IRC logs for #baserock for Monday, 2014-08-18

*** fay_ [] has joined #baserock06:45
*** franred [] has joined #baserock06:49
*** radiofree [radiofree@unaffiliated/radiofree] has quit [Remote host closed the connection]07:52
*** radiofree [radiofree@unaffiliated/radiofree] has joined #baserock07:54
*** jonathanmaw [] has joined #baserock08:21
*** ssam2 [] has joined #baserock08:39
*** franred [] has quit [Remote host closed the connection]08:43
*** franred [] has joined #baserock08:51
straycatssam2, As I understand it you can't bind two processes to the same port, I don't know of any way to have multiple services running on the same port.09:05
ssam2oh, that's a shame09:06
ssam2I wonder if there's anything else we can do to make it possible to use Morph on public wifi, then09:06
ssam2this limitation made it impossible to actually give a demo when I gave a talk about Baserock at the weekend09:06
*** franred [] has quit [Ping timeout: 272 seconds]09:07
ssam2I wonder if one lighttpd process could route requests to both cgit.cgi and morph-cache-server as appropriate09:08
straycatI guess we could do, that's what we do with gitano on http.09:13
straycatAhh except that the git stuff runs as the "git" user and the cache stuff runs as the "cache" user.09:19
persiaRouting different URIs to different engines from the same server on the same port is fairly normal.09:20
persiaOne way to get around the multiuser issue is to have the server instance proxy to some prefix, which ought just be a line of config somewhere09:21
*** franred [] has joined #baserock09:21
persia(if not, we can move cgit to 7080, and run any of the many lightweight frontend web servers to proxy both services at different prefices.09:21
KinnisonWe could go to a proper httpd which can handle multiple users, or as persia says, we can run a bundle of servers against localhost and then reverse-proxy from :8009:23
KinnisonCherokee is pretty easy to get to build09:23
KinnisonBut keeping the current separation might be better09:24
persiaThe aforementioned "proper httpd" does the same thing, just tries to hide it from users (and tends to use unix sockets rather than real ports)09:24
KinnisonCan use TCP too09:24
KinnisonCherokee is pretty flexible09:24
KinnisonNot that I'm an upstream on it09:24
Kinnison[* may not be true]09:25
KinnisonAnd I definitely didn't do the work to allow multiple uid/gid support while still dropping privs on the main process, oh no09:25
* persia doesn't really see the point of claiming to be a monolithic server than then proxying to oneself over TCP, but leaves that sort of architectural decision to proper upstreams of the relevant servers09:26
KinnisonHigh performance across a wide range of platforms often comes at the expense of architectural elegance :-(09:28
*** franred [] has quit [Ping timeout: 260 seconds]10:05
*** franred [] has joined #baserock10:48
paulsherwoodrichard_maw: Did folks agree with my suggestion to merge franred's script on its own, then run it on master, and merge the result?11:26
* paulsherwood is excited :)11:28
radiofreethis 3.16 built the kernel in 27 minutes, 17 minutes slower than it should11:49
radiofreequite a lot of oops messages (relating to kernel) in the dmesg log as well11:50
paulsherwoodthat's not lovely11:50
radiofreethankfully since i upgraded it using baserock i can just reboot on 3.10 to build the artifacts11:51
paulsherwoodi've been building with 3.15 on x86 with no problems... but maybe i should try 3.1611:51
paulsherwooddo you think what you're seeing is jetson specific?11:52
radiofreepaulsherwood: this is a jetson issue11:52
* paulsherwood must try upgrading soon11:53
paulsherwoodhas anyone done any work to make for minimal size devel system? eg low-hanging-fruit stratum-splitting magic12:30
radiofreei'm hoping this fixes the kernel issue12:31
radiofreethe CPU_FREQ stuff doesn't appear to be very well implemented12:31
radiofreeso it was defaulting to ondemand (so a low clock speed) then crapping out when you try and do anything with it (i.e increase it), since the cpu0/cpufreq/ stuff wasn't being exported12:32
