cradek> also the email in the from and signed-off-by seems wrong - these need to be your correct email
[10:21:14] <cradek> with those changes I will be happy to push this to master for you
[10:22:31] <als> ok thanks
[10:23:34] <cradek> glad to have you back - I see you contributed some stuff in 2009-10, thank you
[10:24:33] <als> ya its been a while but linuxcnc is working great thank to all of you
[10:24:41] <cradek> terrific
[11:23:10] <skunkworks> So if I try to install the install linux-image-rt-amd64
[11:23:10] <skunkworks> from backports - I get
[11:23:25] <skunkworks> linux-image-4.4.0-0.bpo.1-rt-amd64 but it is not installable
[11:23:41] <seb_kuzminsky> jessie?
[11:23:57] <skunkworks_> yes
[11:24:29] <skunkworks_> it is like the amd64 image doesn't exist.
[11:25:17] <seb_kuzminsky> i also do not see it in backports
[11:25:41] <seb_kuzminsky> though the linux-image-rt-amd64 package is there
[11:25:54] <skunkworks> right - ok - I am not crazy
[11:26:04] <seb_kuzminsky> but that one doesn't contain the kernel, it just depends on the most recent actual kernel package
[11:27:48] <skunkworks> maybe I do have something else broken?
[11:28:15] <skunkworks> http://pastebin.com/zwq08cum
[11:33:25] <seb_kuzminsky> skunkworks: no, it's a bug in the backports repo
[11:33:42] <seb_kuzminsky> the dependency package is there, even though the real package it depends on is not
[11:34:22] <CaptHindsight> TomP said he has elephants and there is plenty of room for a Fest :)
[11:37:03] <seb_kuzminsky> skunkworks: i just updated my backports apt source and linux-image-rt-amd64 disappeared (maybe i got a different mirror this time)
[11:37:10] <seb_kuzminsky> CaptHindsight: i saw that!
[11:37:13] <seb_kuzminsky> sounds like fun :-)
[11:53:51] <skunkworks_> ah - so they removed it. darn
[12:11:38] <jepler> yeah that's the second time they've taken away the RT kernel from jessie-backports
[12:11:43] <jepler> jessie-backports now has kernel 4.5 in it
[12:39:12] <skunkworks_> Thought I would try a different rt kernal to see if I could get virtualbox working.
[12:39:15] <skunkworks_> Not a rush
[12:56:41] <pcw_home> Why not just build a RT kernel?
[12:58:42] <skunkworks> I am scared to.. ;)
[12:59:23] <pcw_home> Unlike RTAI, Preempt-RT is pretty easy
[13:01:28] <skunkworks> hmm - doesn't jt have directions somewhere?
[13:02:56] <pcw_home> I have a stupid script I use
[13:04:58] <pcw_home> But I am just directly installing it, I think JTs script may make a .deb
[13:05:21] <skunkworks> can I have your script?
[13:06:01] <pcw_home> sure, you may have to update the names
[13:07:55] <als> cradek, here is the patch with the docs changed hope its ok http://pastebin.com/ZG7qZQHp
[13:10:25] <pcw_home> http://freeby.mesanet.com/makert4.4.7-rt16 is for the latest available Preempt-RT kernel
[13:10:26] <pcw_home> also http://freeby.mesanet.com/makert3.1.8 (but I would update it to the latest 3.18 kernel/patchset)
[13:12:30] <pcw_home> I've had some issues with 4.4.X either stability, or worse latency than 3.18
[13:12:32] <pcw_home> _BUT_ 4.4.X has worked the best on newer hardware like my Zotax N3150 based machine
[13:12:36] <skunkworks> Thanks!
[13:12:49] <skunkworks> I will try the 4.4.x on this 2nd gen i7
[13:14:38] <cradek> als: great, I tweaked the log message (removed the filenames) and fixed a typo in the docs. you want to review or is it ok to just push it?
[13:15:06] <als> send it thanks
[13:16:02] <KGB-linuxcnc> 03Al Smart 05master 15be328 06linuxcnc 10docs/src/gui/pyvcp.txt 10lib/python/pyvcp_widgets.py allow horizontal radiobuttons in pyvcp * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=15be328
[13:16:14] <pcw_home> on the G3258:
[13:16:16] <pcw_home> http://freeby.mesanet.com/3.18.11-rt7.png
[13:16:17] <pcw_home> http://freeby.mesanet.com/4.4.7-rt16.png
[13:18:38] <skunkworks> looks like they have 4.5.2?
[13:20:23] <pcw_home> Im just going by whats in /pub/linux/kernel/projects/rt
[13:20:39] <skunkworks_> ah - ok
[13:29:07] <skunkworks> didn't you also have a kernel config?
