Switch to DuckDuckGo Search
   April 22, 2020  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | >

NOTICE: This channel is no longer actively logged.

Toggle Join/Part | bottom
[00:01:48] *** am11 <am11!54f985dd@dsl-hkibng22-54f985-221.dhcp.inet.fi> has quit IRC (Remote host closed the connection)
[00:04:44] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has quit IRC (Ping timeout: 256 seconds)
[00:05:20] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has joined #illumos
[00:19:01] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has quit IRC (Quit: Textual IRC Client: www.textualapp.com)
[00:26:16] *** wiedi <wiedi!~wiedi@ip5b4096a6.dynamic.kabel-deutschland.de> has quit IRC (Read error: Connection reset by peer)
[00:26:52] *** wiedi <wiedi!~wiedi@ip5b4096a6.dynamic.kabel-deutschland.de> has joined #illumos
[00:35:48] *** andy_js <andy_js!~andy@> has quit IRC (Quit: andy_js)
[01:01:02] *** phyre <phyre!~phyre___@> has quit IRC (Remote host closed the connection)
[01:04:26] *** varna <varna!~varna@> has quit IRC (Ping timeout: 265 seconds)
[01:04:42] *** varna <varna!~varna@> has joined #illumos
[01:42:12] *** ypankov <ypankov!~ypankov@> has quit IRC (Ping timeout: 256 seconds)
[01:51:51] *** LeftWing <LeftWing!~jclulow@vortex.sysmgr.org> has quit IRC (Quit: upgrading)
[01:52:09] *** LeftWing <LeftWing!~jclulow@vortex.sysmgr.org> has joined #illumos
[02:50:12] *** ypankov <ypankov!~ypankov@> has joined #illumos
[02:55:24] <gitomat> [illumos-gate] 12450 Add support for BCM57765 family devices to bge -- Robert Mustacchi <rm at fingolfin dot org>
[02:55:25] <gitomat> [illumos-gate] 12496 bge mac address initialization is wrong -- Robert Mustacchi <rm at fingolfin dot org>
[02:56:51] <gitomat> [illumos-gate] 12482 Have /usr/bin/awk point to /usr/bin/nawk -- Dan McDonald <danmcd at joyent dot com>
[02:57:00] <danmcd> It's raining pushes!
[02:57:15] <rmustacc> Well, I'm glad I snuck that in so I didn't have to rebuild, haha.
[02:58:00] <danmcd> Yours was entirely confined to bge(7D) and packages, so I didn't feel compelled, though I should respin now that I'm back.
[02:58:16] <rmustacc> No worries. I have another build going for the next thing to push.
[02:58:54] <danmcd> I also feel bad for the OmniOSce gang... '034 is dropping soon (and yes, I'll admit, I wanna run stock dehydrated on '034).
[02:59:29] <danmcd> 151034's gonna be a big one here... I can consider what NFS I serve to move to purpose-built zones.
[03:00:02] <danmcd> (And honestly, if I should move some services to SMB3.)
[03:00:06] <rmustacc> Just means it'll be time to add more functionality to make you want to get onto bloody. ;)
[03:04:08] <LeftWing> I am excited for a stable OmniOS that includes 7119 so that I can make images more easily haha
[03:04:42] <rmustacc> We just need the rest of the gcp image bits as well.
[03:04:57] * LeftWing crumples into a ball of embarrassment
[03:05:27] <rmustacc> Sorry! That wasn't what I meant.
[03:05:42] <LeftWing> Digital Ocean's additional disk support is Virtio SCSI as it turns out, so I'll probably get that in next
[03:05:48] <LeftWing> Because I want to use those haha
[03:05:52] <rmustacc> We have the worst part of it sorted, which was really huge.
[03:06:07] <LeftWing> Yeah, I mean, the PIT stuff being out of the way is good news
[03:08:53] <danmcd> I'm not telling WEP, "Hey, your CCC backups are going on bits I'm changing at least twice-weekly!"
[03:10:37] <LeftWing> haha
[03:10:38] <rmustacc> Haha, that's fair.
[03:28:48] <gitomat> [illumos-gate] 12508 ndi_devi_alloc() and friends could take const char * names -- Robert Mustacchi <rm at fingolfin dot org>
[03:44:29] <LeftWing> danmcd: I think I must have hit the same awk thing with dehydrated. I have a /opt/dehydrated/workaround with a awk->nawk and grep->ggrep link that I prepend to PATH
[03:45:28] <danmcd> bahamat says that the grep one isn't so bad and can be dismissed.
[03:45:41] <danmcd> I have s/awk/nawk/g in my dehydrated for now.
[03:46:57] <ypankov> what is the grep issue?
[04:05:31] <LeftWing> ypankov: Good question! Must have been some flag we do not have
[04:05:38] <LeftWing> I should have made a note
[04:05:49] <LeftWing> I'll take a look
[04:09:57] <ypankov> yes, I'm wondering if it's a flag or some regex issue
[04:10:15] <jbk> hopefully not \< \> :)
[04:16:33] *** jcea <jcea!~Thunderbi@2001:bc8:2ecd:caed:7670:6e00:7670:6e00> has quit IRC (Quit: jcea)
[04:21:39] <danmcd> "grep -o", IIRC.
[04:22:20] <danmcd> "kebe" is MacOS X
[04:22:23] <danmcd> kebe(~)[134]% grep -o Kebe
[04:22:23] <danmcd> Kebe was here
[04:22:23] <danmcd> Kebe
[04:22:25] <danmcd> kebe(~)[0]% grep --version
[04:22:27] <danmcd> grep (BSD grep) 2.5.1-FreeBSD
[04:22:29] <danmcd> kebe(~)[0]%
[04:22:32] <danmcd> So I guess it's not just GNU grep.
