IRC logs for #baserock for Wednesday, 2014-11-26

*** genii [~quassel@ubuntu/member/genii] has quit [Read error: Connection reset by peer]00:54
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock01:42
*** zoli_ [~zoli_@linaro/zoli] has quit [Ping timeout: 264 seconds]01:46
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock04:15
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]04:35
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock05:12
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]06:29
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock06:29
*** zoli_ [~zoli_@linaro/zoli] has quit [Ping timeout: 250 seconds]06:34
*** genii [~quassel@ubuntu/member/genii] has joined #baserock06:50
*** genii [~quassel@ubuntu/member/genii] has quit [Read error: Connection reset by peer]06:51
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock07:03
*** cyndis [] has quit [Ping timeout: 272 seconds]07:49
*** cyndis [] has joined #baserock07:51
petefothI haven't pushed the branch containing my ‘ Documentation for KVM Write Extension’ patches as I'm not in baserock-writers08:40
*** rdale [] has joined #baserock08:48
paulsher1oodpetefoth: do you want to be? it's a git-intensive role, and quite a responsibility :)08:57
*** sherm_ [] has joined #baserock08:59
petefothpaulsher1ood: I don’t think i do. To have a remote where I can push branches for review, is it best to use my github account, or gitlab or …? Maybe some guidance on this could be in the ‘Contributing’ section of w.b.o (if it isn’t already - I’ll go and look) 09:00
*** mariaderidder [] has joined #baserock09:00
*** bashrc [] has joined #baserock09:01
paulsher1oodpetefoth: any public git repo host would be fine, i think. preferably one with cgit web ui or similar09:12
paulsher1ood+1 for any improvement to the wiki :)09:12
petefothpaulsher1ood: thanks09:12
*** tiagogomes [] has joined #baserock09:19
*** jonathanmaw [] has joined #baserock09:26
pedroalvarezhm... mason-x86-64 has panicked (I believe)09:40
pedroalvarezthiis is the log I can get before rebooting:
pedroalvarezlooking at the journal, the only thing I know is that it was trying to build ruby09:47
persiaWas there filesystem pressure (e.g. not enough space)?  That seems to be one of many causes of this sort of thing.09:49
pedroalvarezis the first thing I checked when I rebooted the system09:49
pedroalvarezit has 20 G free, 09:50
pedroalvarezand seems to be working ok now :/09:50
persiaThen the problem is not likely to be about filesystem extents :(09:50
pedroalvarezworth noting that this system has glibc-2.20 and systemd 21709:50
pedroalvarezis the only one in "production" with those09:51
pedroalvarezwell, both Masons09:51
persiaWhich kernel?09:52
* pedroalvarez looks at the other mason and finds surprises09:52
pedroalvarezyeah, 3.1509:53
pedroalvareznote, the machine of the second log, hasn't panicked09:53
persiaThat's a lot f segfaults.09:55
*** Krin [] has joined #baserock09:59
radiofreelooks like i was too late to the qt5 reviewing party10:16
radiofreeif we want to use this for wayland we'll have to create qt5-tools-wayland, qt5-tools-qtmultimedia-wayland... qt5-tools-qtwebkit-wayland...10:16
*** locallycompact [] has joined #baserock10:17
*** locallycompact [] has quit [Read error: Connection reset by peer]10:18
pedroalvarezreagarding the sefault, nothing important10:19
*** locallycompact [] has joined #baserock10:19
persiaEverything I can find for that class of error seems to be internal kernel functions spinning10:19
pedroalvarezlooks like we are having a ssh dictionary attack, and everytime that shadow tries to get information of a non-existent user, the segfault happens10:20
persiaSolutions seem to involve 1) disabling NUMA, 2) disabling SMP, 3)  disabling the offending kernel function (when one finds it).10:20
persiaPerhaps it is time to add fail2ban to the Mason systems?10:20
pedroalvarezs/Mason/the/ ?10:21
persiaWell, anything public-facing.10:22
