[00:15] <sfeole> quince: hold the side power button down and it should reset
[00:16] <sfeole> quince: thats a known issue (onscreen keyboard becoming unresponsive)
[00:50] <quince> sfeole: I wasn't able to reset it by holding the power button down (for about 5 seconds -- perhaps not long enough?)
[00:52] <lilstevie> quince, what do you mean by "(but not using fastboot -- why?)"
[00:54] <quince> lilstevie: I tried running sudo fastboot reboot, and that printed "< waiting for device >" but had not other obvious effect
[00:55] <quince> *no
[00:55] <lilstevie> quince, but everything else you typed would suggest you were in userland
[00:55] <sfeole> quince: yea not long enough, 15 seconds will reboot it.
[00:56] <quince> userland?  not sure what you're referring to in this context
[00:56] <quince> sfeole: ok
[00:56] <lilstevie> quince, you were running ubuntu
[00:57] <lilstevie> from what it sounds like
[00:57] <lilstevie> or android
[00:57] <quince> on which device?
[00:57] <quince> both my desktop and the nexus 7 were running ubuntu
[00:57] <lilstevie> yeah I mean on your nexus 7
[00:57] <lilstevie> fastboot only works in fastboot mode
[00:57] <quince> I was running fastboot on the desktop
[00:57] <lilstevie> which is in bootloader
[00:58] <lilstevie> ok, let me simplify this
[00:58] <quince> ok
[00:58] <lilstevie> I will differentiate both, fastboot-util for the thing you run on your computer
[00:58] <lilstevie> fastboot for on device
[00:58] <quince> think you alrready explained...
[00:59] <lilstevie> to issue commands from fastboot-util your device has to be in fastboot mode
[00:59] <lilstevie> when you are booted into an OS you are not in this mode
[00:59] <quince> right -- was clear from "only works in fastboot mode" :-)
[01:00] <lilstevie> :)
[02:20] <raster> m00
[05:40] <sbaugh> Is Ubuntu on the Nexus 7 at a point where it's usable without an external keyboard/mouse?
[05:43] <achiang> define "usable" ? :)
[05:49] <sbaugh> achiang: Good point :) If I just wanted to do some reading and surf the web and do some tinkering, would I be able to spend a good couple hours without critical bugs?
[05:50] <achiang> sbaugh: you'll probably still hit the "button1 stuck" bug, which is frustrating
[05:50] <achiang> but anecdotally, if you don't exercise the unity elements too quickly (like opening/closing the dash too quickly or opening the indicators) you'll be ok for a while
[05:50] <sbaugh> achiang:  https://bugs.launchpad.net/ubuntu-nexus7/+bug/1068994 is this it?
[05:50] <ubot2`> Launchpad bug 1068994 in ubuntu-nexus7 "button1 gets stuck after a while" [Critical,Confirmed]
[05:51] <achiang> sbaugh: yeah, basically it stops responding to touch input :-/
[05:51] <achiang> it's deep in the guts of X so hard to figure out
[05:53] <sbaugh> achiang: Is that the only major bug I might run into?
[05:53] <achiang> well, there are lots of others, but i don't think they are show stoppers, as bad as button1 stuck
[05:54] <sbaugh> Then, it is time to flash!
[05:55] <achiang> good luck. :)
[06:23] <sbaugh> Hey, heh, I tried flashing but I'm getting a password propmt which rejects my password even though I'm sure I'm entering it right
[06:23] <sbaugh> For permission to run ubuntu-nexus-installer
[06:28] <achiang> should be your typical sudo password
[06:28] <achiang> do you have sudo perms?
[06:31] <sbaugh> Yes
[06:32] <sbaugh> even changing user to root and running it, it still fails
[06:35] <sbaugh> earth : Dec 12 01:11:30 : root : 3 incorrect password attempts ; TTY=pts/3 ; PWD=/home/spencer/Downloads/UbuntuNexus7 ; USER=root ; COMMAND=/usr/bin/[ 1 ]
[06:35] <sbaugh> Is the mail sent to root to notify of the attempt
[06:37] <achiang> interesting, this is not my area of expertise
[06:37] <Ethernin> sbaugh, dude use fastboot
[06:37] <achiang> vanhoof will wake up in a few hours. :)
[06:38] <Ethernin> sbaugh, fastboot is way better and you can actually see what's going on
[06:38] <Ethernin> just download the latest raring image off the wiki
[06:38] <sbaugh> Ethernin: Do you mean do it manually? Fastboot seems to detect my device but it isn't letting me do anything.. mhm
[06:38] <Ethernin> it should
[06:38] <Ethernin> u r root?