[04:22:58] * danmcd has to leave, but wonders if -o might be a good RFE to file?
[04:38:27] *** am11 <am11!54f985dd@dsl-hkibng22-54f985-221.dhcp.inet.fi> has joined #illumos
[04:43:26] <am11> jbk: one of the failing test was related to exception handling: `terminate called after throwing an instance of 'PAL_SEHException'`. i read about it https://groups.google.com/forum/#!topic/envoy-users/tBxG321pk1Y, that order of link (libc before libgcc_s matters), i recompiled with that change, but failure is persisting. any ideas what else to look
[04:43:26] <am11> for
[04:43:28] <am11> ?
[04:44:47] <am11> (exit code: 134)
[04:45:05] *** hemi770 <hemi770!~hemi666@unaffiliated/hemi770> has quit IRC (Ping timeout: 250 seconds)
[04:45:28] <jbk> not offhand
[04:46:47] *** hemi770 <hemi770!~hemi666@unaffiliated/hemi770> has joined #illumos
[04:51:42] <am11> would a self-contained repro help investigation? the test is exception_handling.pal_sxs.test1: https://github.com/dotnet/runtime/blob/4f9ae42/src/coreclr/src/pal/tests/palsuite/exception_handling/pal_sxs/test1/dlltest2.cpp#L43 (macro expansion is bit of a roller-coaster ride).
[05:10:51] <LeftWing> danmcd: Do you have the whole nightly log
[05:14:23] *** tsoome <tsoome!~tsoome@> has quit IRC (Read error: Connection reset by peer)
[05:15:13] *** tsoome <tsoome!~tsoome@> has joined #illumos
[05:49:03] <am11> coredump stack looks like this http://sprunge.us/ZJkgSv
[06:22:42] *** PMT <PMT!~PMT@pool-100-33-69-155.nycmny.fios.verizon.net> has quit IRC (Ping timeout: 256 seconds)
[06:23:26] *** jeffpc <jeffpc!~jeffpc@josefsipek.net> has quit IRC (Ping timeout: 256 seconds)
[06:23:41] *** jeffpc <jeffpc!~jeffpc@josefsipek.net> has joined #illumos
[06:27:52] *** PMT <PMT!~PMT@pool-100-33-69-155.nycmny.fios.verizon.net> has joined #illumos
[06:43:57] *** tsoome <tsoome!~tsoome@> has quit IRC (Read error: Connection reset by peer)
[06:44:43] *** tsoome <tsoome!~tsoome@> has joined #illumos
[07:14:19] *** neirac <neirac!~neirac@pc-4-149-45-190.cm.vtr.net> has quit IRC (Ping timeout: 260 seconds)
[07:14:40] *** neirac <neirac!~neirac@pc-4-149-45-190.cm.vtr.net> has joined #illumos
[07:26:18] *** nde <nde!uid414739@gateway/web/irccloud.com/x-rodtxgnkqqyyluhy> has quit IRC (Quit: Connection closed for inactivity)
[07:41:34] *** neirac <neirac!~neirac@pc-4-149-45-190.cm.vtr.net> has quit IRC (Ping timeout: 240 seconds)
[07:45:58] *** arnoldoree <arnoldoree!~arnoldore@2001:d08:2100:62cf:574c:2307:6b59:748b> has joined #illumos
[07:46:44] *** neirac <neirac!~neirac@pc-4-149-45-190.cm.vtr.net> has joined #illumos
[08:04:54] <gitomat> [illumos-gate] backout: 12579 parallelize crypto-test build (needs more work) -- Joshua M. Clulow <josh at sysmgr dot org>
[08:15:23] *** BOKALDO <BOKALDO!~BOKALDO@> has joined #illumos
[08:25:19] *** pwinder <pwinder!~pwinder@> has joined #illumos
[08:43:36] *** phyre <phyre!~phyre___@> has joined #illumos
[08:48:07] *** amrfrsh <amrfrsh!~Thunderbi@> has quit IRC (Ping timeout: 260 seconds)
[08:53:03] *** alanc <alanc!~alanc@inet-hqmc01-o.oracle.com> has quit IRC (Remote host closed the connection)
[08:53:32] *** alanc <alanc!~alanc@inet-hqmc01-o.oracle.com> has joined #illumos
[09:00:25] <sjorge> LeftWing yeah i can log in on code.illumos.org, so I think it should be fine
[09:01:06] <LeftWing> Great!
[09:07:00] *** leoric <leoric!~leoric@> has joined #illumos
[09:21:31] *** wacki <wacki!~wacki@i577A563D.versanet.de> has joined #illumos
[09:27:43] <leoric> hm.. ctfmerge: failed to open main.32.o: File does not contain CTF data
[09:30:57] <leoric> dmake: Warning: Command failed for target `sha256_32_kcf'
[09:35:55] *** jimklimov <jimklimov!~jimklimov@ip-86-49-254-26.net.upcbroadband.cz> has joined #illumos
[09:36:07] <leoric> peerhaps, needs just git clean...
[09:39:33] <leoric> 12579 parallelize crypto-test build ?
[09:41:28] <leoric> pmooney: ^
[09:49:02] <wacki> leoric: 12579 has been backed out this morning
[09:50:41] <leoric> Ah, so it's our issue
[10:01:39] <wacki> Probably not; it was 8:05 am. So after our jenkins build has been run.