persiaI don't think it belongs in minimal or GENIVI10:22
pedroalvarezI'd put it on network-security stratum10:23
persiaIt's not precisely related to either of the things already there, but that wouldn't be bad.10:25
* persia wants hierarchical nested component definitions, and decides to respond again to the ybd thread10:26
pedroalvarezI've never used it. Is something that requires configuration to start working? or can it have a default configuration for ssh attacks>10:26
persiaLooking at Ubuntu, it appears it can ship with a default configuration, and there are samples for many attacks.10:27
persiaTake a look at
DavePageYeah, fail2ban is inelegant but seems to work :)10:29
pedroalvarezIt has been in my things-to-do plate for months, but I wasn't to worried about it since I don't use passwords :)10:32
jjardonSshguard is another possibility10:35
radiofreedon't allow people to login with a password?10:35
pedroalvarezradiofree: that is my approach10:36
*** franred [] has joined #baserock10:38
*** franred [] has quit [Client Quit]10:38
radiofreehmmm having to build a whole lot of stuff again when testing javiers weston branch10:39
*** franred [] has joined #baserock10:39
radiofreei built it last night though10:39
radiofree(with an older version of morph)10:39
radiofreei take it isn't up-to-date then?10:40
pedroalvarezradiofree: take it as: only have cache for devel systems10:42
radiofreewell i built this exact system last night10:43
persiaDid you update definitions since?10:43
radiofreei've upgraded morph and switched branches10:43
radiofreepersia: no10:43
pedroalvarezhm.. maybe is misbehaving?10:43
persiaUpgrading morph can change cache keys (we should probably test that more aggressively).  Switching branches can change definitions causing cache to be invalid.10:44
radiofreewell i switched back to the branch i want to test10:44
persiapedroalvarez: For a build last night, the cache should be local, no?10:44
* radiofree sighs and just builds loads of stuff again10:45
pedroalvarezpersia: yeah, you are right10:45
pedroalvarezradiofree: OOI, what is it building now?10:45
persiaDo we have any data on the load on the distbuild networks that Mason is driving?10:45
radiofreei'm building javier's weston system 10:45
pedroalvarezradiofree: but what chunk10:46
radiofreeit wants to rebuild from bsp-jetson10:46
pedroalvarezradiofree: hm.. anyway, nothing depends on the bsp, so maybe you are lucky10:46
radiofreeit built the bsp, now it's on nss10:46
radiofreewhich is in network-security10:47
radiofreei'll remove llvm from this system actually10:47
radiofreesince it's not needed on arm10:47
radiofreeoh btw, i was thinking about sending a patch for the llvm chunk10:48
radiofreeif it's on arm, just return "true" for everything, to skip the build10:48
radiofreesince mesa depends on it10:48
pedroalvarezsounds hacky10:48
radiofreewell, the mesa strata depends on it10:48
radiofreeyeah it's hacky, but better then spending hours building something that isn't used at all10:49
richard_mawuntil we have parameterised morphologies, that's the hack we need10:49
radiofreeor do we want to go back to mesa-arm, mesa-x, mesa-wayland....10:49
radiofreei'll think i'll try this out now then!10:49
pedroalvarezradiofree: btw, network-security stratum is not in the devel system10:50
radiofreepedroalvarez: i know, i copied javiers weston branch10:54
radiofreeoh weird... it's not building at all now10:55
*** cosm [] has quit [Ping timeout: 265 seconds]10:59
jonathanmawlooking at, devel-systems were replaced with build-systems. Is there any difference other than the name?11:00
*** wdutch [] has quit [Ping timeout: 265 seconds]11:02
persiajonathanmaw: the "build" systems strive to be just enough to run morph, whereas the "devel" systems contain a bit more, making it a slightly more pleasing environment in which to do development.11:03
radiofreeso apparently it failed last night as well11:06
jonathanmawpersia: Will baserock be delivering more devel systems, or is baserock-14.40 the last time a devel system will be delivered?11:06
