=== zz_chihchun is now known as chihchun === nashpa_ is now known as nashpa [07:58] good morning [07:59] <[mbm]> yep. [08:16] [A [08:16] morning [10:52] I have "Linux ubuntu 2 3.2.0-23-omap #36-Ubuntu Tue Apr 10 20:24:21 UTC 2012 armv7l armv7l armv7l GNU/Linux", I am following the instruction here http://ubuntuforums.org/showthread.php?t=263136 and I am typing the wpa_supplicant command and get this output http://paste.ubuntu.com/1357684/ [10:52] does anyone know that if I should anything to use wpa_supplicant on my board? [10:53] *I should download anything [11:09] angs, well apt-get install wpasupplicant [11:10] brendand: thanks, I was trying wpa_supplicant to install it:) [11:17] angs, seriously, thats nearly 7 years old ... [11:17] angs, http://wiki.debian.org/WiFi/HowToUse#WPA-PSK_and_WPA2-PSK [11:17] use that one, wpa got lors easier over the years [11:17] *lots [11:18] ogra-cb: I have been using linux for 5 months, thank you for the documentation :) === chihchun is now known as zz_chihchun === XorA|gone is now known as XorA [12:02] Who woulda thought dist-upgrade would have been a bad idea on my chromebook ;_; [12:03] it isnt on mine [12:03] * ogra-cb_ runs raring happily over here === ogra-cb_ is now known as ogra-cb [12:05] my chromebook is fine with raring [12:05] how did you guys get to raring? [12:05] orik: I did precise -> raring in one run [12:05] i used the proper upgrade methid via do-release-upgrade though [12:05] ahhhh [12:05] apt-get update;apt-get dist-upgrade for me [12:05] thats what i did for quantal to raring [12:06] I was doing upgrade-manager -d [12:06] and I was going to stop at 12.10 inbetween [12:06] precise quantal i did via do-release-upgrade [12:06] but never booted after 12.10 [12:06] yeah, update-manager is the same as do-release-upgrade [12:06] I want to modify Jay Lee's script to use my own tarball but I can't find the rootfs I want >.< [12:07] oh, i know what it is, you need to disable the plymouth jobs in /etc/init/ [12:07] hrw: will that take me straight to 13.04? I'm going to try again here shortly [12:07] reboot into chromes developer mode, mount the ubuntu partition and put .override files in place for all plymouth jobs you find in /etc/init/ [12:08] the console=none option the bootloader sets makes plymouth hardlock the device on boot [12:08] * ogra-cb just remembered he had to do that [12:09] ogra-cb; man I wish I understood half of what you just said [12:10] so I boot in dev mode, go to chrosh and then into shell, mount the ubuntu partition, and then I do *something* to *somefiles* in /etc/init/ [12:10] echo manual|sudo tee -a /etc/init/plymouth.override [12:10] thats what you need to do [12:11] man that makes me wish I hadn't already blown ubuntu off this thing xD [12:11] or just echo "manual" >/ubuntu-mountpoint/etc/init/plymouth.override [12:11] not sure if you are actually root in the chros shell [12:12] you can always also just press ctrl-alt-f2 to get into shell [12:12] do the same for every file you find in /ubuntu-mountpint/etc/init/ whith plymouth in their name [12:13] I'm glad this channel is logged so I can check back later xD [12:13] ogra-cb: thx [12:13] ogra-cb: you know that you can change kernel cmdline too? [12:13] hrw, console_none is set by the bootloader [12:13] *console=none [12:14] ogra-cb: and then 'console=tty1' by kernel [12:14] I have kernel messages on my chromebook [12:14] err, its actually console= [12:14] right, plymouth picks the first console- option though [12:14] and cant haldle an empty value for it [12:15] the kernel is fine here, its plymouths fault trying to attach to a non existing tty [12:15] which actually locks yous device hard [12:16] *your [12:17] hrw; did you have to go through this or did you have another work around? [12:19] or is this issue have to do with trying to boot 12.10 [12:19] I think I had to handle plymouth in some way [12:21] ogra@chromebook:~/packages$ ls /etc/init/*.override [12:21] /etc/init/plymouth-log.override /etc/init/plymouth-splash.override /etc/init/plymouth-stop.override /etc/init/plymouth-upstart-bridge.override /etc/init/plymouth.override [12:22] :D thank's for putting up with me guys [12:25] later down