[10:02:20] <LeftWing> Sorry about that folks, it appears to be a race that only happens on some systems
[10:02:34] <LeftWing> If it is not fixed with the backout please let us know
[10:03:22] <tsoome_> I was hit by race on my t4 in one run and clean on another:D
[10:04:16] *** neirac <neirac!~neirac@pc-4-149-45-190.cm.vtr.net> has quit IRC (Ping timeout: 256 seconds)
[10:05:34] <wacki> Parallelism :D
[10:05:51] *** am11 <am11!54f985dd@dsl-hkibng22-54f985-221.dhcp.inet.fi> has quit IRC (Remote host closed the connection)
[10:07:19] *** neirac <neirac!~neirac@pc-4-149-45-190.cm.vtr.net> has joined #illumos
[10:09:53] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has joined #illumos
[10:14:57] *** man_u <man_u!~manu@89-92-19-81.hfc.dyn.abo.bbox.fr> has joined #illumos
[10:38:09] *** andy_js <andy_js!~andy@> has joined #illumos
[10:40:30] *** vila <vila!~vila@laubervilliers-659-1-133-94.w80-15.abo.wanadoo.fr> has quit IRC (Remote host closed the connection)
[10:43:31] *** hemi770 <hemi770!~hemi666@unaffiliated/hemi770> has quit IRC (Remote host closed the connection)
[10:43:49] *** tsoome <tsoome!~tsoome@> has quit IRC (Read error: Connection reset by peer)
[10:44:41] *** vila <vila!~vila@laubervilliers-659-1-133-94.w80-15.abo.wanadoo.fr> has joined #illumos
[10:45:23] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[10:50:26] <sjorge> OK, I think I got a untouched gate to build on OmniOS
[10:50:31] <sjorge> now to try with the patch on top :)
[10:50:38] <sjorge> En erm then open a cr
[10:54:35] *** BOKALDO <BOKALDO!~BOKALDO@> has quit IRC (Read error: Connection reset by peer)
[10:54:49] *** Riastradh <Riastradh!~riastradh@netbsd/developer/riastradh> has quit IRC (Ping timeout: 264 seconds)
[10:54:54] *** BOKALDO <BOKALDO!~BOKALDO@> has joined #illumos
[10:56:54] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Ping timeout: 256 seconds)
[10:58:06] *** hemi770 <hemi770!~hemi666@unaffiliated/hemi770> has joined #illumos
[11:09:23] *** amrfrsh <amrfrsh!~Thunderbi@> has joined #illumos
[11:10:54] *** EisNerd <EisNerd!~manschwet@mail.cs-software-gmbh.de> has quit IRC (Ping timeout: 240 seconds)
[11:27:37] *** Exagone313 <Exagone313!exa@3131313.xyz> has quit IRC (Ping timeout: 265 seconds)
[11:28:35] *** ypankov <ypankov!~ypankov@> has quit IRC (Remote host closed the connection)
[11:31:35] *** Exagone313 <Exagone313!exa@3131313.xyz> has joined #illumos
[11:50:23] *** Exagone313 <Exagone313!exa@3131313.xyz> has quit IRC (Quit: see ya!)
[11:51:15] *** Exagone313 <Exagone313!exa@3131313.xyz> has joined #illumos
[11:52:18] *** EisNerd <EisNerd!~manschwet@mail.cs-software-gmbh.de> has joined #illumos
[11:55:34] *** kovert <kovert!~kovert@> has quit IRC (Ping timeout: 240 seconds)
[11:58:23] *** EisNerd <EisNerd!~manschwet@mail.cs-software-gmbh.de> has quit IRC (Ping timeout: 265 seconds)
[12:19:06] *** kovert <kovert!~kovert@> has joined #illumos
[12:49:16] <sjorge> didn't we have a docs page somewhere on how to submit a change to code.illumos.org? I remember vaguely about adding that as a new remote and doing some git commit --amend to add a special ID before pushing
[12:49:29] <sjorge> But I find no mentioned of it on illumos.org or on the dev-guide repo
[12:49:50] <jlevon> I think just an old email from josh
[12:53:08] <sjorge> ah
[12:55:56] <sjorge> ofcourse looking for code.illumos.org or gerrit returns over a 100 mails
[12:59:26] <andyf> sjorge: the help in gerrit is pretty good too
[12:59:52] <sjorge> What I did for the old joyent one didn't seem to work though
[13:00:10] <sjorge> But my lunch break is over now, will poke more at it later
[13:00:23] <andyf> yes, this is newer and needs a change id, any you always push to refs/for/master
[13:00:28] <andyf> https://code.illumos.org/Documentation/intro-gerrit-walkthrough.html
[13:01:00] <sjorge> Got it I think https://illumos.topicbox.com/groups/advocates/T4b033899883e06a3-Mdaf0e17883570db142edd4ff/heads-up-changing-git-urls-for-gerrit-cutover
[13:02:58] <jlevon> https://github.com/jlevon/grot/blob/master/gerrit-change-id is handy
[13:06:04] <sjorge> That is handy indeed :)
[13:12:06] <sjorge> I think I did it alright https://code.illumos.org/c/illumos-gate/+/572
[13:12:19] <sjorge> gerrit is not happy with my key vs my e-mail though
[13:12:41] <sjorge> something about key comment using my private domain but my e-mail/gerrit email being my public domain, but oh well...
[13:15:16] <sjorge> So I now leave this up for a few weeks to get feedback and update it with a new patchset once I have processed the feedback and after a few weeks of no new feedback send an e-mail to the list?
[13:15:25] <sjorge> if enough people +1 it that is
[13:20:10] <jlevon> you don't need to wait weeks
[13:20:25] <jlevon> if you have >= 1 code review and you're ready for RTI go ahead.
[13:21:21] <jlevon> if it were more contentious you might want to wait around a while
[13:21:30] <jlevon> you sent email out with the link right?