[06:38] <sbaugh> by anything, I mean restart
[06:38] <sbaugh> ah
[06:38] <sbaugh> no
[06:38] <sbaugh> i wasn't root when i executed that command
[06:38] <Ethernin> ok 1 - be root
[06:38] <Ethernin> 2 - is your nexus in fastboot mode?>
[06:38] <Ethernin> ?
[06:38] <sbaugh> okay, oyes
[06:38] <Ethernin> k
[06:39] <Ethernin> you should be able to do -
[06:39] <Ethernin> fastboot erase boot
[06:39] <Ethernin> fastboot erase userdata
[06:39] <Ethernin> fastboot flash boot raring*.bootimg
[06:39] <Ethernin> fastboot flash userdata raring*.img
[06:39] <Ethernin> fastboot reboot
[06:40] <sbaugh> oh, haha
[06:40] <sbaugh> how odd
[06:40] <sbaugh> i did fastboot reboot
[06:40] <sbaugh> it seems it's already been flashed
[06:40] <Ethernin> is your bootloader unlocked?
[06:40] <sbaugh> i must have gotten through the first time and not noticed
[06:43] <sbaugh> great, it's booting
[06:44] <Ethernin> saweeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeet
[06:44] <Ethernin> ;-)
[06:44] <Ethernin> which nexus do u have?
[06:44] <Ethernin> 16 or 32?
[06:44] <Ethernin> gb?
[06:44] <sbaugh> 16
[06:45] <Ethernin> word yeah i have one of those
[06:49] <sbaugh> hanging for quite a while on wireless
[06:49] <sbaugh> in the first-start system config
[06:49] <sbaugh> would it hurt to just power cycle?
[06:55] <Ethernin> omg
[06:55] <Ethernin> yeah
[06:55] <Ethernin> i had to deal with that fucking thing like 10 times
[06:55] <Ethernin> did u install the latest raring image?
[06:56] <sbaugh> yes
[06:56] <sbaugh> and now i'm
[06:56] <Ethernin> the initial menu is SUPER buggy when it comes to connecting to the wifi
[06:56] <Ethernin> i usually skip it and connect after
[06:56] <sbaugh> stuck at the account setup
[06:56] <sbaugh> can't seem to enter any text
[06:56] <Ethernin> yeah, u may want to plugin a keybaord
[06:56] <Ethernin> yeah keep tapping
[06:56] <Ethernin> another bug for sure
[06:56] <sbaugh> onboard is up
[06:56] <Ethernin> just keep tapping
[06:57] <sbaugh> okay
[06:57] <Ethernin> the input fields
[06:57] <sbaugh> ok
[06:57] <Ethernin> super buggy but u should be able to
[06:57] <Ethernin> i had the same thing happen like 10 times
[06:57] <sbaugh> is there a way at this point to connect remotely
[06:57] <sbaugh> and do this?
[06:57] <Ethernin> yeah attach a external usb keyboard via OTG cable
[06:57] <Ethernin> and tap tap tap away
[06:57] <sbaugh> no, i mean, ssh or vnc or something
[06:58] <Ethernin> they really need to fix the initial menu, and stop a bunch of services like cups
[06:58] <Ethernin> unity is super slow too
[06:58] <Ethernin> i find xfce and lxde to be the best
[06:58] <Ethernin> gnome classic is pretty good too
[06:58] <Ethernin> yeah but u have to install them first
[06:58] <sbaugh> i was planning on lxde
[06:58] <sbaugh> oh
[06:58] <sbaugh> drat
[06:58] <Ethernin> yeah lxde is very nice
[06:58] <sbaugh> you're right
[06:58] <Ethernin> ;-)
[06:58] <sbaugh> ubuntu doesn't ship with an ssh server, right?
[06:58] <Ethernin> apt-get install openssh-server
[06:58] <Ethernin> once yer in
[06:59] <Ethernin> i knwo the initial menu is really buggy....
[06:59] <sbaugh> isn't there some serial connection thing
[07:02] <sbaugh> yes, I see on here: https://wiki.ubuntu.com/Nexus7/UsingTheDevice
[07:03] <sbaugh> oh god
[07:03] <sbaugh> what is the default accout name and password? ubuntu/ubuntu doesn't seem to work?
