[02:00] Has anyone done any work to be able to boot into fastboot from linux? (like adb reboot bootloader) [02:01] on n7 [02:10] from a n7 running ubuntu? [02:10] you can't yet, youd have to just power off + on while holding volume down [02:11] it shouldn't be too difficult to manually do though [02:11] (reboot into fastboot) [02:11] from memory it is a bootloader message on misc [02:14] if all you have is ssh and you're locked out of where it is, it's helpful [06:57] <[mbm]> uberushaximus: 'adb reboot bootloader' just runs the command 'reboot bootloader' and reboot then passes 'bootloader' as an argument to the kernel reboot function, so really it's all about which kernel was used [11:32] hrw, lol, did you notice that if you copy the while content of /opt from the cros disk to ubuntu, copy two additional libs and run /opt/google/chrome/chrome, you actually get the chromeos desktop in a window [11:33] s/while/whole/ [11:33] it seems to require working gles though, its ultra slow [11:35] but i bet with working 3d accel it will work fine, so you could use the whole browser with all plugins (incl flash and hangouts) under ubuntu [12:00] ogra-cb: nice [12:01] ogra-cb: copy /usr/lib/lib{mali,EGL,GLES,GL}* and you may have opengles [12:01] tried already, isnt enough [12:01] I was able to run glmark2-es once [12:01] i need wiorking armsoc for that i think [12:01] ah [12:01] have to put it into ppa [12:01] * ogra-cb is on fbdev [12:01] let me create package [12:01] I am on armsoc [12:02] give me 30 minutes :) [12:02] i didnt upgrade the system from precise yet [12:02] and tryig to build it myself failed due to missing deps [12:02] (libdrm-omap or so) [12:02] apt-getbuild-dep xf86-video-omap or sth like that [12:03] thats not in precise [12:03] ok [12:03] I run raring [12:03] i know [12:03] :) [12:04] wasnt breave enough to upgrade yet [12:04] the touchpad works flawless for me with evdev btw [12:04] including all the multitouch stuff [12:05] i had to adjust accel and snesitivity a bit through the UI tool but after that its not worse than in cros [12:05] share accel/sens values? [12:06] oh, weeme it doesnt persist over reboots, heh [12:06] *seems [12:06] so it use the defaults since a while [12:06] nevermind [12:07] ;D [12:07] package sent to ppa [12:09] oh, cool, i just found out that two finger scrolling works [12:09] both directions [12:09] just needed to enable it in the gnome gui [12:09] 1,2,3,4 finger tap too [12:09] no idea what for 4tap is anyway [12:09] it defaults to edge scrolling, that doesnt really work right [12:10] 4 tap brings up the dash in unity [12:10] so likely emulates the super key in other envs [12:10] * lilstevie is jealous [12:11] the tf dock trackpad refuses to work properly with anything [12:11] sad [12:11] even with mtrack [12:12] mtrack works, but gives no scrolling ability at all [12:12] at which point one might as well force single touch mode in the driver [12:12] rouchpads are dead anyway in the aera of bathroom windows^W^W windows 8 [12:12] touch [12:12] heh [12:14] ogra-cb: super key is on 'find' key [12:14] yeah [12:14] which gets pretty annoying in unity over time [12:14] i always accidentially hit it [12:14] remap it? [12:15] yeah, i will if that goes on [12:15] I would love to have usable xkb editor [12:15] lack of pgup/down suxx [12:15] i actually would like to get unity3d in raring to work though, then i'll likely need it [12:15] as a comparison system for the nexus [12:16] ++ [12:16] you could try to remap shift up/down tp pgup pgdn [12:16] or something like that [12:17] beond the mapping issues i'm really impressed by the kbd [12:19] ogra-cb: you have UK or US one? [12:20] uk [12:20] I will work on using find up/down for pgup/down [12:20] shift+up/down are already useful [12:55] hwwhat ppa did you upload to ? [12:55] hrw, ^^ [12:55] hrw:my-own-packages [12:56] or rather: ppa:hrw/my-own-packages [12:56] k, i dont see anything there [12:56] alsa-lib is the latest [12:56] I see [12:57] * ogra-cb finds it funny that the battery meter can act as a load meter [12:57] if i scroll a lot in firefox the remaining time drops significantly, as soon as i stop scrolling it bumps up again [12:58] ;) [12:58] I know why no ppa [12:58] you uploaded to raring directly ? [12:58] :) [12:59] forgot -sa [12:59] ah [13:00] I need to sort out versioning [13:00] armsoc driver has upstream, fork, different versions etc [13:01] ah, i see it now [13:01] I am thinking about using 0.0+gitYYYYMMDDrSHAID version [13:02] Linaro upstream is at 0.5.1, Chromium fork is 0.0.1 but has code from 0.4* etc [13:02] That's safe as an upstream version [13:02] yeah, sounds sane [13:02] (only 0.0.0~ or similar breaks stuff) [13:02] so i wonder if i can make that build on precise [13:02] ogra-cb: if you get b-d packages then it will work [13:03] or if i have to upgrade and lose my desktop [13:03] ywah, that might take some effort [13:03] I want to replace my patch with patches from xf86-video-omap x11 update but need some time to fix it [13:10] ah, i just needed to pull libdrm from quantal [13:11] bah, or not [13:11] failed [13:12] omap_dri2.c: In function 'OMAPDRI2ScreenInit': [13:12] omap_dri2.c:586:4: error: unknown field 'ReuseBufferNotify' specified in initializer [13:13] :( [13:19] hrw, did you just dist upgrade from precise to raring or did you properly use update-mamanger ? [13:24] haha [13:24] "your graphics may not be fully supported in 12.10, you might end up with a slow desktop ...." [13:25] apt-get dist-upgrade as usual [13:26] well, i'm going the proper route now === ben1066_ is now known as ben1066 [13:59] Hi, is there somebody involved with the Nexus 7 port of Ubuntu? I'd like to discuss dual-booting ubuntu with android. [14:02] [14:03] Tassadar, no plans to work on it ... see https://wiki.ubuntu.com/Nexus7/FAQ#Do_you_plan_to_support_dual_booting_Ubuntu_and_Android.3F [14:04] No, I mean I did it, and I'd like to ask some details about how to change some things in ubuntu to make it work better [14:05] it will definitely break with one of the next kernel upgrades [14:05] (or with the next package that triggers regeneration of the initrd, whatever happens first) [14:06] Yeah, I know, that is why I am here :) [14:06] it would require massive changes to the design of flash-kernel [14:07] the ramdisk which it flashes must be generated from something, no? [14:07] update-initramfs from the initramfs-package dooes that [14:08] any package can trigger such a rebuild and the last step of update-initramfs is to flash the new initrd to the boot partition [14:08] i'd only need to move init -> main_init, add my own init binary, busybox and one symlink to the ramdisk [14:08] using flash-kernel [14:09] that wont prevent flash-kernel from flashing over the android kernel; [14:09] that is okay [14:09] it will trash your android every time it runs in the current setup [14:10] I use different boot.img for both android and ubuntu [14:10] ogra-cb, "Can we remove the Google Logo from the bootloader? There are no firm plans to try this, although it may be possible. Patches welcome!" <-- most likely not, the bootloader is sigchecked on flashing, even while unlocked [14:10] my init then asks you what to boot, and it will either proceed with boot, or flashes new boot.img and restarts [14:11] it is the only solution I could think of, besides kexec, which is pretty hard to make working :/ [14:11] you flash on every boot ? [14:11] kexec would be fine with the hardboot hack [14:11] that is what I do for the tf201 [14:11] that will kill your MMC at some point [14:11] exactly why I didn't go with something like that :p [14:12] lilstevie: that is not much faster than reboot, but I'll look into that [14:13] ogra-cb: no, not every, only when it is needed, eg. android->ubuntu and vice-versa [14:13] well, indeed, thats what i meant [14:14] Tassadar, it is