radiofreeforcing it to PERFORMANCE looks ok, but i'll know in about 7 minutes12:32
paulsherwoodwhere are you setting that, radiofree ?12:33
radiofreein the kernel config12:33
radiofreedoesn't look good12:46
paulsherwoodoh dear12:56
radiofreemove back to 3.1012:56
radiofreebut then docker might not work?12:58
paulsherwoodprobably not13:01
paulsherwoodradiofree: are you using the configs i added for docker? if so, maybe they are the problem?13:02
ssam2paulsherwood: haven't done any work for a minimal size dev system, would love to13:04
ssam2there's some issues with splitting that need resolving though, like the one I mailed the list about last week13:04
radiofreepaulsherwood: not that13:09
radiofreeit's the 3.16 kernel i'm building with the baserock config, no docker stuff in it13:10
radiofreecpu isn't maxed out13:10
paulsherwoodradiofree: so are you saying that 3.16 works nicely for graphics, but has issues for compile/build?13:18
cyndisare you using 3.16 on jetson?13:34
pedroalvarezcyndis: I believe he is13:35
cyndisthen it is no wonder that cpufreq isn't working :)13:36
pedroalvarezradiofree: ^^13:36
cyndissince the series that adds it isn't yet even in -next, just on the ML13:36
pedroalvarezthat's good to know, thanks13:39
cyndisif you're interested in tegra support in the upstream kernel, please join #tegra. most of the devs are there.13:40
radiofreecyndis: thanks :)13:41
cyndis:) great to see that people are actually using it!13:42
cyndislots of patches in the pipeline at the moment13:42
paulsherwoodcyndis:  aha! :)13:44
paulsherwoodradiofree: ^^13:45
*** bjdooks_ [] has joined #baserock14:06
pedroalvarezDo we have an example of morphology of max-jobs?15:37
*** ssam2 [] has quit [Ping timeout: 246 seconds]15:41
radiofreepedroalvarez: qt-webkit15:50
radiofreeany chance on getting another +1 on my tegra-3.10 merge?15:51
radiofreepaulsherwood: took me 9 minutes to flash using usb passthrough in qemu16:03
radiofreepaulsherwood: also your board is now working again16:03
*** jonathanmaw [] has quit [Quit: Leaving]16:32
paulsherwoodradiofree: w00t! :)16:39
richard_mawradiofree: have a +116:41
*** ssam2 [] has joined #baserock16:41
radiofreemore patches sent!16:47
richard_mawradiofree: I would have preferred the commit message say why we need to revert, but I overheard enough conversation to have a sufficient idea of the problem16:51
pedroalvarezradiofree: thanks for the example (max-jobs)16:53
pedroalvarezSeems like I need to set max-jobs to  in audiomanager16:54
richard_mawradiofree: +2, on the condition that you update the commit in your definitions change from 123cc68c33ff537748a244f4932640083b4b6825 to de3664efca390a8d2fe525a583d17772deee8500 (the merge commit)16:54
richard_mawlet me know when it's done, so I can include it in the morphology rearrangement patch16:57
paulsherwoodrichard_maw: do you mean the chunks-in-definitions rearrange? is that going live today?16:59
richard_mawthat's the plan16:59
radiofreerichard_maw: merged16:59
radiofreeyes merged16:59
* paulsherwood can hardly wait16:59
pedroalvarezI wonder if I can get a quick review on this:
pedroalvarezor should I send it to the ML?17:00
pedroalvarezcurrently audiomanager doesn't build on a Jetson board17:01
richard_mawpedroalvarez: give it a few minutes, and include the audiomanager morphology change in definitions.git17:01
pedroalvarezoh! good to know!17:01
paulsherwoodradiofree: you missed my comment on your comment :)17:02
richard_mawpedroalvarez: is audiomanager running out of memory, or is the build-system failing to parallelise properly?17:02
richard_mawthe former is minor grumbling, the latter brings shame upon audiomanager's build system17:02
pedroalvarezrichard_maw: I believe it's running out of memory17:03
pedroalvarezOh, I managed to build it, it doesn't happens always :)17:03
paulsherwoodpedroalvarez: well +117:05
paulsherwood(better to fix it, avoid tripping up someone else in future)17:06
