=== amitk is now known as amitk-afk
xnoxppisati: morning. Yeah, overlay works fine here afterall =)08:29
ppisatixnox: thanks, i saw your email08:29
ppisatixnox: i'll send a pull req today08:30
ppisatixnox: thanks for testing :)08:30
xnox\o/ awesome.08:30
ppisatii realy would like to get my nexus7 back in shape08:30
xnoxppisati: well, there is a way to repack .bootimg & add preseed and wifi config with the hope for it to come up with wifi network and ssh server installed.08:33
ppisatixnox: any doc around?08:33
xnoxas i understand it's the screen brightness that is preventing you from installing?08:33
ppisatixnox: yesterday i played a bit with .img08:33
ppisatixnox: modified it, repacked and reflashed08:33
xnoxppisati: http://blog.surgut.co.uk/2013/02/flash-nexus7-like-rock-star.html08:33
ppisatixnox: but flash didn't go well08:34
xnoxthat auto packs a presseed (modify as wanted) & copies network-manager's wifi config.08:34
xnoxthis does not have a success-command to install ssh server though. I'll try to add that today for you ;-)08:34
ppisatibug 112683608:35
ubot2Launchpad bug 1126836 in ubuntu-nexus7 "The error "df: warning: cannot read table of mounted filesystems: No such file or directory" is shown. Splash screen appears, then black screen." [Wishlist,Confirmed] https://launchpad.net/bugs/112683608:35
ppisatithis is exactly my problem08:35
=== smb` is now known as smb
=== amitk-afk is now known as amitk
* smb yawns08:38
xnoxppisati: ouch, I flashed with 11th march image last, and it was fine here with a 16gb wifi model here.08:39
smbppisati, You sound like having fun with the n7... not08:42
ppisatismb: eh08:53
ppisatixnox: did you try the phablet img?08:53
xnoxppisati: yeah, that works fine as well.08:54
ppisatixnox: here too08:54
ppisatixnox: i can go back to android or to the phablet img08:54
ppisatixnox: the only one NOT working, is the ubntu desktop one :(08:54
ppisatixnox: anyhow, coffe, overlayfs pull req and than i'll play with your script08:55
apwppisati, how far does it get08:55
ppisatiapw: desktop img on nexus?08:55
apwppisati, yes08:56
ppisatiapw: dies after plymouth08:56
ppisatiapw: completely black screen08:56
ppisatiexactly like in lp112683608:56
apwppisati, so the kernel is up and working, does serial work at that point ?08:56
apwbug 1126836   08:56
ubot2Launchpad bug 1126836 in ubuntu-nexus7 "The error "df: warning: cannot read table of mounted filesystems: No such file or directory" is shown. Splash screen appears, then black screen." [Wishlist,Confirmed] https://launchpad.net/bugs/112683608:56
smbapw, or even ssh (remember the fb fun)08:57
ppisatiapw: yes, but there's no user defined08:57
ppisatiapw: that's why i need to roll a custom img and flash it08:57
ppisatiapw: it's still an unconfigured ubuntu installation08:57
apwdidn't rtg complain about the same thing, and say that 1/10 reboots it worked and then he could customise it, and then it was fine ?08:57
ppisatiapw: dunno08:58
apwworth a try08:58
apwsmb, yeah it does feel a little like that, but i don't think we have an old framebuffer in these devices to tickle the issue08:59
smbapw, true. just something about the symptoms. though clearly can be anything else after and in the bug report ppisati mentions the title sounds like the fs is a problem and mountall09:00
* smb tries to understand the sentence he just wrote09:01
ppisatiback in a bit09:57
apwppisati, ok i have my n7 up and working with the latest kernel11:26
apwppisati, on the ubuntu desktop image11:26
ppisatiapw: nice11:27
apwso it must be an installation issue i think11:28
apwwhich fits with rtg's experience11:28
ogra_whats rtg's experience ?11:31
ogra_apw, ^^11:45
* ogra_ still looks for somoebody that can reproduce ppisati's black screen issue11:46
ppisatiogra_: rtg has it11:47
ogra_ppisati, did he also have phablet installed before ?11:50
ppisatiogra_: dunno11:50
ppisatiogra_: let's see when he awakes if he can confirm those issues11:51
ppisatiogra_: btw, people reported it in the past, i asked if they experience it with latest images11:51
ppisatiogra_: let's see if they respond11:51
ogra_i'll prepare an initrd for you with init=/bin/bash so you can set a rootpw 11:51
ogra_i saw your comment on the whoopsie bug, yeah11:51