the road I'm going to play around with the Chrubuntu script and have it put on a debian root fs or something [12:25] I'll be back on in ~2 hours on 12.04 and 2 hours later on 13.04 [12:26] * ogra-cb hopes you mean 12.10 [12:26] ;) [12:27] can I not apt-get update, dist-upgrade to 13.0? [12:27] or just update-manager -d twice? [12:27] 13.04* [12:36] update-manager didnt know about 13.04 last week when i tried it [12:36] and you shouldnt skip releases on upgrading [12:36] we dont test that, it could break heavily [12:39] alright [12:39] sounds good [12:40] so, I'm writing out the steps I'm going to take before I disconnect here [12:41] what's the syntax for the tee command? [12:41] the only examples I find have to do with the output of another command [12:42] so, if I wanted to rename /dev/mmcblk0/etc/init/plymouth as /dev/mmcblk0/etc/init/plymouth.overide [12:43] would mv work? [12:47] time to go break things; thansk for the help again [12:47] hrw pulled off precise to raring so I'll try it, and if I break things too bad I can try again slower [14:25] so hrw; is xserver 13 going to run at all before I patch my driver after i update to 13.04? [14:36] orik: you compiled chromium armsoc driver under precise? [14:37] not yet; let me look into that [14:37] I just finished testing glmark2 [14:37] got a 5! [14:39] 112 iirc was my result from glmark2-es2 [14:40] I'm grabbing es2 and running it again, I want a good before and after picture [14:43] are there any large performance increases from 12.04 to 13.04 on ARM bases systems? I figured there would be with the push to get the Nexus 7 running [14:43] but if I can just patch this driver in 12.04 I might leave it at that [14:44] 13.04 is still very young [14:44] the release will surely see a lot of preformance improvements [14:44] I like to romantacize myself as a "Bleeding Edge" sorta guy but deep down I'm not [14:45] through glmark2-es2 I got 4 [14:45] well, with 13.04 ubuntu follows a new dev model that should result in a lot less breakage during development releases [14:45] so it shouldnt be to harmfule to live on the edge anymore [14:46] thats good to hear [14:46] btw, jut mv'ing the upstart jobs wont help [14:46] the .override file needs to only contain one line: manual [14:46] Oh [14:47] so the... whatever it is [14:47] looks for plymouth [14:47] .override tells upstart "use this file instead of the job" [14:47] goes Hey, There's an overide, lets use that instead [14:47] very cool. [14:47] if you just mv the .override is exactly the same as the job file was [14:48] I was under the impression that you were renaming it to .override as to make it so the original file would not be found [14:48] sorta like a .ini.backup or something [14:48] so> echo "manual"| sudo tee -a /path/to/override/file [14:48] mmmm that makes more sense [14:49] hrw: do you think it would make more sense for me to upgrade to 13.04 then use your .deb for the patch or to manually patch it first [14:50] slangasek, i was actually taking a brief look at the plymouth issue and came up with something like http://paste.ubuntu.com/1357935/ ... not sure how accurate that is, my C is a bit rusty (it should avoid using console=none and an empty console=) [14:50] orik: any way you like [14:50] I'm going to try to patch it first so I can say I learned something [14:51] rather than just taking all the scraps off the table :P [14:51] orik, alternatively send beer to hrw to make him upload the package to 13.04 [14:51] orik: package cgpt and kernel signing tools [14:51] get rid of that ugly start screen? [14:51] ogra-cb: need to add /dev/mali stuff and number fix [14:51] that would be lovely [14:51] yeah [14:52] orik: there are two options for 'ugly start screen' [14:52] 1. flash developer firmware of chromium os [14:52] i dont find the start screen ugly though, i could live with it if i had not to start at it for 30sec every time i boot [14:52] 2. flash u-boot instead of u-boot+vboot combo [14:52] ogra-cb: no one told you about "ctrl-d" shortcut? [14:52] s/atart/stare/ [14:52] I'll consider whoever does that a brave man [14:52] hrw, lol, nope [14:53] ogra-cb: HAHA!!!! [14:53] orik, why? [14:53] ogra-cb: ctrl-u boots from