[13:33:54] *** amrfrsh <amrfrsh!~Thunderbi@> has quit IRC (Quit: amrfrsh)
[14:01:15] *** EisNerd <EisNerd!~manschwet@mail.cs-software-gmbh.de> has joined #illumos
[14:07:24] <sjorge> err no? I did not send an e-mail with the link
[14:07:49] <sjorge> It's not super clear what to do next after creating the it on gerrit
[14:09:58] <jlevon> sjorge: if you know certain people should review, add them on gerrit. either way, email the list with the link and ask for reviews.
[14:12:28] <sjorge> ah wait wait yeah that makes sense, mail -developers for reviews, collect reviews -> mail -advocates for RTI
[14:14:24] <jlevon> yes
[14:16:54] *** jcea <jcea!~Thunderbi@2001:bc8:2ecd:caed:7670:6e00:7670:6e00> has joined #illumos
[14:21:27] *** phyre <phyre!~phyre___@> has quit IRC (Quit: Saliendo)
[14:22:25] <sjorge> done
[14:22:58] <sjorge> And just pushing to the same remote is so much easier than the older gerrit @ joyent
[14:39:30] *** amrfrsh <amrfrsh!~Thunderbi@> has joined #illumos
[14:47:46] *** am11 <am11!54f985dd@dsl-hkibng22-54f985-221.dhcp.inet.fi> has joined #illumos
[14:48:46] <wilbury> sjorge: i left you some comments inline.
[14:59:26] <sjorge> wilbury it actually works with and without the / in front for the allow= or deny=
[14:59:30] <sjorge> I double checked
[15:00:25] <wilbury> oh!
[15:02:33] <sjorge> My first instinct was to add it too
[15:02:43] <sjorge> double checking all other pam_* man pages now
[15:03:32] *** KungFuJesus <KungFuJesus!~adam@> has quit IRC (Quit: Lost terminal)
[15:06:27] <sjorge> TIL pam_smbfs_login is a thing
[15:09:18] <sjorge> nice, no other thing seem to have a configurable path that gets passed
[15:09:29] <sjorge> So updating it won't be inconsistant :)
[15:12:35] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has joined #illumos
[15:15:55] <wilbury> sjorge: allowdeny_filename is being fopen()'d so i'd say that slash should be there.
[15:16:35] <sjorge> I guess wether it works would depend on the cwd for the process
[15:17:26] <wilbury> yes, in that case, yes. without leading / it only would work with cwd /
[15:32:39] <sjorge> Now to do some more work and look at doing the list.c comments later tonight
[15:33:26] *** kovert <kovert!~kovert@> has quit IRC (Ping timeout: 256 seconds)
[15:35:44] *** kovert <kovert!~kovert@> has joined #illumos
[15:36:18] *** eki <eki!~eki@dsl-hkibng41-54f858-46.dhcp.inet.fi> has quit IRC (Quit: leaving)
[15:36:43] <andyf> We'll keep you busy :)
[15:38:05] *** eki <eki!~eki@dsl-hkibng41-54f858-46.dhcp.inet.fi> has joined #illumos
[16:06:13] *** fanta1 <fanta1!~fanta1@p200300F76BC47B0011E9931A6A8199BF.dip0.t-ipconnect.de> has joined #illumos
[16:16:56] <sjorge> Hmmm... I changed realloc to malloc and droped the if before the free and now get
[16:16:57] <sjorge> ../list.c:219 pam_sm_acct_mgmt() warn: possible memory leak of 'grbuf
[16:17:03] <sjorge> And compile fails
[16:18:10] <andyf> That's because there are several places in the code where it returns
[16:18:28] <andyf> It's a real buf
[16:18:30] <andyf> *bug
[16:18:40] <sjorge> So why was it not complaining with realloc?
[16:18:56] <andyf> Once you have allocated the memory for grbuf(), you need to make sure you do not return from the function without freeing it
[16:19:21] <andyf> Probably a limitation of whatever is warning you now
[16:20:19] <sjorge> I guess whereever '(void) fclose(fd);' is called is a good indication
[16:20:27] <sjorge> The are all over the place
[16:20:32] <sjorge> Usually in error handling bits
[16:20:38] <andyf> If you move the malloc bit down to just above the start of the fgets loop, that will skip two of the cases
[16:20:42] <sjorge> But odd it doesn't even compile anymore
[16:20:57] <sjorge> Originally I was doing the alloc and free inside the loop
[16:21:00] <andyf> that is probably smatch that's complaining
[16:21:02] <sjorge> So for each group entry
[16:21:22] <sjorge> But I think jbk said that was too slow? let me look up in the logs
[16:21:42] <andyf> It's definitely better to allocate once and there are a number of ways to skin it (like everything)
[16:21:43] *** tsoome <tsoome!~tsoome@> has joined #illumos
[16:22:08] <andyf> if you move the calloc() down a bit I think there are just two places where you now need to add a free(grbuf) after fclose(fd)
[16:22:16] <sjorge> Imagine me being the person who has never touched a gun... that you give a loaded shutgun and is pointing it at his face
[16:22:18] <sjorge> that is me right now
[16:23:20] <sjorge> heh you're right moving it to just above the while makes smack happy
[16:24:17] <sjorge> would calloc over malloc make sense?
[16:24:27] <sjorge> We're reusing the grbuf in the loop right for every group entry?
[16:25:02] <andyf> Somebody else suggested calloc() over malloc().. I don't think it's important for this code
[16:25:03] <sjorge> so it could be all zero, filled half with group_a and then nearly full with group_b because it has a ton of members, then back to nearly nothing with group_c ?
