=== h2o64_ is now known as h2o64 === \b is now known as benonsoftware === plars_ is now known as plars === wavis_ is now known as wavis === tgm4883_ is now known as tgm4883 === Ursinha_ is now known as Ursinha === pitti is now known as Guest47071 === elijah is now known as Guest16052 === XDS2010_ is now known as XDS2010 === Azelphur_ is now known as Azelphur [02:31] mariogrip: still working on that OPO port? [02:45] I just wanted to let you know that this link on the porting guide is broke http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-utopic.git;a=shortlog;h=refs/heads/mako === ajalkane_ is now known as ajalkane === pipedrea1 is now known as pipedream === dkessel_ is now known as dkessel === lan3y is now known as Laney === Laney is now known as Guest38885 === Guest38885 is now known as Laney [11:18] ogra_: Ping [11:18] SturmFlut, yes ? [11:18] ogra_: Seriously, do you even sleep? [11:18] at the end of the month usually :) [11:19] now that's commitment! [11:19] (seen G+ ??) [11:20] * ogra_ just stumbled over the krillin on the mediamarkt page ... [11:20] ogra_: I would ask cking directly, but he isn't online, so I'm looking for somebody who knows a bit about our phone kernels in regard to /proc/timer_stats. The arale kernel tries to lie to me about a kernel thread having PID 1, which makes the question the content of the whole file [11:20] s/makes the/makes me/ [11:20] the source is on github, thats all i know ... (i never built it ... ) [11:21] and they are build the "android way" with special MTK scripts and all [11:21] I heavily suspect this Android stuff broke something fundamental again [11:22] yeah [11:22] looks at your system loda in top ;) [11:22] *look [11:22] *load [11:22] geez ... [11:23] (it only watches one core i think) [11:23] ogra_: Hm, that load is a bit low [11:24] Seriously, a kernel thread with PID 1, what comes next? Ubuntu fridges? [11:25] flying ubuntu fridges with lasers ... convergence, y'know [11:25] So Snappy Skynet IS real [11:25] ssshhhh ! [11:26] the converged phone might be delayed though ... bq still hasnt found a way to integrate the fridge in the case [11:27] Hmmmmmmmmm, the error even shows up in an example in the in-kernel documentation for /proc/timer_stats [11:27] hah [11:37] But it seems to be correct on krillin and my 15.04 laptop [11:37] are you sure you are not seeing the android container btw ? [11:38] not sure how lxc's call of the android init will show up [11:39] Hmmmm [11:41] The Android init inside the container shows up as a normal process in the "lxc-start -n android -F -- /init" hierarchy, on my device it is currently PID 879 [11:44] Type to exit the console, to enter Ctrl+a itself [11:44] root@arale:/ # ps [11:44] USER PID PPID VSIZE RSS WCHAN PC NAME [11:44] root 1 0 716 156 c0153934 0001c328 S /init [11:44] root 3 1 580 76 c0153934 0001c328 S /sbin/ueventd [11:44] inside the container it has PID1 [11:45] I would expect that, things break if it doesn't [11:45] right === boud_ is now known as boud [11:50] * SturmFlut has to switch trains [12:40] ogra_: okay, I've asked on #kernelnewbies and nobody seems to be immediately sure what's happening === LarrySteeze|Away is now known as LarrySteeze === BOHverkill_ is now known as BOHverkill [14:09] Hmmm, I rebooted my arale, started a QML game, Telegram and the browser and memory usage is at 1.2 GB. But userspace only accounts for about 700 MB. Are there any drivers that allocate large chunks of memory, e.g. the GPU? [14:15] ogra_: ^^ [14:59] I wonder if webapps will boom with web assembly === dobey_ is now known as dobey [16:29] ogra_: Have you ever looked at memory management on arale and krillin? On krillin everything seems to add up fine, but on arale reported memory consumption is about 500 MB higher than what the userspace actually needs [16:30] and I'm not sure if this is just a change in reporting or if something in the kernel actually claims half a gigabyte of memory [16:31] what numbers do you compare? [16:31] or .. sum.. [16:33] anpok_: cking's smemstat outputs the proportional set size for every process, the sum of that should be what all userspace processes are actually using [16:33] (if I'm not mistaken, which would explain some things) [16:38] if I reboot krillin, unlock the phone and immediately run "free", it reports 395 MB used (without buffers and caches) and smemstat outputs a sum of 380 MB. That sounds quite reasonable. [16:41] SturmFlut, sounds like a bug ... and no, i havent looked deeper, my bug about the load avg. was set to low prio by the guys working on the device tarball for the arale [16:41] ogra_: Hmmmm [16:41] file one ;) [16:42] To give you the numbers for arale: after a reboot "free" gives me 552 MB, but smemstat just 292 MB [16:42] the browser and webapps surely behave a bit weird (bein OOMed earlier than they should) [16:42] *being [16:44] ogra_: The whole memory management feels strange, to be honest. Sometimes I switch from Telegram to the browser, do something, switch back, and Telegram was OOMed and has to be restored. On arale. A device with 2 GB of RAM. Even more you take the zram swap into account [16:44] *if you [16:44] yep [16:45] as i said, file a bug (against canonical-system-image for a start) with your findings [16:46] is it "normal" that arale r1 kills my adb sessions all the time? [16:46] no, but it is normal that the USB works very badly ... watch your dmesg on the PC [16:46] ogra_: I'll dig a bit deeper before filing it, it may be some kind of reporting change with kernel 3.10 [16:47] might be, but it is a mediatek device with a mediatek kernel ... so it wouldnt be surprised at all [16:47] *I [16:47] hehe === Stskeepz is now known as Stskeeps [18:20] Is any one currently in here? [18:51] any developers I have an idea [18:52] Rugby worldcup coming in the next few months [18:52] need apps [18:52] lol === gcollura_ is now known as gcollura [18:58] you have few months to learn how to code :P [18:58] arhhhhh lol [19:27] hi! how can i export the device SMS database? [19:30] nvm, i think i found the database [20:44] any new ota on July ? thanks [20:46] will we see tethering soon? thanks again === sergiusens_ is now known as sergiusens