=== prp^2 is now known as prpplague | ||
=== doko_ is now known as doko | ||
alfonsojon | hrw: Hey, olga_ sent me here | 10:18 |
---|---|---|
alfonsojon | ogra_* | 10:18 |
alfonsojon | not olga... | 10:18 |
ogra_ | heh | 10:19 |
alfonsojon | sorry ogra | 10:19 |
alfonsojon | honest mistake | 10:19 |
alfonsojon | ogra looks like olga | 10:19 |
ogra_ | :) | 10:19 |
=== chihchun_afk is now known as chihchun | ||
alfonsojon | so ogra_ | 10:22 |
alfonsojon | (not olga) | 10:22 |
alfonsojon | Is that rumor true or no? | 10:22 |
ogra_ | what ? | 10:22 |
ndec | ogra_: not sure you look like this ... https://www.google.fr/search?q=olga&source=lnms&tbm=isch... | 10:22 |
ogra_ | well, i kind of have a similar haircut | 10:23 |
alfonsojon | lol | 10:23 |
alfonsojon | What I'm referring to is that I think phoronix or some other site said that 13.10 is bringing better support for the ARM Chromebook with an accelerated graphics driver | 10:24 |
ndec | if phoronix said it, it must be true. | 10:24 |
ogra_ | well, i think hrw uploaded the armsoc driver ... | 10:24 |
ogra_ | but you still need to get the GLES libs manually in place | 10:25 |
ogra_ | iirc they cant be packaged due to missing licensing | 10:25 |
alfonsojon | darn | 10:25 |
hrw | and we have too old kernel for current ones | 10:25 |
ogra_ | (but my info might be ouotdated, hrw is really the better person to answer this) | 10:25 |
alfonsojon | oh hi hrw | 10:25 |
hrw | no change for opengles | 10:25 |
alfonsojon | = no minecraft / games / anything | 10:25 |
hrw | I have decided to not work on packaging them at all | 10:25 |
alfonsojon | ? | 10:25 |
* tgreer puts a brick through his tegra 2 board | 10:26 | |
ogra_ | hrw, the nexus10 ones work as well, btw | 10:26 |
hrw | alfonsojon: I find your lack of faith disturbing | 10:26 |
alfonsojon | minecraft: pi edition could be ported | 10:26 |
alfonsojon | maybe? | 10:26 |
ogra_ | so pepole can just download the android GLES package from google | 10:26 |
hrw | ogra_: my chromebook went to repair | 10:26 |
ogra_ | yeah, i saw | 10:26 |
alfonsojon | Actually, couldn't the surfaceflinger driver be ported or is it proprietary? | 10:27 |
hrw | ogra_: does it has any useful license? | 10:27 |
hrw | alfonsojon: feel free to do it | 10:27 |
alfonsojon | I know nothing but shell | 10:27 |
alfonsojon | I would if I could :/ | 10:27 |
hrw | I decided to help getting base stuff working. will not work on opengles, openmax etc stuff | 10:27 |
ogra_ | hrw, it is a shellscript merged with binary data ... as all these android driver packages ... you have to agree to the scripted eula before it even unpacks to disk | 10:27 |
hrw | no binary blob related stuff | 10:28 |
hrw | ogra_: eula or 'google terms of service'? | 10:28 |
ogra_ | the latter i think | 10:28 |
hrw | I do not consider it license | 10:28 |
ogra_ | isnt that the same thing, just with different name ? | 10:28 |
hrw | ogra_: ask ubuntu archive maintainers | 10:28 |
ogra_ | well, see http://cdimage.ubuntu.com/ubuntu-touch/daily-preinstalled/20130612/ | 10:29 |
ogra_ | (the legal notice) | 10:29 |
hrw | ogra_: opengles for chromebook got rejected from NEW with g-t-o-s used as license | 10:29 |
ogra_ | we will have to have the N10 blobs in the archive after all | 10:29 |
ogra_ | for touch | 10:29 |
hrw | ogra_: if they will work on chromebook then small package providing symlinks and we are done | 10:30 |
alfonsojon | Here's what I'm getting out of this (I'm not very technical compared to you guys) | 10:30 |
