#linuxcnc-devel Logs

Jul 09 2019

#linuxcnc-devel Calendar

04:25 PM -!- #linuxcnc-devel mode set to +v by ChanServ
05:41 PM memfrob: andypugh, might be better to talk on this :)
05:41 PM andypugh: ‘Appen
05:43 PM memfrob: so the no-calibration test works on your hardware?
05:45 PM andypugh: Maybe
05:45 PM andypugh: The latency test runs
05:46 PM andypugh: Ah! I forgot to patch LinuxCNC…
05:46 PM memfrob: As for the make menuconfig thing, you really shouldn't need it, even before that commit I posted on the mailing list about the linux dir fix if you're specifying it manually.
05:48 PM memfrob: and lastly, you said you had a different answer about the problem with detecting the absolute full kernel version string?
05:49 PM andypugh: No, not as such
05:50 PM memfrob: I'm just open to better ideas than sourcing the kernel config file for CONFIG_LOCALVERISON
05:51 PM memfrob: If only there was a way to tie into make deb-pkg params without using ~/.bash_history lol
05:52 PM andypugh: I seemed to be able to make the debian configure happy (as that needs a version and a flavour) by using rtai-amd64 in the kernel config.
05:52 PM memfrob: ah it needs a flavour too
05:53 PM andypugh: But I didn’t bother with tha LOCALVERSION= in tha deb-pkg command.
05:53 PM memfrob: then there'll be a remnant + sign in it..
05:53 PM andypugh: No… Linux rm-one 4.14.132-rtai-amd64 #1 SMP PREEMPT Mon Jul 8 23:47:38 BST 2019 x86_64 GNU/Linux
05:54 PM memfrob: I get a plus in it.
05:54 PM memfrob: I'm also using git sources not tarballs and i think that might be where its coming from
05:54 PM andypugh: Even with a - in the “Local version - append to kernel release” ?
05:55 PM memfrob: Are you using 4.14 tarball or kernel git 4.14.y branch?
05:55 PM andypugh: tarball
05:55 PM memfrob: ok so you're lucky
05:55 PM andypugh: This only needs to work once…
05:56 PM andypugh: (or, only needs to really work for one person, who makes the packages)
05:56 PM memfrob: well yeah but by practice it shouldn't be a PITH haha
05:56 PM andypugh: mozmck: Do you think we shoud aim to launch on Buster?
05:56 PM memfrob: *PITA
05:57 PM memfrob: In case anyone uses git like me, packagers should just use LOCALVERSION="" anyway to cancel whatever else debian wants to throw in, out.
05:58 PM memfrob: what keyboard are you using btw? Your ... shows up as one character.
06:00 PM andypugh: Apple aluminium ones
06:00 PM memfrob: andypugh, I just had an idea!
06:01 PM andypugh: I am having terrible trouble with git.
06:01 PM memfrob: For the kernel version string and all that, tie that to ./configure as an AC_CONFIG_LINK!
06:01 PM memfrob: Will Debian work that way though?
06:02 PM andypugh: Every time I try to pull from your repository I get merg conflicts, despite the fact that all I have done is. 1) Delete my own RTAI repo, 2) reset —hard to yours, then pull to be sure, 3) cherry pick the debian folder commit 4) push the new branch to my RTAI.
06:03 PM memfrob: I was having issues with git myself and rebased a few which is why that happened
06:03 PM memfrob: there is no more of that
06:04 PM memfrob: Basically in rules.in -- ./autogen.sh fills in FLAVOR like so: FLAVOR=@AC_FILLS_THIS_IN@
06:05 PM memfrob: configure: AC_FILLS_THIS_IN="Flavor here for Debian"
06:05 PM memfrob: I think that's a bad idea come now that I said that out loud.
06:05 PM andypugh: I was still trying to work out what it would do.
06:06 PM memfrob: autotools would replace what debian would do to rules.in or at least act like a helper
06:06 PM memfrob: but there's no way debian would like tha
06:06 PM memfrob: t
06:10 PM andypugh: I am recompiling LinuxCNC against 14.132 with the 132 patch from your RTAI
06:10 PM memfrob: core release 6?
06:10 PM memfrob: master or no-cal branch?
06:13 PM mozmck: andypugh: Buster would make the most sense to me. I noticed they use something different than iptables now though - nftables? Has anyone run an ethernet Mesa card with that yet?
06:14 PM andypugh: memfrob: no-cal first
06:14 PM mozmck: Seems like Mesa ethernet cards do something with iptables, but I may be mistaken.
06:14 PM andypugh: Maybe PCW has tried it?
06:15 PM mozmck: or pcw_home
06:15 PM mozmck: ;-)
06:16 PM memfrob: nftables is the new iptables but debian can use both
06:17 PM memfrob: nftables uses netfilter underneath, same stack as iptables, but different syntax
06:18 PM memfrob: what is the head of your no-cal branch? latest - 4773844?
06:19 PM memfrob: for ethernet mesa cards, your best bet is disabling iptables and or nftables entirely
06:21 PM memfrob: you can even use iptables syntax in nftables using iptables-nft
06:24 PM andypugh: memfrob: 4773844d5c46e9883b95a1fd22d2f94f7bee3f1f
06:25 PM andypugh: memfrob: Currently LinuxCNC compile against no-cal is failing “In file included from /usr/include/rtai/rtai.h:27:0,
06:25 PM andypugh: from /home/andypugh/linuxcnc-dev/src/rtapi/rtai_rtapi.c:77:
06:25 PM andypugh:
06:25 PM memfrob: mozmck: http://dpaste.com/15GKXGY
06:26 PM memfrob: this thread might be handy for you too: https://www.linuxquestions.org/questions/linux-security-4/nftables-rules-and-tables-dissapiered-after-reboot-4175649319/
06:26 PM memfrob: andypugh, need to see more of the error..
06:27 PM andypugh: #include <rtai_config.h>
06:27 PM mozmck: I don't think you want to disable nftables if you have the machine on the internet.
06:27 PM andypugh: Oh, I see, IRC was swallowing the / : /usr/include/rtai/asm/rtai.h:22:25: fatal error: rtai_config.h: No such file or directory
06:27 PM memfrob: linuxcnc needs that header??
06:27 PM memfrob: oh crap.
06:28 PM memfrob: i ran git grep in linuxcnc and didn't see a usage so i pulled it out..
06:28 PM memfrob: i thought rtai_config.h was used internally by RTAI only
06:28 PM memfrob: so i removed it from the install..
06:29 PM memfrob: wait it says: /usr/include/rtai/asm/rtai.h:22:25 not linuxcnc
06:29 PM memfrob: andypugh, this might have been error on my part brb
06:35 PM memfrob: bad commit.
06:39 PM andypugh: Bad Commit is a spa town in Germany?
07:01 PM memfrob: haha i wish
07:02 PM memfrob: ok, i'm fixing it right now, this is the last time i'm going back and re-writing things
07:02 PM andypugh: Anyway, my ambition for today is to get to sleep before the sun comes up, so I am stopping now.
07:02 PM memfrob: so you'll need to re-clone i apologize for this mess
07:02 PM memfrob: another day though!
07:03 PM memfrob: after that, it should be done
07:03 PM andypugh: Yes, I will pull master tomorrow, run through the prcess again and see where we are.
07:28 PM memfrob: all fixed now