[16:25:21] <andyf> given the way it's used, malloc() should be fine IMO
[16:25:42] <sjorge> I check the pam code and it's mostly malloc with a few exceptions where it uses calloc
[16:28:30] <sjorge> oops :p most calls are in list.c and are the ones I just added
[16:28:35] <sjorge> might be best to use calloc then
[16:29:49] <sjorge> I guess I should also change grbuflen back to size_t then as calloc seems to want a size_t not an int
[16:30:12] *** jimklimov1 <jimklimov1!~jimklimov@ip-86-49-254-26.net.upcbroadband.cz> has quit IRC (Read error: Connection reset by peer)
[16:31:07] <sjorge> I'm guessing smack isn't perfect wrt branch prediction?
[16:31:27] <sjorge> As I can see some return paths where grbuf is not freed
[16:31:47] <andyf> sometimes it is cleverer than us..
[16:32:21] <andyf> for example, LIST_COMPAT_MODE and check_group together is not supported..
[16:32:40] <andyf> so it will actually never his those return statements if grbuf was allocated..
[16:32:44] <andyf> *hit
[16:32:47] <am11> regarding the atan2 bug, it is confined to unoptimized code, i.e. `curl -s http://sprunge.us/WqQiP1 | gcc -xc -O2 - -lm; ./a.out` returns the same value as linux and macos on smartos. without -O2, we get `hard-coded 3.14159 vs. soft-coded 0`.
[16:32:52] <sjorge> hmmm... you're right haha!
[16:33:00] <andyf> but you should still add them I think
[16:33:24] *** tru_tru <tru_tru!~tru@> has quit IRC (Ping timeout: 265 seconds)
[16:33:27] <sjorge> It's not much effort to do
[16:34:03] <am11> asm diffs of unoptimized code: https://paste2.org/LfbL5MDk.
[16:34:18] <am11> with -O2, no delta is produced.
[16:37:59] <sjorge> andyf: well good news is, I think i address the comments and everything still works :D
[16:38:28] <sjorge> on the condition I actually use the correct password :D
[16:39:50] <sjorge> @wilbury I was also able to verify the broken ness with etc/users.allow by forcing a cwd of non /
[16:39:51] <sjorge> ldap1 sudo[4381]: [ID 266943 auth.error] pam_list: fopen of etc/users.allow: No such file or directory
[16:40:01] <sjorge> So the error is correct, the man is indeed wrong
[16:41:12] <wilbury> told'ya! :-) if you'll need a bit of help, just shout.
[16:42:10] <sjorge> I think I got most list.c feedback applied
[16:42:17] <sjorge> retresting all them things again
[16:43:57] <sjorge> I both love and hate you don't need to restart anything after editing pam.conf
[16:44:02] <sjorge> ... gues who just locked himself out
[16:46:23] *** KeiraT <KeiraT!~k4ra@gateway/tor-sasl/k4ra> has quit IRC (Ping timeout: 240 seconds)
[16:46:43] *** cantstanya <cantstanya!~chatting@gateway/tor-sasl/cantstanya> has quit IRC (Ping timeout: 240 seconds)
[16:49:38] <sjorge> @am11 for your opimization question, hopefully someone who how the compiler stuff works will get back to you soon
[16:52:30] <sjorge> andyf is there any easy way to run cstyle and smack just for a single file?
[16:53:00] <andyf> for cstyle and other checks, `git pbchk -p master` if you are in a build environment (via bldenv)
[16:53:24] <andyf> for smatch, you need to build but you can usually change to the right directory and just build the one thing you're working on
[16:53:44] <andyf> `cd usr/src/lib/pam_modules; dmake` will probably do it
[16:53:49] <sjorge> ack, I assume dmake -e install in usr/src/lib/pam_modules/list was Ok then
[16:53:52] <andyf> again, within a build environment
[16:54:44] <sjorge> do I use /opt/onbld/bin/bldenv or /build/illumos-gate/usr/src/tools/scripts/bldenv
[16:54:55] <andyf> the second one is safest
[16:55:04] <andyf> you should always use one matching the code you're working on
[16:55:32] <andyf> but cd /build/illumos-gate; ./usr/src/tools/scripts bldenv /opt/onbld/env/omnios-illumos-gate
[16:55:35] <andyf> something like that
[16:55:47] <sjorge> hmmm it does complain about CODEMGR_WS
[16:56:21] <sjorge> i'm an idiot
[16:56:31] <sjorge> forgot the .
[16:56:44] <wilbury> sjorge: guess who locked himself out yesterday when fixing friend's pf.conf on his freebsd box? :-)
[16:57:16] <sjorge> wilbury: It's why I run a small rpi2 with freebsd that has serial console access to all my servers
[16:58:33] <wilbury> sjorge: yeah, that was a server at friend's home so not a big deal. he spent 2 days trying to figure out why pf is not working on 13-current. and the culprit was in "pfctl -si". "Status: Disabled" :-D after pfctl -e, everything started to work.
[16:58:36] <sjorge> andyf might be worth adding that to the buidling illumos on omnios page too
[16:58:59] *** arnoldoree <arnoldoree!~arnoldore@2001:d08:2100:62cf:574c:2307:6b59:748b> has quit IRC (Ping timeout: 252 seconds)
[16:59:04] <andyf> Could do - it's covered in the illumos developers' guide but extra notes can't hurt
[16:59:33] <andyf> pkg refresh
[16:59:40] <andyf> ^ oops
[17:01:39] <sjorge> Any idea what to do with "WARNING: skipping paragraph macro: PP after SH"
[17:01:59] <sjorge> Lots of warnings for the pam_list.5 man page, but doesn't look like stuff I introduced
[17:02:27] <jlevon> delete them
[17:02:35] <jlevon> we clean those up as we change man pages
[17:02:47] <sjorge> So just delete 'PP' ?