[07:05] <infinity> sbaugh: There is no default account.
[07:05] <infinity> sbaugh: First boot after flashing it should give you an installerish deal that prompts to create a user.
[07:06] <sbaugh> infinity: I'm having trouble with that so I thought i'd bypass it with the serial connection.
[07:06] <infinity> Well, see above.  There's no default user.
[07:07] <Ethernin> infinity, the initial menu config after install is SUPER buggy, often u have to tap many times to actually input text into the user/pass fields, and sometimes it doesnt even work
[07:07] <sbaugh> There's no way I can get on the device without going through this setup? :(
[07:07] <Ethernin> infinity, just an FYI if u didn't already know
[07:07] <sbaugh> Because like Ethernin said it's barely functional
[07:08] <infinity> xnox: You know anything about the current uniquity-versus-input issues?
[07:08] <sbaugh> Ethernin: I should be tapping the field itself, correct? And when will I know I am able to input text?
[07:09] <Ethernin> sbaugh, yeah that's what worked for me
[07:09] <sbaugh> (as opposed to the keyboard)
[07:09] <Ethernin> yeah u will have to show onboard too if it isn't
[07:10] <infinity> sbaugh: If you can boot with an arbitrary commandline, you could always boot with init=/bin/sh and set stuff up via serial.  I'm on my way to bed, though, so can't be particularly helpful on that score.
[07:10] <infinity> (Obviously, us fixing whatever's currently plaguing the installer is the better option here)
[07:11] <Ethernin> infinity, thank you!
[07:12] <sbaugh> infinity, I would ask how I would go about booting in such a way, but Ethernin, maybe you can explain?
[07:13] <Ethernin> if it really doesn't work, try rebooting, if that doesn't work u can always reflash
[07:13] <Ethernin> but really u should get an OTG usb cable
[07:13] <Ethernin> so u can plug in a usb keyboard ect
[07:13] <Ethernin> that's always my backup at this point ;-)
[07:13] <sbaugh> yes, I should, but I don't have one right at the moment :)
[07:14] <sbaugh> Would I need to reflash to change init or etc.
[07:15] <Ethernin> honestly I would try reflashing - it's the simplest quickest option
[07:16] <sbaugh> Okay, so how would I flash differently this time around?
[07:17] <Ethernin> no
[07:17] <Ethernin> same
[07:18] <Ethernin> just follow the wiki, erase, and flash
[07:18] <Ethernin> and reboot
[07:18] <Ethernin> before that u could try rebooting the device
[07:18] <Ethernin> by holding the power button
[07:18] <Ethernin> and maybe it will prompt u with the menu again
[07:18] <Ethernin> try that before reflashing
[07:19] <sbaugh> also I seem to have accidentally rotated the screen, is the hotkey super-something?
[07:19] <Ethernin> um....not sure about that one!
[07:25] <sbaugh> I'm not sure I just want to reflash and hope. Isn't there anything more direct I could do?
[07:26] <kulve> sbaugh: http://ruedigergad.com/2012/12/09/nexus-7-easily-flashaccess-entire-userdata-partition/
[07:26] <kulve> you can mount the partition over usb to your PC and add root passwd there. I think then you can log in as root
[07:36] <sbaugh> kulve: Do you know how I would do this more specifically? I'm having trouble telnetting in (I'm also unsure how I would mount the user partition)
[07:38] <kulve> what parts have you done now? Did you already boot with the moslo images?
[07:38] <sbaugh> Yes
[07:38] <sbaugh> But that's as far as I've gotten
[07:39] <gurgalof> it should come up as a usb massstorage device...
[07:39] <kulve> now you should already see the big partition in your Linux box. "dmesg" should tell you the /dev/sdX device for it?
[07:42] <sbaugh> ah, okay, it's appeared
[07:43] <kulve> sbaugh: use "mkpasswd" to create a password hash for you and add it to etc/shadow file on the mounted file system
[07:44] <kulve> to my understanding that's enough to enable the root user
[07:45] <sbaugh> shadow should currently be blank?
[07:46] <sbaugh> ah
[07:46] <sbaugh> whoops
[07:46] <sbaugh> nevermind, ignore me
[07:47] <sbaugh> so I should just put the hash after root? Or make a new line for a new user?