ogra_ | well | 10:30 |
ogra_ | they will likely be shipped inside dteh packaged android ... | 10:30 |
ogra_ | and you dont want the container running on a real chromebook | 10:31 |
alfonsojon | The display driver for the samsung display driver is not open source | 10:31 |
hrw | alfonsojon: o rly? | 10:31 |
ogra_ | so i guess some hackery will be needed | 10:31 |
alfonsojon | But the one for android is, but that runs on display flinger? | 10:31 |
hrw | alfonsojon: define 'display driver for the samsung display driver' please | 10:31 |
ogra_ | (installing the container, but not running it and have a postinst that sets the links for xorg) | 10:31 |
alfonsojon | I feel I'm getting confused | 10:31 |
alfonsojon | hang on | 10:31 |
hrw | ogra_: get it into archive and then we can take a look | 10:31 |
ogra_ | GLEs isnt depending on the display server | 10:32 |
alfonsojon | the mali t604 | 10:32 |
hrw | alfonsojon: so please explain to my why I have both text console and X11 running on my FOSS chromebook? | 10:32 |
ogra_ | well, it is ... but less than you think :) | 10:32 |
alfonsojon | ? | 10:32 |
alfonsojon | to my why I have | 10:33 |
alfonsojon | confusion | 10:33 |
hrw | alfonsojon: start 13.04/13.10 on your chromebook | 10:33 |
alfonsojon | oh | 10:33 |
hrw | alfonsojon: then get x11 working | 10:33 |
alfonsojon | I have a blinking cursor on 13.04 and 13.10, nothing else. Plymouth just dies and I can't even get to tty | 10:33 |
alfonsojon | Plymouth "takes over" tty | 10:34 |
hrw | alfonsojon: you will still do not have any binary blob (other then ones in linux-firmware) | 10:34 |
rbasak | There's a plymouth bug | 10:34 |
ogra_ | remove the plymouth upstart jobs | 10:34 |
alfonsojon | No but even without plymouth xorg dies | 10:34 |
alfonsojon | I did that | 10:34 |
alfonsojon | xorg dies | 10:34 |
rbasak | bug 1082742 | 10:34 |
ubot2` | Launchpad bug 1082742 in plymouth (Ubuntu) "plymouthd: ply-terminal.c: 611 ply_terminal_open: Assertion `terminal != ((void *)0)' failed." [Undecided,Confirmed] https://launchpad.net/bugs/1082742 | 10:34 |
ogra_ | rbasak, well, its more of a console-setup race :) | 10:34 |
alfonsojon | that's it | 10:34 |
rbasak | Workaround is to disable plymouth. My chromebook-plymouth-hack package in the PPA does it for you | 10:34 |
ogra_ | we have one open since 2 years or so for it | 10:34 |
alfonsojon | in which PPA? | 10:35 |
hrw | alfonsojon: arm-chromebook one | 10:35 |
ogra_ | the proper workarounf is to use an initrd with FRAMEBUFFER=Y set | 10:35 |
rbasak | ppa:chromebook-arm/ppa | 10:35 |
hrw | ogra_: we do not have initrd | 10:35 |
ogra_ | if you cant use an initrd, just add overrides for all plymouth jobs | 10:35 |
hrw | anyway once I will get speakers replaced (which may take weeks) I plan to move to U-Boot instead of signed kernels | 10:36 |
rbasak | Also, you need to fire a plymouth-started event or something like that for lightdm to start. My package does all of that :) | 10:36 |
ogra_ | after all someone needs to fix the race ... console-setup needs to be executed before plymouth starts, then everything is fine | 10:36 |
alfonsojon | U-Boot :D | 10:36 |
ogra_ | but if you hardcode that it will delay the boot | 10:36 |
ogra_ | which is why nobody worked on a fix yet | 10:36 |
alfonsojon | okay so I tried arch arm | 10:37 |
ogra_ | s/delay the boot/delay the boot for non affected systems/ | 10:37 |
alfonsojon | hated it | 10:37 |
alfonsojon | there was nothing to work with except pacman | 10:37 |