radiofreejjardon: network manager doesn't build for me11:06
jonathanmaws/will be/was/11:06
radiofreejjardon: 2014-11-26 10:53:01 ERROR OUTPUT FROM FAILED BUILD: No package 'libsystemd-login' found11:06
rdalesystemd has changed and doesn't have lots of .pc files anymore11:07
jjardonradiofree: you have to build master11:07
*** wdutch [] has joined #baserock11:07
jjardonAnd also add readline >= 611:07
pedroalvarezjonathanmaw: we are delivering now the build system which is almost the same as the devel system was11:08
persiajonathanmaw: In my ideal world, there would be a daily snapshot available.  Currently, Mason doesn't push to when it successfully builds a devel system.  I'd +1 a patch that fixed that, but suspect it to be better to wait for the next generation Mason (coming RSN).11:08
jjardonSome things changed since I sent the branch, that's why I said yesterday I will rebase and send again ;)11:08
persiaBut yes, as pedroalvarez says, if you want devel, you would do better to download build and build devel (possibly tweaking to your preferences).11:09
pedroalvarezyeah, upgrading your build system to a devel sytem can take around 20 minutes (depending on your internet connection)11:10
radiofreejjardon: do *i* have to add readline, or is it your new branch?11:10
radiofreealso it is mostly master,   "strata/x-generic.morph: Upgrade to latest xserver (1.16.2)" is what i cherry-picked your previous branch onto11:11
*** cosm [] has joined #baserock11:11
jjardonradiofree: I meant you need to build network manager master11:12
jjardonAnd for that you need readline611:12
jjardonThe current one in base rock is too old11:13
radiofreeah, i see you removed it from your weston system in master11:13
radiofreesorry, in your rebased branch11:13
SotKIn which repository should the turbo-hipster plugins I'm writing for Mason live? The system-tests repo or somewhere different?11:18
persiaThere's a related architectural question: do you imagine these being useful for a wide range of system tests, or are they tightly tied to test implementations?11:20
persiaIf the former, they should be in a separate repo, so that users can point at an arbitrary repo to find tests to apply when gating.11:20
persiaIf the latter, putting them in system-tests serves as a good example, and users can derive from system-tests to generate their own system acceptance tests.11:20
SotKthey are just implementations, so I'll put them in system-tests11:22
persiaThat makes sense.  So users with different systems can examine system-tests, and either branch it, or create their own using that as a model.11:23
radiofreejjardon: linux-pam fails11:44
radiofreepam_unix_acct.c:106:19: error: 'RLIMIT_NOFILE' undeclared (first use in this function)11:44
jjardonradiofree: is not the same version we are using in weston-genivi?11:47
radiofreejjardon: no11:48
radiofreeweston-genivi is using 1.1.8, you're using 1.1.511:49
radiofreei'll change the sha and soldier on11:51
jjardonOh, ok, as i told you i was rebasing the branch, didnt have time to compile it yet, so thanks for the pretesting ;)11:52
jjardonradiofree: yeah, same compilation error here11:55
radiofreejjardon: and congratulations, the whole point of your patch series doesn't build either :D11:58
radiofreeconfigure: error: Package requirements (xcb xcb-xfixes xcb-composite xcursor cairo-xcb) were not met:11:58
radiofreei think you need --disable-xwayland --disable-x11-compositor11:58
jjardonradiofree: its a working branch, its not finished11:59
radiofreeit builds when you disable xwayland, x11-compositor anyway12:07
radiofreeyou'll probably want to disable-x11-compositor?12:07
radiofreehmm my lorries don't seem to have lorried12:11
radiofreedid i do something wrong?12:11
richard_mawradiofree: I'll have a poke, it usually means there's a syntax error in the lorry file12:12
radiofreerichard_maw: spotted it12:13
radiofree+ "nvidia/tegra-pinmux-scripts.git" {12:13