SD [14:53] hrw, thx !! [14:53] you made my day ;) [14:53] and you mine :) [14:53] hahahaha [14:53] * hrw -> food etc [14:53] that was really annoying me, i tried things like esc indeed [14:54] lilstevie: this is my only working computer right now, and I'd hate to see someone brick their machine trying to flash a different bootloader onto it [14:54] but not any modifier key combos [14:54] ogra-cb: read all posts by ojn on his google+ stream [14:54] orik, the first 2MB of spi are ReadOnly, within that 2MB there is a fair bit of "brick recovery" stuff [14:54] I heard these samsung micro ssd's have pretty good throughput so booting off of them won't suck [14:55] orik, by that I mean depending on whether you are trying to boot recovery mode or normal it loads a different u-boot from the SPI [14:55] at least that is what we have discovered [14:55] * ogra-cb waits for really small but really fast USB sticks to show up [14:55] O: [14:55] (USB 3.0 i mean) [14:55] you can't boot from 3.0 [14:55] :'c [14:55] that should be on par with a sata disk [14:56] sure you can [14:56] your kernel cant live on usb [14:56] your rootfs can [14:56] orik, by "you can't boot" you mean the kernel can't :p [14:56] and thats all that counts [14:56] I am still really undecided about the chromebook [14:56] ogra-cb: http://www.codinghorror.com/blog/2012/11/a-ssd-in-your-pocket.html [14:56] its really awersome [14:56] lilstevie: whatcha mean? [14:56] i just with GLES would work fully [14:57] *wish [14:57] orik, I mean, I don't know if I want one or not [14:57] but you can create a ramdisk to preload the kernel first and then boot from that? [14:57] if I were you [14:57] I'd just [14:57] buy one from best buy or someplace with a [14:57] like [14:57] month long return policy [14:57] use it [14:57] if you don't like it [14:57] return it [14:58] orik, i wait for something about 1/10th of that size, but yeah [14:58] orik, you are assuming I am in the states [14:58] something that doesnt stick out of the case so much [14:58] I am, I am :/ [14:58] but if you grab them out of the states [14:59] orik, I would have to import the chromebook [14:59] you get a fucked keyboard [14:59] rumors say you do [14:59] or uh [14:59] lilstevie, how about uk ? [14:59] i ordered mine there [15:00] ogra-cb, tbh both the UK and US are as equally difficult to order from [15:00] land downunder and all that crap [15:00] :p [15:01] oh, yeah [15:01] dont you also need an upside down kbd layout ? :P [15:01] :p [15:02] :P [15:02] hey can I just uh [15:02] update to 12.04 [15:02] i mean 12.10 [15:02] and then start messing with the Plymouth files [15:02] sure [15:02] before I even reboot [15:02] sudo nautilus and throw in new ones [15:02] well, mess with them at anty point you like [15:03] so I can use that easymode graphical interface [15:03] its just that the cb doesnt use an initrd so if you screw up you have to go back into cros and fix it from there [15:04] instead of having a chance to fix it from an initrd prompt [15:04] am I changing things in /etc/init/ or etc/init.d/ [15:06] well if nothing else comes out of my night, I am glad to finally have volume control :D [15:07] orik, /etc/init [15:08] alright, this will be take two of upgrading to 12.10 [15:10] I wish there was a website that collected glmark2 benchmarks, so you could compare systems [15:11] orik: have you tested phoronix? [15:11] do they have more than just news? [15:11] like, do they have a database going? === Quintasan_ is now known as Quintasan [15:11] OH; it's a test suite... [15:11] all their news is generated using http://www.phoronix-test-suite.com/ [15:11] very cool. [15:12] I'll run it after getting graphics drivers [15:12] I feel like running it now would be a bit to painful [15:12] can I upgrade to 12.10 and then to 13.04 without rebooting inbetween? [15:14] orik: linaro is also working on something for their validation runs https://blueprints.launchpad.net/linaro-graphics-dashboard/+spec/linaro-graphics-dashboard-glmark2 [15:16] can't wait [15:16] :D [15:21] infinity, along with the kernel, we also need a firmware package for the nexus7. It's currently in the staging PPA, in case you prefer a review before it