[17:03:59] <sjorge> Yep that seems to do it
[17:04:08] <jlevon> aye
[17:04:22] *** KeiraT <KeiraT!~k4ra@gateway/tor-sasl/k4ra> has joined #illumos
[17:04:48] <sjorge> Do I also need to fix "no copyright claim for current year found" ?
[17:05:26] <wilbury> i also have such warnings for my webrev for dispadmin.1m
[17:05:33] *** cantstanya <cantstanya!~chatting@gateway/tor-sasl/cantstanya> has joined #illumos
[17:05:51] *** jamtorus <jamtorus!~quassel@s91904423.blix.com> has joined #illumos
[17:05:52] *** jamtorus <jamtorus!~quassel@s91904423.blix.com> has quit IRC (Client Quit)
[17:05:52] *** jellydonut <jellydonut!~quassel@s91904423.blix.com> has quit IRC (Ping timeout: 256 seconds)
[17:06:14] <andyf> You don't have to
[17:06:32] <andyf> CDDL requires something but that is covered these days by the git commit metadata
[17:06:50] <andyf> but you should feel free to add yourself to the copyrights
[17:06:58] *** jellydonut <jellydonut!~quassel@s91904423.blix.com> has joined #illumos
[17:07:15] *** jellydonut <jellydonut!~quassel@s91904423.blix.com> has quit IRC (Client Quit)
[17:08:06] <wilbury> even for small changes?
[17:08:08] *** jellydonut <jellydonut!~quassel@s91904423.blix.com> has joined #illumos
[17:08:31] <sjorge> Same question basically, I don't think is a super small change
[17:08:43] <sjorge> I never did it for the tiny changes to zhyve in SmartOS either
[17:08:46] <rmustacc> Copyright warnings are option.
[17:08:48] <rmustacc> *optional
[17:09:04] <rmustacc> They are there for folks for whom policy dictates that (which came from Sun and is true for a number of folks)
[17:09:41] <andyf> sjorge: do update the man page date though
[17:10:00] <sjorge> andyf: will do, no pbchk warning for that though... perhaps there should be?
[17:13:05] <rmustacc> Tooling can certainly be improved (though there is usually a distinction between new content and say typos)
[17:13:22] <sjorge> new content in this case
[17:13:29] <sjorge> and cleanup, and some fixes
[17:17:03] *** arnoldoree <arnoldoree!~arnoldore@> has joined #illumos
[17:18:04] <sjorge> thanks for all the feedback so far
[17:18:14] <sjorge> also nice, gerrit allows to add a note to the update patch set!
[17:27:16] <am11> and with clang-9, both hard-coded and soft-coded values are 0 with and without -O2 (whereas, the expected value is the value of pi).
[17:27:58] <am11> sjorge: yup, just posting the results of some investigations. maybe someone will pick up, if the situation is unknown.
[17:29:18] <gitomat> [illumos-gate] 12480 long mblk chain will cause mlxcx to stop sending -- Paul Winder <pwinder at racktopsystems dot com>
[17:32:15] <rmustacc> There's probably a bug somewhere with the libm bit. We'll need to figure out what's going on and double check the varoius C / POSIX standards and whether this is something where an ASM optimization in libm is going wrong or not.
[17:35:58] <am11> i got the same feeling that libm (vai libc) are somehow going wrong. clang vs. gcc with/without -O2 giving three unique set of results.
[17:36:51] <rmustacc> The difference in asm that you shared was between whether it's being optimized out or not.
[17:37:39] <rmustacc> Someting for someone to investigate.
[17:42:22] <am11> yup, that diff was captured on SmartOS, using gcc7 (-m64 default) with and without -O2. I then reconfirmed on OpenIndiana, with `-m64` (as `-m32` is default there for some reason..) , gcc7 and gcc9 produce the same results for with/without -O2. And clang9 on OI with -m64 produces 0,0 (whereas it should be valueof(pi),valueof(pi)).
[17:49:01] <pwinder> I just did a build with the awk changes - the build went fine. When I checkout another branch I got "error: The following untracked working tree files would be overwritten by checkout: usr/src/man/man1/nawk.1"
[17:50:06] <andyf> It may be that there are some clobber rules that need updating
[17:51:24] <andyf> or did you run `dmake clobber` before switching branches?
[17:51:38] <andyf> I would just remove that file and then switch branch
[17:51:57] *** Riastradh <Riastradh!~riastradh@netbsd/developer/riastradh> has joined #illumos
[17:56:38] <wilbury> do we have something like "nmdm"?
[17:56:46] *** pwinder <pwinder!~pwinder@> has quit IRC (Quit: This computer has gone to sleep)
[17:56:52] <wilbury> if not, i'd like to port it from freebsd as a exercise
[17:57:09] <wilbury> software nullmodem.it is very handy for bhyve serial consoles
[17:57:36] <andyf> How does it work? We currently bhyve wired up so that its serial console is connected to the zone console
[17:57:45] <andyf> so `zlogin -C <VM>` gets you on the console
[17:58:24] <wilbury> andyf: upon opening /dev/nmdmXXz (XX = 0..99, z = A|B) the A and B parts are "connected" together
[17:58:51] <wilbury> you specify like A part as a serial console for bhyve and connect to B part using tip/picocom/cu/... from the host system
[17:59:46] <wilbury> i run my openindiana: /data/sysop/bin/vmrun.sh -c 4 -m 8G -d /dev/zvol/zroot/data/bhyve/oi1/disk0 -d /dev/zvol/zroot/data/bhyve/oi1/disk1 -d /dev/zvol/zroot/data/bhyve/oi1/disk2 -t tap5 -w -E -P 5951 -C /dev/nmdm51B oi1
[18:00:13] <wilbury> so when i cinnect cu -e -s -s 115200 -l /dev/nmdm51A, i'm on oi1's serial console
[18:00:28] <wilbury> (with properly configured console and grub, ofc)
[18:01:24] <wilbury> (host is freebsd)
[18:02:19] <sjorge> wilbury: nmdm is pretty nice yeah
[18:02:43] <wilbury> and it's of a pretty much use with bhyve
[18:02:55] <sjorge> wilbury: I mostly use socat with unixsocks on SmartOS, but nmdm would be nice, especially if it works in a zone too
[18:03:53] <wilbury> it's a simple character devices, could be made zone aware, too. indeed.