radiofreesorry about that, should i fix it in master?12:13
richard_mawyeah, go ahead and push the fixup directly12:13
radiofreeso cryptopp, if there's a requirement to have the jetson flashing scripts work in a devel system, will need to be added to a devel system12:20
radiofreealthough i could git clone it and build as part of the script12:20
richard_mawis there any reason we can't build the jetson flashing scripts in morph too?12:20
radiofreesure, it requires some of those nvidia tools12:21
radiofreebut then you'd have to add "jetson-baserock-flash" stratum to a devel system12:22
radiofreewhich i don't think paulsher1ood wanted12:22
paulsher1oodi'm not against it, really?12:22
radiofreei don't know, are you?12:22
paulsher1oodi think we could make it more generic - flashing-tools? and start with jetson stuff?12:23
radiofreewell in theory i think my script would work on a wandboard as well12:23
radiofreeor any arm board with usb gadget support in u-boot12:24
richard_mawthere's potentially room for fastboot and heimdall in flashing-tools too12:24
paulsher1ood+1 for this idea, then12:25
radiofreeok, flashing tools it is then, i'll make my life easier...12:25
radiofreeif you try to use it on a desktop machine i'll just say "you need to build foo from bar first"12:25
paulsher1oodradiofree: i was thinking about flashing-tools as stratum name... you can keep this script jetson specific if you prefer12:26
radiofreei will, i'll put it in a flashing-tools stratum though12:26
*** mariaderidder [] has quit [Ping timeout: 240 seconds]12:59
*** mariaderidder [] has joined #baserock13:15
*** jjardon [sid723@gateway/web/] has quit [Ping timeout: 258 seconds]14:14
*** jjardon [sid723@gateway/web/] has joined #baserock14:15
*** radiofree [radiofree@unaffiliated/radiofree] has quit [Ping timeout: 258 seconds]14:20
*** radiofree [radiofree@unaffiliated/radiofree] has joined #baserock14:20
radiofreeturns out new version of cryptopp are a nightmare to compile14:31
radiofreei think it's working now, but i had to grab stuff from
radiofreespecifically the libcrypto++_5.6.1-6.debian.tar.gz changes14:31
radiofreewhat's the convention in baserock for including these changes? just make a branch?14:32
franredradiofree, yes, the convention is to create a branch with the patches on it14:35
franredindicating from where they come from14:35
radiofreedidn't work that time either14:45
*** mariaderidder [] has quit [Ping timeout: 245 seconds]14:50
*** mariaderidder [] has joined #baserock15:05
*** genii [~quassel@ubuntu/member/genii] has joined #baserock15:05
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]16:05
radiofreecancel that hmm16:09
franredradiofree, it has been canceled16:12
radiofreecan someone with a baserock x86 vm please run a script for me, and give me the output?16:15
radiofreeoutput is a file16:15
pedroalvarezradiofree: yup16:15
radiofreeok pedroalvarez, will e-mail you!16:17
radiofreepedroalvarez: biff16:20
pedroalvarezradiofree: done16:25
radiofreeyou're a star!16:30
radiofreelooks like it's working on x8616:30
pedroalvareznp :)16:33
radiofreei don't suppose you have a 32-bit system around?16:52
*** sherm_ [] has quit [Quit: Leaving]17:16
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock17:27
*** jonathanmaw [] has quit [Quit: Leaving]17:34
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]17:36
*** zoli_ [] has joined #baserock17:51
*** zoli_ [] has quit [Changing host]17:51
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock17:51
*** Krin [] has quit [Quit: Leaving]17:52
*** locallycompact [] has quit [Ping timeout: 265 seconds]17:55
*** mariaderidder [] has quit [Quit: Ex-Chat]17:59
*** bashrc [] has quit [Quit: Lost terminal]17:59
*** petefoth [] has quit [Quit: petefoth]18:10
*** tiagogomes [] has quit [Quit: Leaving]18:10
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]21:31
*** genii [~quassel@ubuntu/member/genii] has quit [Remote host closed the connection]23:22

Generated by 2.15.3 by Marius Gedminas - find it at!