enters the NEW queue. [15:21] https://launchpad.net/~ubuntu-nexus7/+archive/staging/?field.series_filter=raring [15:21] ogra-cb, you may want to take a look if you're curious and have time [15:22] infinity, it is separate as it needs to present the license agreement to the user before installing [15:26] janimo, looks fine to me, only FW for eth wlan chip for now ? [15:26] *the [15:26] wlan and BT right? [15:26] well, there is other stuff we might want to add over time [15:26] ogra-cb, the same we had in the patched PPA for 12.10 [15:26] for a start its fine i guess [15:26] sure, we can add more latee [15:27] right [15:57] janimo: it's now finally decided that ubuntu can redistribute the firmware for wifi/bt for the nexus7? [15:58] morphis, it can redistribute I think if it pops up the license agreement as seen in that ppa [15:58] I would not say finally decided about anything involving binary blobs :( [15:59] ok [15:59] but current thinking is we're fine if we show the same text the .sh installers from google show [16:02] janimo: ok, will take this for discussion over to webos-ports :) [16:05] morphis, I honestly do not think Google or any of the partners would care at all if we or webos just distributed without any legalese, but at least Canonical needs to do this properly for image's sake if not else [16:06] with all the GPL infringers putting out binary blobs it would be weird of them to start minding about license issues :) [16:06] janimo: yes, but it's better to try to play well (as much as possible) to not cause any harm later [16:07] s/harm/extra work/ I'd say [16:08] janimo: yeah [16:09] janimo: at least you only have to deal with the wifi/bt bits we need to use the graphic bits too ... [16:10] they are freely distributable [16:11] morphis, we also ship the tegra drivers but those are not firmware [16:11] with aproper license [16:11] ogra_: for the nexus7 they are [16:11] well, for tegra2 and tegra3 [16:11] ogra_: but over at webos-ports with the galaxy nexus we have a completely different situation :) [16:11] morphis, the LICENSE agreement needs to be popped up only for the 5 or so binaries that can be downloaded from Google [16:11] PVR has a usable license as well [16:12] we ship PVR preinstalled on pandaboard images [16:12] ogra_: you want to know the whole story? [16:12] morphis, he does [16:12] surely [16:12] ogra_: PVR preinstalled, really? [16:12] ogra_: ok, let me summarize: in openwebos the central rendering process is the LunaSysMgr; no X11 or something like this [16:13] we are limited to a 3.0.x kernel [16:13] recent PVR drivers like you ship with ubuntu for the pandoboard needs omapdrm [16:13] 3.0.x does not have omapdrm [16:13] oh, yeah [16:13] ndec_, yes,m since unity-2d was dropped [16:13] furthermore it has different omap code used for android devices [16:13] not what is in mainline ... [16:14] why are you limited to that ancient kernel ? [16:14] ogra_: hmm... i missed that. in which archives are you pushing the pvr package, then? [16:14] ogra_: that is what google provides with Android [16:14] ah, crap, i thought all android would at least be at 3.1 already [16:14] no [16:14] where would the fun be then? [16:15] nexus7/10 has already a 3.4 kernel [16:15] :) [16:15] morphis, nexus7 3.4 kernel? [16:15] I heard something like this [16:15] with the recent android update 3.4 afaik [16:15] nexus10 maybe [16:15] we'Re still on 3.1 [16:15] s/3.4/4.2/ [16:15] that would be nice [16:16] a little ahead of time though [16:16] the 4.2 OTA should come any hour now [16:16] anyway, just merging in 3.1 causes too much merge conflicts to resolve :) [16:16] https://android.googlesource.com/kernel/tegra/ this still only has 3.1 [16:16] so we're using directly android graphic drivers [16:16] and this is the tree we work from [16:16] morphis, ouch, painful [16:16] maybe the nexus10 being a completely different SoC is on 3.4 [16:16] with a approach called libhybris which allows you to use bionic-linked libs [16:17] ogra_: yes, but it works .... [16:17] just a compositing solution is missing [16:17] morphis, we had other PVR drivers n earlier releases in ubuntu, with a little digging you might be able to find something