[07:48] <kulve> change it for root
[07:49] <kulve> adding a new user would need editing passwd and group files as well as adding a home directory. So it's easier to enable root user :)
[07:50] <kulve> if you get that working, you can then adduser in a live system (although doing that graphically might be better idea in ubuntu)
[07:50] <sbaugh> kulve, and i'm replacing the * in root:*:151414:etc.?
[07:52] <kulve> yes
[07:52] <kulve> I hope I remember this correctly as I haven't done that many times..
[07:53] <kulve> sbaugh: and do make sure you edit the device's file system and not your real one ;)
[07:53] <sbaugh> good advice :)
[07:54] <kulve> I often cd to the mount point and then vi /etc/foobar instead of vi etc/foobar ..
[07:55] <dholbach> good morning
[07:57] <sbaugh> kulve: as before it is immediately saying "Login incorrect" if I enter root as my login (not letting me get to password)
[07:58] <sbaugh> Do I also need to enable logging in as root, perhaps?
[07:58] <kulve> How are you logging in as I assume you don't have wifi up?
[07:59] <sbaugh> The serial connection mentioned here https://wiki.ubuntu.com/Nexus7/UsingTheDevice
[08:00] <kulve> I think you need to make sure ttyACM0 is in /etc/securetty to allow root login through ttyACM0
[08:00] <sbaugh> well then
[08:00] <kulve> it might not be there by default
[08:08] <sbaugh> great!
[08:08] <sbaugh> awesome, thanks kulve
[08:08] <kulve> np :)
[08:08] <kulve> that moslo stuff is quite neat, if you want copy something big over there
[08:09] <sbaugh> nothing I really need
[08:09] <sbaugh> will the installation wizard still come up if there's a user set up?
[08:10] <kulve> no idea
[08:35] <cortexA9> hello
[08:37] <sbaugh> excellentttttttttt
[09:05] <cortexA9> testing 12.10 to 13.04 on nexus 7 :)
[09:12] <victorp> ogra_, ping
[09:19] <sbaugh> Okay, the issues with the first-run configuration wizard seem to be issues with Unity; I managed to run it under LXDE and it worked fine.
[09:19] <sbaugh> (Or possibly just the window managers)
[09:22] <xnox> infinity: on nexus7 with ubiquity on compiz, ogra reported onboard "fighting" for the screen estate with ubiquity somehow. But I am yet to look into that.
[09:29] <cortexA9> mmmm
[09:30] <cortexA9> preparing to replace bluez..
[09:30] <cortexA9> very strange
[09:30] <cortexA9> it takes few mins
[09:31] <cortexA9> i try to upgrade to 13.04
[09:31] <cortexA9> from 12.10
[09:55] <cortexA9> i have problem with bluez
[09:55] <cortexA9> can i remove it ?
[09:56] <smartboyhw> cortexA9, don't they recommend that you NOT upgrade but instead to refresh the image?
[09:56] <smartboyhw> cortexA9, anyway you can install the 13.04 image using the instructions here: https://wiki.ubuntu.com/Nexus7/Installation
[09:56] <cortexA9> smartboyhw: i tried
[09:57] <cortexA9> but with multirom doesn't work
[09:57] <cortexA9> and now i try to upgrade
[10:03] <cortexA9> lol
[10:03] <cortexA9> i remove it
[10:03] <cortexA9> and now seems work
[10:03] <cortexA9> :)
[10:03] <smartboyhw> cortexA9, :D
[10:04] <cortexA9> i don't care bluetooth :P
[10:32] <cortexA9> no.. doesn't work :P
[10:35] <smartboyhw> :P
[10:36] <cortexA9> wait
[10:36] <cortexA9> try another method
[10:38] <cortexA9> lol
[10:38] <cortexA9> now i don't have wifi :(
[11:25] <cortexA9> there is a problem with wifi on nexus 7 ?
[11:25] <cortexA9> because i can't see that after upgrading the packages.
[11:38] <Laney> Can somebody try running rhythmbox with gstreamer1.0-plugins-ugly on their nexus 7 please?
[11:38] <Laney> See if it crashes like mine does ...
[11:38] <Laney> with ... *installed*, that is
[11:41] <Laney> http://paste.debian.net/215757/
[11:52] <kulve> looks like there's some armv4 assembly that's not supported on the nexus 7's armv7 anymore?