alfonsojon | basically, it's like getting debian with the GNU utilities and apt-get | 10:37 |
alfonsojon | nothing else | 10:37 |
alfonsojon | So will I be able to get unity up and running with the plymouth workaround? | 10:38 |
ogra_ | i run it here ... thogh i'm still using the chrubuntu install | 10:38 |
rbasak | AIUI, Unity won't work, due to Google's supplied GLES stuff not supporting compositing. Or has somebody managed to get it to work? | 10:38 |
alfonsojon | I used Chrubuntu | 10:38 |
ogra_ | and dont expect it to be fast | 10:38 |
hrw | unity on chromebook? | 10:38 |
alfonsojon | yep | 10:39 |
ogra_ | yeah, worls fine | 10:39 |
ogra_ | *works | 10:39 |
alfonsojon | Ogra, did you install Chrubuntu 12.04? | 10:39 |
hrw | ogra_: on mesa I think? | 10:39 |
ogra_ | hrw, half half it seeems | 10:39 |
rbasak | I worked out my own install instructions yesterday based on 13.04 Ubuntu Core | 10:39 |
hrw | alfonsojon: you want ubuntu on chromebook? 13.04 or later | 10:39 |
alfonsojon | Okay | 10:39 |
hrw | rbasak: https://github.com/jay0lee/chrubuntu-script/commits/master | 10:39 |
ogra_ | some bits seem to be accelerated, for some it seems to fal back to mesa | 10:39 |
alfonsojon | oh boy :D | 10:40 |
ogra_ | if it would be fully accelerated it would fly on that HW | 10:40 |
alfonsojon | it should | 10:40 |
alfonsojon | this is a beast of an ARM-based machine | 10:40 |
ogra_ | Mir will solve that for us ... | 10:40 |
hrw | ogra_: add "--use-gl=egl" to chromium browser and check how it will display | 10:40 |
rbasak | I've not tidied it up yet but will post when it's ready. | 10:40 |
ogra_ | but desktop Mir is still two releases away | 10:40 |
hrw | ogra_: 14.10? | 10:40 |
rbasak | hrw: thanks. I didn't know that existed. I'm doing it differently - setting up an SD card on a host machine. | 10:40 |
ogra_ | hrw, it does fine, thats what i'm using | 10:40 |
ogra_ | just cant use flash | 10:40 |
ogra_ | that collides with compiz' compositing | 10:41 |
hrw | ogra_: I had too many issues with opengles on chromebook so decided to skip it | 10:41 |
alfonsojon | Flash is just | 10:41 |
alfonsojon | ugh | 10:41 |
hrw | rbasak: jaylee created that repo on my request | 10:41 |
hrw | rbasak: check amount of changes done after that d: | 10:41 |
ogra_ | well, apart from the "x crashes if you switch ttys" i have no issues at all here | 10:41 |
hrw | rbasak: but it supports all chromebooks | 10:41 |
ogra_ | and thats not a GLES issue but something with the Xserver | 10:42 |
alfonsojon | hrw: have you considered bitbucket? | 10:42 |
alfonsojon | just curious | 10:42 |
hrw | alfonsojon: for hosting git? | 10:42 |
alfonsojon | Yep | 10:42 |
hrw | alfonsojon: for public stuff I use github cause it is popular and works well. for private I have own server | 10:42 |
alfonsojon | ah | 10:43 |
rbasak | hrw: my approach is to minimise dependencies on Chrome OS. Enabling developer mode, USB boot and grabbing binaries - that's all I want to do there. | 10:43 |
hrw | alfonsojon: I do not care is their backend open or not | 10:43 |
rbasak | Then it's easier to test and develop, since I can automate everything. | 10:43 |
hrw | rbasak: I plan to open case, remove write protect, flash uboot, use as any arm device | 10:43 |
rbasak | I don't want to open the case :) | 10:43 |
alfonsojon | If it is uboot-based, could I boot off the USB 3.0 port? | 10:43 |
hrw | alfonsojon: no | 10:44 |
alfonsojon | darn | 10:44 |
alfonsojon | I have a 32 GB USB 3.0 flash drive, and I can't even use its full speed | 10:44 |