*** fay_ [] has quit [Remote host closed the connection]17:07
richard_maw is live17:07
paulsherwoodwhoa!!! :)17:08
pedroalvarez           )17:08
pedroalvarez          (.)17:08
pedroalvarez          .|.17:08
pedroalvarez          l8J17:08
pedroalvarez          | |17:08
pedroalvarez      _.--| |--._17:08
pedroalvarez   .-';  ;`-'& ; `&.17:08
pedroalvarez  & &  ;  &   ; ;   \17:08
pedroalvarez  \      ;    &   &_/17:08
pedroalvarez   F"""---...---"""J17:08
pedroalvarez   | | | | | | | | |17:08
pedroalvarez   J | | | | | | | F17:08
pedroalvarez    `---.|.|.|.---'17:08
* rjek wondered what that was going to be at first.17:09
pedroalvareza cake?17:09
* paulsherwood manages to run morph build post 2b10a9677817377618e33d2b07ab7caefe31663317:13
* pedroalvarez had an ERROR, but he forgot to upgrade his morph first17:15
pedroalvarezno rebuild needed after 2b10a9677817377618e33d2b07ab7caefe31663317:20
pedroalvarezamazing stuff! thanks!17:20
paulsherwood2014-08-18 17:24:21 [Build 484/554] [linux-x86-64-generic-devel] Starting actual build: linux-x86-64-generic-devel 7d1311b17:24
* paulsherwood likes that v3.17-rc1 => 7d1311b17:25
paulsherwoodrichard_maw: maybe there should be an announcement about this? i should probably make a couple of new tutorials too :)17:27
straycatneat :)17:29
*** ssam2 [] has quit [Quit: Leaving]17:44
petefothrichard_maw: and franred: belated congratulations - great stuff18:31
*** flatmush [] has quit [Ping timeout: 240 seconds]19:04
*** franred [] has quit [Ping timeout: 260 seconds]19:05
*** flatmush [] has joined #baserock19:05
*** franred [] has joined #baserock19:07
*** franred [] has quit [Ping timeout: 240 seconds]19:12
*** franred [] has joined #baserock19:13
*** franred [] has quit [Ping timeout: 255 seconds]19:21
*** flatmush [] has quit [Ping timeout: 272 seconds]19:23
*** flatmush [] has joined #baserock19:32
*** flatmush [] has quit [Ping timeout: 245 seconds]19:37
*** flatmush [] has joined #baserock19:40
*** flatmush [] has quit [Ping timeout: 260 seconds]19:45
*** flatmush [] has joined #baserock19:49
*** franred [] has joined #baserock19:51
*** flatmush [] has quit [Ping timeout: 250 seconds]19:56
*** franred [] has quit [Ping timeout: 255 seconds]19:57
*** franred [] has joined #baserock20:00
*** flatmush [] has joined #baserock20:05
*** flatmush [] has quit [Ping timeout: 255 seconds]20:10
*** franred [] has quit [Ping timeout: 245 seconds]20:10
*** flatmush [] has joined #baserock20:27
*** flatmush [] has quit [Ping timeout: 240 seconds]20:33
*** flatmush [] has joined #baserock20:40
*** flatmush [] has quit [Ping timeout: 244 seconds]20:46
*** flatmush [] has joined #baserock20:52
*** flatmush [] has quit [Ping timeout: 255 seconds]20:56
*** flatmush [] has joined #baserock20:57
*** flatmush [] has quit [Ping timeout: 264 seconds]21:02
*** flatmush [] has joined #baserock21:13
*** flatmush [] has quit [Ping timeout: 246 seconds]21:17
*** flatmush [] has joined #baserock21:18
*** flatmush [] has quit [Ping timeout: 255 seconds]21:23
*** flatmush [] has joined #baserock21:27
*** flatmush [] has quit [Ping timeout: 260 seconds]21:36
*** De|ta_ [~arc@] has joined #baserock22:05
*** De|ta [~arc@] has quit [Ping timeout: 246 seconds]22:10
*** ratmice__ [] has quit [Ping timeout: 246 seconds]22:10
*** ratmice__ [] has joined #baserock22:11
*** flatmush [] has joined #baserock22:14
*** flatmush [] has quit [Ping timeout: 260 seconds]22:19
*** flatmush [] has joined #baserock22:29
*** flatmush [] has quit [Ping timeout: 245 seconds]22:38
*** flatmush [] has joined #baserock22:40
*** flatmush [] has quit [Ping timeout: 245 seconds]22:44
*** flatmush [] has joined #baserock22:48
*** flatmush [] has quit [Ping timeout: 240 seconds]23:21
*** flatmush [] has joined #baserock23:24
*** flatmush [] has quit [Ping timeout: 264 seconds]23:32
*** flatmush [] has joined #baserock23:34
*** flatmush [] has quit [Ping timeout: 255 seconds]23:51
*** flatmush [] has joined #baserock23:52

Generated by 2.15.3 by Marius Gedminas - find it at!