[11:53] <kulve> well, the function is still _armv7_tick..
[11:53] <suihkulokki> kulve: the code might be catching sigill to figure out what instructions are supported
[11:55] <Laney> yeah might be likely; if I continue I get to the real segfault
[11:56] <suihkulokki> ..and that is indeed what is happening, looking at armcap.c
[11:56] <Laney> Program received signal SIGSEGV, Segmentation fault.
[11:56] <Laney> [Switching to Thread 0x4d5e9280 (LWP 3168)]
[11:56] <Laney> 0x4ab93760 in mad_layer_III () from /usr/lib/arm-linux-gnueabihf/libmad.so.0
[11:57] <suihkulokki> kulve: catching SIGILL is btw one common way detect (un)supported instructions
[11:57] <kulve> ok
[11:57] <Laney> gdb > handle sigill pass nostop
[11:57] <suihkulokki> Laney: aye
[12:01] <Laney> https://bugs.launchpad.net/ubuntu/+source/libmad/+bug/989846
[12:01] <ubot2`> Launchpad bug 989846 in libmad (Ubuntu) "libmad causes mpd to seg fault on update" [Undecided,Confirmed]
[12:13] <dholbach> cortexA9, it works for me
[12:19] <dholbach> Laney, yes, it crashes for me too
[12:20] <dholbach> still surprised I don't get anything in /var/crash
[12:20] <Laney> not surprised :-)
[12:20] <Laney> I don't get any apport goodness either
[12:20]  * Laney will have a poke at it now - building libmad debug package atm
[12:20] <Laney> no ddeb for some reason
[12:21] <dholbach> janimo, do you know if apport should run on the nexus7?
[12:22] <dholbach> (like catching crashes)
[12:25] <cortexA9> dholbach: on 13.04 ?
[12:25] <dholbach> cortexA9, yes
[12:25] <cortexA9> oh nice
[12:25] <cortexA9> how you do that ?
[12:25] <dholbach> cortexA9, I tried both an upgrade from 12.10 and a fresh installation - wifi worked for me in both cases
[12:25] <cortexA9> you have multirom too ?
[12:26] <dholbach> cortexA9, what does "dpkg -l | grep linux-firmware" give you?
[12:26] <dholbach> I'm afraid I don't know what multirom means
[12:26] <cortexA9> it's a system for nexus 7
[12:26] <cortexA9> for dual boot
[12:26] <cortexA9> ubuntu and roms
[12:27] <dholbach> cortexA9, ah, no - no dualboot
[12:29] <cortexA9> dholbach: linux-firmware 1.98
[12:29] <cortexA9> firmware for linux kernel drivers
[12:30] <dholbach> you might be missing this one
[12:30] <dholbach> linux-firmware-nexus7                     0.2
[12:30] <dholbach> best to download it here https://launchpad.net/ubuntu/+source/linux-firmware-nexus7/0.2/+build/3986720/+files/linux-firmware-nexus7_0.2_all.deb and copy it over with a usb-stick or something
[12:31] <dholbach> it's a problem which has been fixed since a couple of days already - it was a missing recommends of another package
[12:31] <cortexA9> how can i copy it ?
[12:31] <cortexA9> mmm
[12:31] <cortexA9> i can't with a stick..
[12:31] <cortexA9> i have no connection
[12:33] <dholbach> https://wiki.ubuntu.com/Nexus7/UsingTheDevice#Attaching_a_USB_Device_to_the_Nexus_7
[12:35] <diwic> dholbach, I believe apport is not yet working on ARM but that pitti is working on it
[12:35] <dholbach> hm ok
[12:37] <cortexA9> mmm
[12:45] <cortexA9> with usb cable not working i think right ?
[12:45] <dholbach> hm?
[12:45] <cortexA9> yes i mean
[12:46] <cortexA9> with the cable
[12:46] <cortexA9> attach on pc
[12:46] <dholbach> put the .deb file on the usb stick on your PC and then connect it to the usb cable to the nexus7
[12:47] <cortexA9> i can't see the nexus 7
[12:50] <dholbach> I used a usb mass storage device (usb stick, memory stick), which I first connected to my laptop and then to the nexus7
[12:51] <cortexA9> oh nice
[12:51] <cortexA9> but you have a special cable ?