[13:29:50] <pcw_home> yeah but you may still need to select 32/64 bits, funny hardware etc
[13:31:19] <pcw_home> most important thing is select fully preempt-able kernel
[13:31:20] <pcw_home> I think I had to enable Atheros wireless on my laptop
[13:32:33] <pcw_home> its easy to forget the config stuff since is uses the previous config file
[13:32:59] <pcw_home> oh and change the make to make -J4 or so
[13:33:59] <skunkworks_> heh
[13:36:39] <skunkworks_> This has an intel wireless in it
[13:36:55] <skunkworks_> and wired
[13:40:41] <pcw_home> on the N3150 I swiped a .config from OSADL
[13:45:08] <pcw_home> It seems to be better than mine but its hard to tell with the statistical nature of latency
[14:01:02] <skunkworks_> churning, churning....
[14:05:25] <skunkworks_> http://electronicsam.com/images/matsuura/MatsVFDside.JPG
[14:05:41] <skunkworks_> a bit of dressing up to do...
[14:10:45] <cradek> that's not much IO. I assume the tool changer doesn't work yet!
[14:17:12] <skunkworks_> no - not yet ;)
[17:07:54] <skunkworks> http://imgur.com/zYHqoWT
[17:25:48] <skunkworks> pcw_home, yes that was pretty painless... (well - with your directions ;)
[17:26:02] <skunkworks> and virtualbox just works
[17:38:13] <skunkworks> heh - the latency looks good until I try to use any of the hot keys like brightness, mute..
[17:38:19] <skunkworks> volume seems ok
[17:39:35] <skunkworks> keyboard light..
[17:39:40] <skunkworks> interesting..
[17:41:04] <skunkworks> that seems to be it
[17:41:21] <skunkworks> other than it doesn't like it if you force it to sleep ;)
[17:48:23] * JT-JA13 builds the dgarr/sync_home_final branch to test on ja
[17:49:06] <skunkworks> heh - it is mute. If I run the volume down using the volume keys - and it goes to mute - the latency spikes
[17:53:16] <skunkworks> Well - I think this will be very usable!
[18:00:46] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/trt_doc_test_v2 2b250fb 06linuxcnc 10docs/src/Submakefile 04docs/src/motion/5*txt 10docs/src/motion/5-axis-kinematics.txt docs/src/Submakefile cleanups forgot -a latexmath * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=2b250fb
[18:00:46] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/trt_doc_test_v2 e98a5e6 06linuxcnc 10docs/src/motion/5-axis-figures/Figure-10.png 10docs/src/motion/5-axis-figures/Figure-11.png 10docs/src/motion/5-axis-figures/Figure-9.png 5-axis-kinematics.txt new Figure-9,10,11 JA * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=e98a5e6
[18:00:46] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/trt_doc_test_v2 71bd41e 06linuxcnc 10docs/src/motion/5-axis-kinematics.txt 5-axis-kinematics.txt fix References items JA * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=71bd41e
[18:01:21] <skunkworks> I will have to hook up a 7i92 tomorrow.
[18:02:09] <skunkworks> (I also think it would be a cool video to walk up to the Matsuura - plug in a network cable to the laptop and run the machine...
[18:02:18] <skunkworks> I think it would open up peoples eyes..
[19:18:47] <PCW> That was my experience with my Del E6420, latency was fine unless you messed with display brightness, or power source (line/battery)
[19:21:08] <PCW> Don't remember if I messed with the volume buttons
[19:22:18] <PCW> I just made a USB power thief cable to run with 7I92
[19:51:28] <skunkworks> yes - unpluggin the charger spikes it too
[19:58:18] <PCW> wonder if that's SMI? (I would be reluctant to disable SMI on a laptop even if possible as it might melt)
[19:59:38] <PCW> I think the cyclictest has some tricks to get a better guess as to what caused a latency spike
[20:46:51] <skunkworks> yah - I think it is usable as is.. (just need to know not to do a few things.)
[20:51:53] <PCW> Yeah I ran the Dell for a day or 2 and it seemed fine
[20:55:21] <skunkworks> as jepler said - good enough for testing
[21:10:25] <skunkworks> runs a bit warm doing realtime.. but not uncomfortable.
[21:47:09] <jepler> http://wiki.osdev.org/APIC_timer#TSC-Deadline_mode
[22:18:28] <mozmck> jepler: I used deadline mode I believe in one config, but iirc it was hard to tell if it improved latency. I'll have to check - maybe I'm still using it.
[22:28:53] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/sync_home_final 022b811 06linuxcnc 10docs/src/config/ini-config.txt 10docs/src/config/ini-homing.txt 10docs/src/getting-started/updating-linuxcnc.txt docs homing: udate for negative HOME_SEQUENCEs JA * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=022b811
[22:28:53] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/sync_home_final 93ba31c 06linuxcnc 10src/emc/motion/homing.c homing.c support for re-home, minor cleanups JA * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=93ba31c
[22:28:53] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/sync_home_final c3cf776 06linuxcnc 10src/emc/usr_intf/axis/scripts/axis.py axis.py always joint mode for unhome defs JA * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=c3cf776