that works with 3.0 [16:17] ogra_: I searched already through the archive but didn't found something usable or working [16:19] we're now on the way implementing some compositing mechanism on top of libhybris so more than one process is able to use acceleration ... [16:21] ogra_: but if you have a direct reference to a pvr driver which does not need omapdrm I would be very happy :) [16:22] all before 12.04 shouldnt need it [16:22] but they need a huge set of kernel patches instead [16:22] for DSS [16:23] hm [16:23] there is also the TI PPA [16:23] yes, I know it [16:23] that was one of the places I was looking at [16:24] huge set of kernel patches should be problematic as well as they will not apply cleanly on an android kernel ... [16:24] yeah [16:25] will surely need work, but somehow it appears cleaner than using libhybris for the whole graphics stack [16:25] ogra_: but it depends on the effort you need to get this implemented [16:25] * ogra_ isnt a big fan of cludge libs [16:26] I love clean solutions as well but with a limited life time of devices and only less time to do some work in spare free time there is often no way around using workarounds like libhybris [16:27] well, both solutions require to write code [16:27] or at least adjust code [16:27] yes [16:27] writing a compositor from scratch might be as much work as adjusting the patches [16:28] ogra_: but are the pvr drivers not related to x11 only? [16:29] as far as I remember for example the pvrctrl binary wasn't part of the later versions of the driver [16:29] the init sequences was then directly implemented in the xorg driver [16:29] with omapdrm, yes [16:30] ok [16:30] but we would still missing the compositor, right? [16:30] it uses XRENDER so you get composite from X [16:31] or does webos not use X ? [16:31] it does not [16:31] thats one of the important problems [16:31] on the legacy HP devices they had two or three framebuffers [16:32] on the Galaxy Nexus we're limited to one [16:33] ah, yeah, thats dfifferent indeed [16:33] i wasnt aware webos doesnt use X [16:35] ok [16:36] ogra_: if you want to see some code: https://github.com/morphis/libhybris/commit/08057ebbef3f62d4ab370d2622fce52acc266b66 [17:04] sfeole: im not sure, i dont think my panda board ever turned display off [17:05] cwayne: ahhh that would be a good test [17:05] cwayne: def something i can do today [17:05] sfeole: awesome [17:05] cwayne: on precise / quantal x86 the display does indeed turn off when rhythmbox is playing audio (radio / mp3) [17:06] mpte that the panda has issues with DPMS [17:06] *note [17:06] sfeole: so that's correct [17:06] perhaps this is a nexus bug then [17:07] cwayne and sfeole are speaking about bug# 1071213 [17:07] (you might end up with a green screen on the panda and have to switch to console and back to get X back) [17:07] lol sfeole refers to himself in the third person [17:08] ;P [17:08] I ended up getting to 13.04, and I was logging in, but it couldn't launch Ubuntu Sesion [17:08] I realized I didn't install any other desktop or window managers to update the driver [17:09] so I get to start from the top again hahaha === doko_ is now known as doko [18:00] janimo: Ew. [18:29] infinity: is that related to our nexus7 kernel? :) [18:30] achiang, related to my person I think :) [18:35] janimo: that i can't help with ;) [18:39] infinity, really. I am open to suggestions, and if it is any consolation I did not enjoy making this packages more than you enjoy reviewing them :) [18:40] janimo: That was specifically for the firmware with the license agreement. I'm looking at both today, though. [18:42] infinity, thanks. I tried using the minimum amount of debconf boilerplate [18:43] I would not mind a bit if someone wrote a dh helper that picked up a debian/package.LICENSE.SHOW file and DTRT generating all debconf machinery [18:44] janimo, oh, you missed to add a skip option we can use during image build [18:44] ogra-cb, ah indeed [18:44] still is there no more obvious envvar than FLASH_KERNEL to suggest we are on a builder? [18:44] something should check for FLASH_KERNEL_SKIP [18:45] Hrm? === Ursinha-afk is now known as Ursinha [18:45] No. [18:45] well, thats definitely the most reliable one we have atm [18:45] ogra-cb, or cannot the live builder silence all debconf somehow? [18:45] it defaults to noninteractive [18:45] on which your preinst would fail [18:45] ah and the default is false here [18:45] If you need to skip it, you preseed the agreement. [18:45] hmm, k [18:46] No need for magic here, the magic already exists. :P [18:46] infinity, but we still need a package change ? [18:46] yeah, but means livecd-rrotfs hackery [18:46] *rootfs [18:46] ogra-cb, can you pont me to a similar one? (or day I say upload a fix to the ppa? ) :) [18:46] That said, if we can't install this for users without them agreeing to said agreement, how does automating it in an image build end up in a sane state? :P [18:46] janimo: No need for a packaging change, this is the whole point of debconf. [18:47] infinity, usb/creator will show the saem license [18:47] infinity, hurray then [18:47] infinity, thats debconf thing is just for users installing the package standalone [18:47] *that [19:06] ogra-cb: well, I don't know about your plymouth patch yet, but I see that add_consoles_from_file() is clearly never used because it's broken and never terminates :P [19:12] <[mbm]> couldn't the kernel just use a hardcoded commandline, avoiding this whole console=none thing? [19:12] ogra-cb: hnngh, I keep reading other bits of the plymouth code for context and finding bugs; command_line_get_string_after_prefix also fails if you have a command line of 'bogus_console=foo console=none' [19:15] ogra-cb: can you clarify the intent of the patch? I don't think it does what you want, but I'm not entirely sure what you want [19:32] Is there any ubuntu openstack image for arm? Or any project for that now? [19:34] Any kind responce please [19:51] Lavvy: Define "openstack image". Most of the openstack bits exist on ARM (except for the ones that can't, like the nova compute backends for Xen and KVM), but there's no specific image built including those packages by default. Easily installed on ubuntu-server, however, just as with x86. [19:58] I'm about to try to patch my video driver; wish me luck [20:32] slangasek, well, avoidingthat plymoputh tries to open /dev/none if there is a console=none set and at the same time skipping to the next console= option if it is either "none" or unset [20:33] ogra-cb: ok... but why would we ever pass either console= or console=none? What's the intended meaning of those commandline options (and what's interpreting it)? [20:34] ogra@chromebook:~$ cat /proc/cmdline [20:34] cros_secure console= console=tty1 debug verbose root=/dev/mmcblk0p7 rootwait rw [20:34] ubuntu@nexus7-roccos:~$ cat /proc/cmdline [20:34] tegra_wdt.heartbeat=30 tegraid=30.1.3.0.0 mem=1022M@2048M android.commchip=0 vmalloc=128M androidboot.serialno=015d2109ce281010 video=tegrafb no_console_suspend=1 console=none debug_uartport=hsport usbcore.old_scheme_first=1 lp0_vec=8192@0xbddf9000 tegra_fbmem=8195200@0xabe01000 core_edp_mv=0 audio_codec=rt5640 board_info=f41:a00:1:44:2 tegraboot=sdmmc gpt gpt_sector=14934015 androidboot.bootloader=3.34 root=/dev/mmcblk0p9 ro console=tty1 fb [20:34] con=rotate:1 quiet splash [20:35] that's not "why" :) [20:35] slangasek, the android bootloaders set such crap ... and it is hardcoded [20:35] ah [20:35] the nexus options *we* set start at root= for example [20:36] ok [20:36] so the only issue I see with your patch is a memory leak [20:36] should i have freed console ? [20:37] o see though that debian has 0.8.5 and i know that i read somewhere that console handling woudl eb better there [20:37] Are interpreted languages making you soft? :) [20:37] yes, you want to free(console) and you don't need to set console=NULL [20:37] Or, rather, garbage-collected languages. [20:38] lazy for sure :) [20:38] yeah, we should look at updating to the new version [20:38] not via Debian though, the packaging is disjoint [20:38] and a merge is hell [20:38] not sure it actually does what it shoudl (only using the very last console= it finds) [20:38] so we should look at pulling the new upstream version, and then worry about a packaging sync separately [20:39] but i have some hope it is at least slightly better [20:39]