[12:51] <dholbach> yes - it's mentioned in https://wiki.ubuntu.com/Nexus7/UsingTheDevice#Attaching_a_USB_Device_to_the_Nexus_7
[12:54] <cortexA9> what about serial debug shell ?
[12:54] <dholbach> I have no experience with it
[12:56] <janimo> dholbach, apport should run on the nexus. I know it was disabled by default on the 12.10 images though
[12:56] <janimo> not sure about raring
[12:56] <dholbach> janimo, /etc/default/apport looks like it's enabled
[12:57] <dholbach> and pitti told me to look this up:
[12:57] <dholbach> daniel@nexus7:~$ cat /proc/sys/kernel/core_pattern
[12:57] <dholbach> |/usr/share/apport/apport %p %s %c
[12:57] <dholbach> daniel@nexus7:~$
[12:57] <janimo> dholbach, are things crashing and apport not running?
[12:57] <dholbach> which supposedly indicates that it's up and running
[12:57] <janimo> I don't know much about this side, never saw the apport dialog on nexus7 either now that i think about it
[12:57] <dholbach> janimo, pitti is reflashing his nexus and looking into it
[12:57] <janimo> although I saw recently some apport/arm related changes from pitti
[12:58] <dholbach> yes, apport seems to be able to retrace crashes from foreign architectures, so LP would know what to do with crashes
[12:58] <janimo> but I think it was disabled very early in the nexus project as it probably came up too often
[12:58] <dholbach> but errors.u.c needs an update still
[13:00] <dholbach> I can pass on any more information I get
[14:10] <dholbach> janimo,
 looks like core dump is disabled on the nexus7
[14:10] <dholbach>  I've set core_pattern to /tmp/core.%e.%p.%h.%t
[14:10] <dholbach>  then as a normal user ulimit -c unlimited
[14:10] <dholbach>  and  sh -c 'kill -SEGV $$'
[14:10] <dholbach>  that should produce a file /tmp/core.sh.*
[14:10] <dholbach>  and there is nothing, while it works on a my desktop
[14:11] <dholbach> not sure though what they're going to do about it
[14:15] <janimo> dholbach, I don't see anything in the ubuntu-defaults-nexus7 package that would disable core
[14:15] <janimo> I had no idea we do that either
[14:15] <janimo> just about apport that was off in 12.10
[14:24] <erupter> hi there
[14:28] <erupter> i'm having some weird problems with ubuntu on a pandaboard
[14:29] <erupter> anybody wanting to help?
[14:33] <erupter> is this the right channel to seek help installing ubuntu-desktop on a pandaboard? it's the 10th time I have to start again after the board fails to boot
[14:33] <highvoltage> erupter: if you'd like help, it's better to be specific. you have to provide details on what you've done and what actually went wrong.
[14:54] <erupter> simple
[14:55] <erupter> follow these steps
[14:55] <erupter> https://groups.google.com/forum/#!topic/pandaboard/7DLabw7amBg
[14:56] <erupter> whenever the board finish the ubuntu-omap-extras installation and I reboot, it fails to boot again
[14:56] <erupter> no error message
[14:56] <erupter> it starts booting, getting up to "uncompressing linux kernel, booting"
[14:56] <erupter> and then restarts and bootloops
[14:57] <erupter> either this, or the unity environment fails to load (no launcher) after the dist-upgrade
[14:58] <erupter> the only way that works is the clean 12.04 preinstalled image
[14:58] <erupter> (without hardware acceleration it's quite slow)
[16:33] <Laney> apparently core dumps are disabled in the kernel
[16:34] <dholbach> Laney, do you know if it's easily fixable?
[16:35] <Laney> i don't know what the limitations are
[16:36] <Laney> let's rock and roll over to #ubuntu-kernel
[17:12] <Laney> well, I found out how to enable it in the kernel I think
[17:12] <Laney> let me try it and see what happens
[17:12] <Laney> http://www.rt-embedded.com/blog/archives/enabling-core-dumps-in-embedded-systems/
[17:21] <achiang> Laney: that's a kernel thing? i thought it was userspace (ulimit) ?
[17:25] <Laney> achiang: If I'm getting this correctly it needs to be enabled in the kernel in the first place
[17:25] <Laney> the nexus 7 kernels indeed don't have it compared to the stock ones which do
[17:25] <achiang> Laney: ah, ok. nice find. that should probably be a bug
[17:25] <Laney> i'm checking it out now
[17:25] <Laney> then we can do a fix however kernel fixes get done
[18:00] <sbaugh> my battery seems to be taking forever to charge
[18:01] <kulve> charging with what?