hrw | alfonsojon: but you can load kernel from emmc/sd/usb and boot to usb3 rootfs | 10:44 |
alfonsojon | unless I do a chroot | 10:44 |
alfonsojon | true | 10:44 |
alfonsojon | Is it normal to say "umount: device busy" right before it repartitions the SSD? | 10:46 |
alfonsojon | I just want to be sure partitioning doesn't go wrong. It's worked fine before but I'm not sure if it's supposed to be there | 10:46 |
alfonsojon | hrw: I'm using your modded chrubuntu script | 10:46 |
hrw | alfonsojon: ? | 10:50 |
hrw | alfonsojon: I never had such one in public place iirc | 10:50 |
alfonsojon | hm | 10:51 |
alfonsojon | It's not your script specifically | 10:51 |
alfonsojon | It happens with the normal chrubuntu script too | 10:51 |
alfonsojon | works fine though | 10:51 |
* hrw afk | 10:52 | |
alfonsojon_ | Hey | 11:55 |
alfonsojon_ | hrw: I got Xubuntu working | 11:56 |
alfonsojon_ | The performance is terrible though | 11:56 |
alfonsojon_ | hrw: Okay, Minecraft's working | 12:13 |
alfonsojon_ | At one FPS | 12:13 |
hrw | oh, ah | 12:13 |
alfonsojon_ | but it's something | 12:14 |
alfonsojon_ | I'm using a fresh Xubuntu install with your chrubuntu script | 12:14 |
hrw | alfonsojon_: one thing: it is not my script. I just did few improvements | 12:15 |
alfonsojon_ | No I know | 12:15 |
alfonsojon_ | :) | 12:15 |
alfonsojon_ | I'm just saying it's "yours" because I don't want to refer to it every time as "the script that you modified" | 12:16 |
alfonsojon_ | lazy typing | 12:16 |
alfonsojon_ | Also what would be the optimal /etc/X11/xorg.conf? | 12:17 |
alfonsojon_ | hrw: *poke | 12:21 |
hrw | alfonsojon_: 'chrubuntu script' is best name | 12:21 |
alfonsojon_ | Okay | 12:22 |
hrw | alfonsojon_: copy /etc/X11/xorg.conf.d/ dir from chromeos | 12:22 |
alfonsojon_ | oh | 12:22 |
hrw | easiest | 12:22 |
alfonsojon_ | also what libs should I copy? | 12:22 |
hrw | none | 12:22 |
alfonsojon_ | oh | 12:22 |
alfonsojon_ | okay | 12:22 |
alfonsojon_ | will this result in some acceleration? | 12:22 |
hrw | I assume 13.04 | 12:22 |
alfonsojon_ | right now I can get Minecraft at one-two FPS | 12:22 |
alfonsojon_ | if I do this, will that break? | 12:22 |
alfonsojon_ | (implying OpenGL is working, but not accelerated) | 12:23 |
hrw | alfonsojon_: if you throw chromebook from 16th floor it will get some acceleration before hitting the ground. | 12:23 |
alfonsojon_ | true. | 12:23 |
alfonsojon_ | lol | 12:23 |
hrw | that's the only acceleration I can offer | 12:23 |
alfonsojon_ | Okay | 12:23 |
hrw | alfonsojon_: http://marcin.juszkiewicz.com.pl/2013/04/15/hardware-acceleration-on-chromebook/ is still up to date | 12:26 |
hrw | ogra_: consider alsa-lib 1.0.27.1 from Debian into Ubuntu - ucm profiles are no longer delta | 12:27 |
alfonsojon_ | sorry to sound dumb but how do I go about pulling the kernel from Chrome OS? | 12:31 |
alfonsojon_ | in order to get the GPU acceleration | 12:31 |
ogra_ | hrw, cool | 12:32 |
suihkulokki | hrw: awesome :) | 12:32 |
hrw | alfonsojon_: what makes you think that kernel differ? | 12:32 |
hrw | alfonsojon_: read my blog post please | 12:33 |
alfonsojon_ | http://marcin.juszkiewicz.com.pl/2013/04/15/hardware-acceleration-on-chromebook/ | 12:33 |
alfonsojon_ | this one? | 12:33 |
hrw | yes | 12:35 |
hrw | it is nearly 2 months old but still uptodate | 12:36 |
alfonsojon_ | Okay | 12:38 |
alfonsojon_ | Wait this is a deb | 12:40 |
alfonsojon_ | just unpacked | 12:40 |
alfonsojon_ | hrw: Isn't this just an unpacked deb? | 12:45 |