[18:11:57] <rmustacc> There have been requests for variants of that at least. e.g. the ability to create a UDS that has visibility cross-zone without playing games with filesystem tricks.
[18:26:22] <am11> jbk: will something like https://paste2.org/Hx1vYtpn close https://www.illumos.org/issues/4963?
[18:27:25] <jbk> i don't know offhand.. i last really looked at it 6 years ago, so i've since forgot a lot of the context
[18:27:38] <jbk> and I seem to recall it involved quite a bit of digging around the first time
[18:28:35] <wilbury> rmustacc: you mean lofi shared across zones?
[18:28:44] <wilbury> sort of
[18:30:09] <rmustacc> No, I do'nt.
[18:30:34] <rmustacc> There's a certian pattern that folks have trying to expose services over say a uds or door in a zone which require entering the zone to create it to know which zone it is.
[18:30:37] <wilbury> i mean "by filesystem tricks"
[18:30:46] <rmustacc> No, they use zlogin and create things.
[18:31:00] <rmustacc> And then watch for zone creation/halt events to try and remove it so they don't block zone shutdown.
[18:31:16] <wilbury> i see.
[18:32:07] *** fanta1 <fanta1!~fanta1@p200300F76BC47B0011E9931A6A8199BF.dip0.t-ipconnect.de> has quit IRC (Quit: fanta1)
[18:32:27] <rmustacc> Your case is a simpler variant of basically establishing a point to point bridde.
[18:32:30] <rmustacc> *bridge
[18:36:43] *** tru_tru <tru_tru!~tru@> has joined #illumos
[18:41:07] <am11> jbk: i was thinking if there is a risk in this change, would the same risk be applied in using `priocntl(2)` directly (i.e. without updating `tsparms_t.ts_uprilim` each time)?
[18:41:34] *** man_u <man_u!~manu@89-92-19-81.hfc.dyn.abo.bbox.fr> has quit IRC (Ping timeout: 256 seconds)
[19:04:30] <jlevon> am11: we should just fix that
[19:05:11] *** pwinder <pwinder!~pwinder@> has joined #illumos
[19:05:29] <jlevon> am11: your fix seems fine to me...
[19:05:53] *** pwinder <pwinder!~pwinder@> has quit IRC (Client Quit)
[19:07:59] *** leoric <leoric!~leoric@> has quit IRC (Quit: Konversation terminated!)
[19:10:27] <am11> jlevon: great, could you please upstream that patch when time permits? i haven't registered the Illumos account. :)
[19:11:00] <jlevon> I'll see if I can find some time to do the necessary testing etc.
[19:11:44] <am11> +1, no rush, thank you!
[19:23:47] *** Kurlon_ <Kurlon_!~Kurlon@cpe-67-253-136-97.rochester.res.rr.com> has quit IRC (Remote host closed the connection)
[19:24:22] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has joined #illumos
[19:27:49] *** BOKALDO <BOKALDO!~BOKALDO@> has quit IRC (Quit: Leaving)
[19:34:54] *** tg2_ <tg2_!~tg2@> has quit IRC (Ping timeout: 240 seconds)
[19:44:36] *** am11 <am11!54f985dd@dsl-hkibng22-54f985-221.dhcp.inet.fi> has quit IRC (Ping timeout: 240 seconds)
[20:16:48] *** arnoldoree <arnoldoree!~arnoldore@> has quit IRC (Quit: Leaving)
[20:30:20] *** ldepandis <ldepandis!~ldepandis@unaffiliated/ldepandis> has joined #illumos
[21:42:31] *** ypankov <ypankov!~ypankov@> has joined #illumos
[22:06:57] *** igork <igork!~igork@> has quit IRC (Read error: Connection reset by peer)
[22:07:14] *** igork <igork!~igork@> has joined #illumos
[22:14:13] *** tg2 <tg2!~tg2@> has joined #illumos
[22:18:44] *** Qatz is now known as DaQatz
[22:19:32] *** DaQatz is now known as Qatz
[22:27:14] *** tomww <tomww!~tom@unaffiliated/tomww> has quit IRC (Ping timeout: 240 seconds)
[22:27:45] *** tomww <tomww!~tom@unaffiliated/tomww> has joined #illumos
[22:27:51] *** leah2 <leah2!~leah@vuxu.org> has quit IRC (Ping timeout: 265 seconds)
[22:30:03] *** leah2 <leah2!~leah@vuxu.org> has joined #illumos
[22:33:04] *** domag02 <domag02!b03fd4e2@catv-176-63-212-226.catv.broadband.hu> has joined #illumos
[22:33:12] <domag02> hi!
[22:37:48] <neirac> domag02 hi
[22:46:10] *** wacki <wacki!~wacki@i577A563D.versanet.de> has quit IRC (Quit: Lingo: www.lingoirc.com)
[22:56:30] *** domag02 <domag02!b03fd4e2@catv-176-63-212-226.catv.broadband.hu> has quit IRC (Remote host closed the connection)
[23:07:14] <toastersonerson1> neirac (IRC): did you end up fixing github.com/docker/docker/pkg/term ?