[18:02] <sbaugh> well, I'm not charging with the bundled nexus 7 charger and cord, but with android it still charged
[18:03] <sbaugh> and, I am assuming it isn't charging from the little battery charge screen when it's off
[18:03] <sbaugh> But when I'm in ubuntu I get conflicting messages about whether it's charged
[18:11] <Laney> achiang: OK that appears to work
[18:11] <Laney> I'm EODing now so if you want to then CONFIG_ELF_CORE=y needs to be set in the kernel
[18:11] <Laney> otherwise I'll follow it up first thing tomorrow
[18:11] <Laney> got a whoopsie popup and everything
[18:30] <tassadar_> heh, /proc/last_kmsg survises device shutdown, and kexec-hardboot data too
[18:30] <tassadar_> well, sort of
[19:05] <[mbm]> tassadar_: guess nobody is clearing the ram
[19:05] <[mbm]> I know some of that was inte tional from the ramconsole patches
[19:06] <sbaugh> there was one really good approach to mouse input I saw in an Android VNC client, it treated the touch screen as a touchpad and left and right as regions, would it be "Write a driver" difficult or "Install a driver" difficult to do this with Ubuntu on the N7?
[19:07] <[mbm]> sbaugh: I've been saying the exact same thing
[19:07] <AndreasWatch> Hi, I'm want to add ubuntu-support to an other tegra3 tablet (lenova A2109) and want to know what are the differences between the N7 android- and linux kernel
[19:07] <[mbm]> sbaugh: scummvm is another go0d example of it
[19:09] <sbaugh> [mbm]: What does scummvm do with it?
[19:09] <sbaugh> [mbm]: And, if you've been saying the same thing, do you have anything that might enable such a feature?
[19:10] <[mbm]> sbaugh: scummvm on android uses a relative movement, treating the surface as a laptop touchpad
[19:10] <sbaugh> oh, ok
[19:10] <[mbm]> sbaugh: doubt anyone has actually had time to lok into it; I certainly haven't
[19:39] <tassadar_> [mbm]: yeah, I always thought that it will loose data soon after power-down, and it was shut down for like 5-10s
[19:39] <tassadar_> there was even enough data to cause kexec-hardboot patch to try run kernel from RAM, but it failed
[19:42] <[mbm]> tassadar_: have you ever seen what happens when the video driver crashes and the lcd goes cloudy and eventually turns white?
[19:42] <[mbm]> what's really happening is a slow discharge; very similar to what happens with ram
[19:43] <[mbm]> and actually the colder the chips are the slower the discharge
[19:44] <[mbm]> there was a research paper written up at one point talking about how you could take a laptop, power it down, use an upsideown can of duster to freeze the ram chips and then put them in another machine and search the ram for encryption keys
[19:44] <tassadar_> yeah, I think I saw something like that
[19:45] <tassadar_> well, I suppose I'll need a bit more knowledge about how the RAM chips
[19:45] <tassadar_> work
[19:46] <[mbm]> also I belive there's an open bug on the nexus7 ubuntu port talkign about how the framebufer isn't erased and will show part of the old desktop when x starts
[19:46] <tassadar_> yeah, same thing happens when you reboot to android from recovery
[19:46] <tassadar_> and because android does not use framebuffer anymore, last frame from the recovery just stays there
[19:47] <[mbm]> well, the basic concept of ram is simple, it's a capacitor that holds a value but will slowly discharge, so there's a refresh where the os will reach the current value and then write out the same value
[19:47] <[mbm]> s/reach/read/
[19:49] <tassadar_> hmmm, there is something I'd like to know - how does the NAND chips do it, that they will keep the data even without power?