no faster than a reboot, just a hell of a lot less mmc wear [14:14] the problem is it would require kernel modification [14:14] that it does [14:14] at this stage [14:14] although we have been looking into some changes to solve that [14:17] somebody on XDA had idea if fastboot boot *boot.img* could be used in some way, but I am do not think that bootloader sources are available [14:18] well we use fastboot for flashing the same way [14:19] i guess thats possible [14:19] but you need a second machine [14:19] and need to boot into fastboot mode [14:20] well, I was thinking that if fastboot just puts the boot image somewhere in ram and then restarts the device, then it could be possible to do the same without fastboot mode, but then again, hard to say without sources, it probably does not even restart the device after fastboot boot :/ [14:21] why would it reboot? [14:21] fastboot boot just loads the kernel into ram, then calls the boot method [14:21] no reason, thats just me hoping it would be so easy) [14:39] lilstevie: I have the kexec-hardboot patch opened, and I see KEXEC_HB_PAGE_ADDR define - i suppose this needs changing, where exactly should it be? The value in patch (0x57fff000) doest not remind me of anything [14:42] Tassadar, is that the original, or the one we have for the tf201. I don't have the source right in front of me this second [14:42] it is the one for epic 4g, so the original I suppose. Is the one for tf201 on XDA? [14:42] it is on github [14:43] gimme sec, have the code in front of me now [14:44] right [14:44] HB_PAGE is the page that the decompressor reads to check for the hardboot flag [14:45] it ideally should be part of standard memory that is out of the way of anything that may run over it during the hardboot cycle [14:46] so something near the ram_console should be ideal [14:46] for the TF201 we have it at 0xBEC00000 [14:46] which is a little before fbmem [14:46] you really need to check what the memory layout is for your device [14:47] originally we had it at ram_console-SZ_1M but that was inside bootloader fb [14:47] it really depends on memory layout [14:48] you shouldn't need to worry about normal kernel operations, or userspace hitting it, it is one of the last things set before the reboot, and one of the first things checked in the decompressor [14:50] yeah, well, I really feel I need a bit more knowledge to do this - so, decompressor == part of bootloader, which decompresses the kernel (zImage)? [14:51] decompressor == part of the kernel [14:51] specifically the first code to be executed from the kernel [14:52] okay, so the kernel like, decompresses itself? [14:53] yes [15:17] lilstevie: how do I check the memory layout? [15:23] hm... opengles works only for root [15:24] libGLESv2.so.2 => /usr/lib/arm-linux-gnueabihf/mali-gles/libGLESv2.so.2 (0x76d9f000) [15:25] ideas? [16:37] lilstevie: could you please send me link to that github repo with hardboot kernel for tf201? [17:09] hrw, yay, unity 3d ! [17:14] hrw, your package needs a udev rule to make /dev/mali0 writable for the user (just hand it to udev-acl) [17:24] hrw, hmm [ 730.601] (EE) AIGLX error: dlopen of /usr/lib/arm-linux-gnueabihf/dri/armsoc_dri.so failed (/usr/lib/arm-linux-gnueabihf/dri/armsoc_dri.so: cannot open shared object file: No such file or directory) [17:50] hrw, /lib/udev/rules.d/69-mali-gpu.rules [17:50] hrw, http://paste.ubuntu.com/1348470/ [17:51] hrw, makes GLES work OOTB here for non root users [17:52] hrw, or better http://paste.ubuntu.com/1348479/ [17:52] (without mentioning tegra :P ) === sv is now known as Genki === Genki is now known as Sv [21:32] hmm, so dpms is really broken with the armsoc driver [22:57] http://www.chromestory.com/2012/10/how-to-install-ubuntu-on-the-new-arm-chromebook/ [22:57] How can I do that except install it to the SSD? [22:57] Don't really need or want ChromeOS on it [23:09] I'll be idling for awhile, just ring me if you know