[23:09:13] <neirac> toastersonerson1 yes, I just fixed it
[23:09:41] <toastersonerson1> goo its the only thing missing for kubelet to compile
[23:09:45] <neirac> now nomad works with the web ui :)
[23:10:20] <neuroserve> \o/
[23:10:28] <toastersonerson1> is the PR still pending then I guess.
[23:11:34] *** Kurlon_ <Kurlon_!~Kurlon@cpe-67-253-136-97.rochester.res.rr.com> has joined #illumos
[23:14:34] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has quit IRC (Ping timeout: 240 seconds)
[23:18:06] <gitomat> [illumos-gate] 12586 zvol_write() can use dmu_tx_hold_write_by_dnode() -- Matthew Ahrens <mahrens at delphix dot com>
[23:21:22] <neirac> toastersonerson1 I need to do more tests there are a couple of issues that the link speed is not detected and the interface, I just tested the raw_exec drive,the zones driver
[23:21:55] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has quit IRC (Quit: Leaving)
[23:27:46] <neirac> toastersonerson1 what's goo?
[23:28:27] <toastersonerson1> good without a d
[23:31:15] <neirac> toastersonerson1 oh, I'm sorry I meant what's missing for kubelet
[23:31:35] <toastersonerson1> only docker term
[23:32:43] <neirac> toastersonerson1 oh, that's great, take the one from nomad/illumos branch and give it a try
[23:33:02] <toastersonerson1> the rest was trivial and I have aPR prepared. as it is the lowest component in the chain the others should compile without a problem
[23:33:11] <neirac> toastersonerson1 let me know if that works and I'll fix what's needed
[23:33:35] <neirac> toastersonerson1 why you were working on kubelet for openaurora right ? you are almost done ?
[23:34:17] <toastersonerson1> only at the begining. starting with getting things compiled. the docker part is usable though
[23:35:40] <neirac> toastersonerson1 work should be start being easier aalong the way.
[23:36:33] <neirac> toastersonerson1 how do I add go-zone as a vendor? I'll need to makae more changes to it, I need to support an illumos brand from omnios that could take a zss as well, I remember I did somethign for lx on go-zone
[23:37:19] <toastersonerson1> go modules should allow you to vendor all dependencies
[23:37:47] <toastersonerson1> or how do you mean?
[23:39:43] <neirac> toastersonerson1 oh, ok I'm sorry I'm have not read about go vendor I'll do that first
[23:41:06] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has quit IRC (Quit: Textual IRC Client: www.textualapp.com)
[23:42:43] <toastersonerson1> the go tooling has all the features the community uses on a often basis :)
[23:44:48] *** BrownBear <BrownBear!~BrownBear@> has quit IRC (Quit: Ping timeout (120 seconds))
[23:45:07] *** deejam <deejam!deejam@irc.nathanic.org> has quit IRC (Remote host closed the connection)
[23:45:22] *** dlg_ <dlg_!~dlg@toy.eait.uq.edu.au> has quit IRC (Read error: Connection reset by peer)
[23:45:29] *** dlg <dlg!~dlg@toy.eait.uq.edu.au> has joined #illumos
[23:45:50] *** EisNerd <EisNerd!~manschwet@mail.cs-software-gmbh.de> has quit IRC (Remote host closed the connection)
[23:46:11] *** EisNerd <EisNerd!~manschwet@mail.cs-software-gmbh.de> has joined #illumos
[23:47:26] *** clapont <clapont!~clapont@unaffiliated/clapont> has quit IRC (Ping timeout: 265 seconds)
[23:47:26] *** MilkmanDan <MilkmanDan!~dan@wilug/expat/MilkmanDan> has quit IRC (Ping timeout: 265 seconds)
[23:47:55] *** PMT <PMT!~PMT@pool-100-33-69-155.nycmny.fios.verizon.net> has quit IRC (Ping timeout: 265 seconds)
[23:48:24] *** PMT <PMT!~PMT@pool-100-33-69-155.nycmny.fios.verizon.net> has joined #illumos
[23:48:30] *** clapont <clapont!~clapont@unaffiliated/clapont> has joined #illumos
[23:49:47] *** MilkmanDan <MilkmanDan!~dan@wilug/expat/MilkmanDan> has joined #illumos
[23:50:10] *** deejam <deejam!deejam@irc.nathanic.org> has joined #illumos
[23:54:42] <neirac> toastersonerson1 what's the issue on delve ? do you have a repo to take a look? I remember that it was easier that update the mdb go support
[23:55:18] <toastersonerson1> delve is missing a bit of c code to locate the process to attach
[23:55:40] <toastersonerson1> pgrep and pstack basicly
[23:56:17] <toastersonerson1> let me find the missing spot
[23:56:32] <neirac> ok
[23:57:18] <toastersonerson1> https://github.com/go-delve/delve/blob/master/pkg/proc/gdbserial/gdbserver_unix.go
[23:57:49] <toastersonerson1> https://github.com/go-delve/delve/blob/master/pkg/proc/native/proc_freebsd.c
[23:58:04] <toastersonerson1> and something that implents the same as this freebsd code
[23:58:21] <neirac> toastersonerson1 let me take notes
[23:59:14] <toastersonerson1> basicly all the native parts need illumos versions like freebsd
[23:59:26] <neirac> oh ok, seems pretty straight forward
[23:59:46] *** BrownBear <BrownBear!~BrownBear@> has joined #illumos

   April 22, 2020  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | >