Back
[09:42:49] <CaptHindsight> jepler: all of memleaks RTAI work over the past few years was done on AMD
[09:43:57] <CaptHindsight> it was the only x86 silicon that allowed you to have your own firmware until last year
[09:45:48] <CaptHindsight> Intel has had a few binaries in the BIOS that are used for backdoor, phone home, remote control etc that operates completely out of band...
[09:46:33] <CaptHindsight> I wonder how it might effect real time when it's accessed
[09:50:11] <jepler> badly I'm sure
[10:54:33] <jepler> apparently debian stable (kernel & graphics) doesn't work too well with skylake, maybe I should go Broadwell instead...
[10:58:38] <pcw_home> Also supposedly you want a 4.4 or later kernel
[10:58:40] <pcw_home> Might be me but I have had trouble with the preempt-RT 4.4 kernels
[10:59:26] <jepler> Linux babs 4.4.0-trunk-rt-amd64 #1 SMP PREEMPT RT Debian 4.4.1-1~0.jepler8~1 (2016-02-16) x86_64 GNU/Linux
[10:59:30] <jepler> 10:33:47 up 6 days, 22:11, 34 users, load average: 0.48, 0.58, 0.58
[10:59:47] <jepler> it's OK on my laptop though I don't actually run linuxcnc much
[10:59:53] <jepler> but for desktop stuff it's stable
[11:00:01] <pcw_home> I either had terrible latency or crashes
[11:01:12] <pcw_home> (with 4.4.4-rt9+)
[11:05:42] <pcw_home> this is on my clunky old core-duo
[11:05:44] <pcw_home> currently running 4.1.15-rt18-rc1 which runs hm2-eth at 4 KHz reliably with
[11:05:45] <Tom_itx> which is broadwell?
[11:05:45] <pcw_home> youtube videos etc
[11:05:49] <Tom_itx> i5?
[11:06:57] <pcw_home> its a technology stepping not a specific CPU
[11:07:25] <Tom_itx> oh
[11:07:48] <pcw_home> its the 5th gen Intel Core series (Skylake is 6 Ivy bridge is 4)
[11:08:34] <Tom_itx> is skylake quite a bit better than i5?
[11:09:00] <Tom_itx> i don't quite understand cpu internals but it has more pipes
[11:09:06] <pcw_home> well there are skylake pentiums, i3s,I5s and I7s
[11:09:09] <Tom_itx> sounds bigger :)
[11:09:35] <Tom_itx> too confusin
[11:09:35] <Tom_itx> g
[11:09:51] <pcw_home> I think its mainly a bit better power consumption and faster memory (DDR4)
[11:11:31] <pcw_home> the current fastest desktop skylake I7 chip is not noticeably faster that the Ivy bridge version
[11:11:33] <pcw_home> but in general they are cooler
[11:12:08] <Tom_itx> is there just one fpga size selection on the 5i25?
[11:12:16] <pcw_home> yes
[11:12:27] <Tom_itx> i was building a bit file for micges yesterday..
[11:12:35] <Tom_itx> wanted to be sure it was ok
[11:13:27] <pcw_home> XC6SLX9blahblah
[11:14:56] <pcw_home> I should put that trick for safely testing a 5i25/7I92 bitfile in the manual
[11:18:35] * jepler hmms
[11:19:24] <jepler> this is write to the so-called "fallback" area, so that a subsequent reinitialization gets you the default area with good configuration?
[11:19:52] <jepler> I wonder if the hostmot2 driver should have a mode where it reboots into the second area at startup, and into the first area at exit
[11:20:22] <jepler> then the normal usage would be to leave the first config untouched
[11:20:38] <pcw_home> yes write to fallback and then reload from fallback
[11:21:56] <pcw_home> all current cheap cards (1 EEPROM, XC9SLX9) actually have room for 4 configs
[11:24:11] <seb_kuzminsky> http://www.phoronix.com/scan.php?page=news_item&px=UbuntuBSD-First-Release
[11:24:20] <jepler> seb_kuzminsky: I saw that
[11:24:23] <jepler> good luck to 'em, I say
[11:25:29] <seb_kuzminsky> yep
[11:25:36] <seb_kuzminsky> bbl
[13:09:35] <emcPT> Hello to all. Sometime ago I made some changes on linuxcnc code that I want to share. What is the best way to share my code?
[13:10:17] <emcPT> Code is simple and tested (over one year) and it is related with tool tip direction.
[13:11:54] <jepler> (I am only here for a moment)
[13:12:37] <jepler> emcPT: if you use github, you can create a fork of linuxcnc, then submit your changes as a pull request. a fork is also a way for interested users to access your modified version even before it is added to "official" linuxcnc.
[13:12:49] <jepler> http://linuxcnc.org/docs/2.7/html/code/contributing-to-linuxcnc.html#_linuxcnc_on_github (the whole document is good to read)
[13:12:53] <jepler> bye
[13:25:51] <emcPT> Well I must say that I try before to use git (about one year ago when the changes were done), but I was not able to. What I did: pulled and changed the source code under a directory named linuxcnc-dev. I also have sure that it builds.
[13:26:30] <emcPT> Should I now use git add filename1 filename2 ... ?
[13:31:07] <mozmck> emcPT: yes, git add will add the files, then you have to commit them.
[13:36:30] <emcPT> Ok, I will try. For now one question: I have made changes both on ioControl_v2.cc and ioControl.cc
[13:36:36] <emcPT> Both files are identical.
[13:37:05] <emcPT> Are both files used, or only the v2 version, that I suppose is newer ?
[13:41:11] <mozmck> I don't know the answer to that
[14:09:48] <emcPT> Ok, I think I have managed to make a commit. I got the following from the command line:
[14:10:16] <emcPT> [master 5dccb59] Added option to select the Fanuc standard for tool tip direction (lathe) User can add this option in the INI file with the line: FANUC_TOOLTIP_ORIENTATION = 1 in the EMCIO section
[14:10:17] <emcPT> .... 5 files changed, 26 insertions(+), 11 deletions(-)
[14:12:19] <emcPT> And now what happens?
[16:41:57] <mozmck> emcPT: Although jepler mentioned using github, it may be easier to send a patch to the mailing list. I think "git format-patch HEAD~" (without the quotes) will make a file with a patch of your last commit.
[16:51:03] <emcPT> Ok regarding the mailing list. Thank you for that info. If git hub was to be used, now that I made the commit, I must do anything else? This is something that is already online for revision or it is only on my PC?
[17:20:17] <jepler> it is only on your pc until you take extra action to share your changes with someone else.
[17:54:03] <emcPT> extra action = ?
[18:07:43] <emcPT> Using mozmck information I created a patch of my last commit. Should I email it to the emc-developers mailing list? or the procedure should be different?
[18:14:06] <emcPT> Sent to emc-developers@lists.sourceforge.net If I did something wrong let me know please.
[19:19:29] <mozmck> huh, does git format-patch output not indicate the sha the patch is against?
[22:18:05] <jepler> if(error_printed == 10)
[22:18:05] <jepler> rtapi_print_msg(
[22:18:05] <jepler> error_printed == 0 ? RTAPI_MSG_ERR : RTAPI_MSG_WARN,
[22:18:05] <jepler> "RTAPI: (further messages will be suppressed)\n");
[22:18:10] <jepler> is there a profoundness here I'm missing?
[22:18:23] <jepler> or is the inner test of error_printed == 0 entirely redundant?
[22:37:51] <jepler> I guess it's just copied from the cases above