hrw | alfonsojon_: ? | 12:46 |
alfonsojon_ | the git repo for this | 12:46 |
alfonsojon_ | https://github.com/hrw/chromebook-mali-driver | 12:46 |
hrw | alfonsojon_: it is source version of package | 12:46 |
hrw | alfonsojon_: not maintained anymore | 12:47 |
alfonsojon_ | so what now? :/ | 12:47 |
alfonsojon_ | hey* | 12:48 |
alfonsojon_ | he ashams | 12:48 |
hrw | alfonsojon_: feel free to take care of it | 12:48 |
alfonsojon_ | just bring in the new files from chrome os and repackage? | 12:49 |
hrw | alfonsojon_: git has whole history how it was done | 12:49 |
hrw | alfonsojon_: I just refuse to do anything with it anymore. | 12:49 |
hrw | alfonsojon_: so please change subject of this discussion | 12:49 |
alfonsojon_ | You know | 12:50 |
alfonsojon_ | I have an idea | 12:50 |
alfonsojon_ | Ridiculous and just off-sounding | 12:50 |
alfonsojon_ | What if we got a hold of someone at Google | 12:50 |
alfonsojon_ | some how some way | 12:50 |
alfonsojon_ | And asked what license it's under? | 12:50 |
hrw | alfonsojon_: as long as you will get it officially etc | 12:51 |
tassadar_ | I would imagin he'd respond something like "How did you get into my house at 2AM?" | 12:51 |
alfonsojon_ | xD | 12:51 |
hrw | alfonsojon_: I can only say that Samsung licenced OpenGLES driver from ARM and then sublicensed it for Google to use with ChromeOS. | 12:52 |
alfonsojon_ | So google is the licensee of the licensee in this case? | 12:52 |
hrw | alfonsojon_: that's kind of answer you can get from Google people in corridor talks | 12:52 |
hrw | alfonsojon_: note "for Google to use with ChromeOS" | 12:53 |
alfonsojon_ | Sometimes I love google, sometimes I hate them | 12:53 |
hrw | alfonsojon_: so Google is allowed to distribute it with ChromeOS. | 12:53 |
hrw | alfonsojon_: it is not Google's fault | 12:53 |
alfonsojon_ | I know | 12:53 |
hrw | this is how licenses work | 12:53 |
alfonsojon_ | It's just | 12:53 |
alfonsojon_ | a pile of crap | 12:53 |
alfonsojon_ | licensing and licensing | 12:53 |
hrw | alfonsojon_: check my git repo and decide which way to go | 12:54 |
hrw | alfonsojon_: it already covers two ways | 12:54 |
alfonsojon_ | I'll take a look | 12:54 |
hrw | http://commondatastorage.googleapis.com/chromeos-localmirror/distfiles/mali-drivers-0.45-r106.run is latest available publically version | 12:55 |
hrw | chromeos may have newer one anyway | 12:55 |
alfonsojon_ | if it works it works | 12:56 |
hrw | it does not with 3.4.0-5 kernel | 12:56 |
hrw | that's also mentioned on my blog | 12:56 |
alfonsojon_ | what kernel does this work with? | 12:57 |
alfonsojon_ | 3.6? | 12:57 |
alfonsojon_ | well | 12:57 |
alfonsojon_ | I would assume 3.2 | 12:58 |
hrw | 3.4.0 but from R25 or newer branch. | 12:59 |
hrw | guess what? that's on a blog | 12:59 |
=== lag` is now known as lag | ||
alfonsojon_ | oh :/ | 13:00 |
alfonsojon_ | sorry if I'm getting on your nerves | 13:00 |
hrw | far from it | 13:09 |
hrw | have to go anyway | 13:11 |
=== tgall_foo is now known as Dr_Who | ||
=== prpplague is now known as prp^2 | ||
alfonsojon | hrw: Okay I think I got acceleration (LightDM works now), but no sessions load | 14:17 |
=== chihchun is now known as chihchun_afk | ||
martin-aa6e | Is anyone running Ubuntu 13.04 on BeagleBoard Black? | 22:43 |
=== martin__ is now known as martin-aa6e | ||
martin-aa6e | (glitch here) still looking for anyone using U13.04 with BeagleBoard Black | 22:49 |
=== chihchun_afk is now known as chihchun |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!