[19:58] <tassadar_> well, wiki says that it is very similiar to mosfet, but can keep it's state even without power from outside, other then that, it's way to complex)
[19:59] <tassadar_> but I guess the proffesors at college will try to teach me how that all works in a few years, so for now it's okay)
[20:02] <[mbm]> flash is a bit more cokplicated; uses nand or nor gates to build latches
[20:03] <[mbm]> bank erase command will set everything to 0xff and then writes will set tue 0 bits
[20:04] <tassadar_> hmm, that is why "empty" memory on ATmega controllers is "0xFF"
[20:04] <[mbm]> the annoyance being that once a byte has a 0 it can't be changed back to a 1 without eraseing tje entire bank
[20:04] <[mbm]> yep
[20:05] <[mbm]> nand chips allow for higher capacity than nor chips but are harder to interface
[20:06] <[mbm]> so often times bootup is done using nor chips
[20:08] <[mbm]> flash is also a few orders of magnitude slower than ram
[20:13] <[mbm]> should point out that on flash, a bank or page is often 256k or higher
[20:13] <tassadar_> that is really lot Oo
[20:13] <tassadar_> no wonder USB drives sux at small writes
[20:13] <[mbm]> that's why SSDs have a TRIM command
[20:15] <[mbm]> trim basically means ignore the old contents and just write the new value
[20:16] <[mbm]> so if the pages are 256k and you only want to write 4k, without trim you'd be reading all 256k into ram, erasing and then writing 4k along with the remainder 0f the original 256k
[20:18] <[mbm]> trim lets you short circuit the process and just 0verwite the full 256k with the new 4k data
[20:19] <tassadar_> when is it called, like is it instead of "delete" command, or during the write operation..?
[20:20] <[mbm]> it means the write command doesn't need to preserve the old data if a write doesn't end on a boundary
[20:21] <[mbm]> early ssds didn't support trim, so the init9al write was fast, but as the drive filled it couldn't simply write the data, it had to do a read, modify, write
[20:22] <[mbm]> so it slowed down over time because it didn't know the os didn't care about the old data
[20:23] <tassadar_> just to be clear - one block == one file?
[20:23] <[mbm]> not quite
[20:24] <tassadar_> I mean, if you have lot of small files, eventually you would have to put more of them into one flash block, which means you have to read it in order not to lose the other file
[20:25] <[mbm]> suppose you had a 16M flash chip, it's probably arranged into blocks of 64k or 128k
[20:25]  * tassadar_ has 16M flash chip
[20:25] <[mbm]> as you get 9nto larger capacities the sizes go up
[20:26] <[mbm]> under linux, the mtd driver is used when dealing with nand or nor
[20:26] <[mbm]> and if you look at /proc/mtd you'll see an erasesize
[20:27] <tassadar_> yeah, it's like that on my LG P500
[20:27] <[mbm]> that's showing you how big the blocks/pages are
[20:27] <tassadar_> it has 0x00020000, which is..
[20:27] <tassadar_> g << 00020000
[20:28] <tassadar_> oh, wait, wrong channel, there is no geordi here)
[20:28] <tassadar_> 128k
[20:29] <[mbm]> flash is rated in terms of how often you can erase and rewrite it before the silicon falls apart
[20:29] <[mbm]> which is on the order of 100,000 - 1,000,000 cycles
[20:30] <tassadar_> wiki says that when MLC chips are used (like in most android devices?), it is much less - around 1-10k
[20:31] <tassadar_> which scares me a bit
[20:31] <[mbm]> when you deal with flash chips hidden behind controllers like sd cards there is a certain amount of reserved space
[20:31] <[mbm]> and bad pages get automatically remapped
[20:32] <[mbm]> when dealing directly with the flash there are special filesystems like jffs2 and yaffs2 which try to distribute the wear
[20:34] <[mbm]> if you were to try to use ext3 directly you'd probably have some files which almost never change and some files which are constantly changing
[20:35] <[mbm]> which results in uneven wear, the areas with tue constant files would have only used a few cycles while the remainder of the filesystem was constantly using up cycles
[20:36] <[mbm]> one of the stupidest things you can do is wget a file directly to a flash filesystem
[20:37] <[mbm]> which will go slow because it has to constantly update the flash with the new file size and data
[20:37] <[mbm]> works much better to download to ram and then once downloaded, write the whole thing to flash once
[21:00] <tassadar_> [mbm]: okay, thanks for the explanation, I feel like It gave me more than the last month and a half in school :) I have to get some sleep now, gn
[21:21] <Ethernin> Hey u guys, can I ask how many of you have ubuntu installed on a Nexus7?
[21:23] <Ethernin> just curious...for those who are running ubuntu on a Nexus7 - what desktop environment has worked the best for you?
[21:23] <Ethernin> I'm finding even tweaking unity it is practically unusable