ogra_i highly doubt thats related to your issue11:52
tseliotapw: do you think it would be wrong if I made kmod treat "alias $module_name off" as a "blacklist $module_name"?11:52
tseliotapw: as it would solve bug #107306211:54
ubot2Launchpad bug 1073062 in kmod (Ubuntu) "modprobe: Assertion `kmod_module_get_initstate(m) == KMOD_MODULE_BUILTIN' failed" [Undecided,Confirmed] https://launchpad.net/bugs/107306211:54
tseliotapw: I guess it wouldn't be the same...12:04
=== amitk is now known as amitk-afk
apwtseliot, hmmm, thinking12:15
apwno i don't think it would be the same at all12:15
tseliotapw: I think we should just make sure that the installation command is set to /bin/true when alias is set to off, as the old module-init-tools did12:16
apwthat might work12:16
apwwe should fix kmod ... sigh12:16
apwi have a patch here somewhere as well12:17
ogra_ppisati, use fastboot flash to flash http://people.canonical.com/~ogra/tegra/nexus7/boot-bin-bash.img ....  have a keyboard attached and set a root pw ... then use fastboot flash with http://people.canonical.com/~ogra/tegra/nexus7/boot.img and log in via serial to debug12:19
apwogra_, could we not have a default user on these images which is removed when oem-config does its thing12:20
ogra_apw, no, that woould mean lots of ugly extra hackery12:20
ogra_apw, oem-config actually drops you to a root shell if it fails so you can do everything you need ... but if you dont have a display at all thats tricky (but also a very special case ... and ut of about 20 people doing installls the last days onlly ppisati (and now rtg)  seemed to have that issue12:25
apwbjf also reported it iirc, he was the one who said "keep rebooting it" to solve it12:26
ogra_how would keep rebooting fix the display ... 12:26
* ogra_ is confused12:26
ogra_it seems nobody who only had installed the desktop image can reproduce it so i'm still blaming the phablet images until someone reports to see it who only ever had the desktop image installed12:27
ogra_but lets see what ppisati finds :)12:31
apwogra_, that seems unlikely give we zap both flashes with new stuff12:34
apwogra_, but as you say ... we'll see12:35
ogra_well, i suspect the phablet kernel or surfaceflinger get the device in a state that either our kernel or xserver dont get along with12:37
apwogra_, even though we go through a flash and full reboot?  i am sceptical12:37
apwi would love to be able to blame phablet obviously12:38
ogra_are you 100% the device is completely been powered off ?12:38
* ppisati is reflashin with today's img12:38
ogra_a reboot will likely not clear everything 12:39
ppisatiand will try to reboot 10 times in a row12:39
ogra_ppisati, well, it would have been nice if yoou used the above boot.img's to debug the issue with the broken image you had ... 12:39
apwis there any way to pwoer down that device if you cannot boot it12:39
ogra_zep, from the bootloader12:40
ogra_*yep even12:40
ogra_phablet doesnt power it down completely 12:40
ogra_even with reboot -p it will still drain the battery as if it fully runs (see the release notes)12:40
ogra_android :)12:43
smblets take the battery out for 5 minutes... oh wait ...12:50
* ogra_ hands smb a hammer12:51
ogra_nothing we couldnt fix ;)12:51
smbogra_, Luckily I am blessed with _no_ pahblet12:51
=== amitk-afk is now known as amitk
tseliotapw: the solution to the problem is much much easier than I thought13:44
apwtseliot, oh, which is13:47
tseliotapw: modprobe looks for the module (which in this case is aliased as "off") in /sys and raises an error if it doesn't find it13:49
tseliotapw: and I wouldn't expect to find anything like "off" in /sys...13:50
tseliotapw: assert(kmod_module_get_initstate(m) == KMOD_MODULE_BUILTIN);13:52
tseliotapw: and we shouldn't do this if the alias is "off"13:52
tseliotthat's it13:52
apwtseliot, right the fix for kmod is pretty simply i think, we should do it for 'null' and 'off'13:54
apwas those are both official aliases for not doing anything13:54
tseliotapw: ah, null? Where is it documented?13:54
tseliotapw: ah, here: http://ccrma.stanford.edu/planetccrma/man/man5/modules.conf.5.html13:56
tseliotapw: ok, let me rework my patch and then I'll send it upstream to debian13:56
tseliotapw: shall I do it for things like "alias iso9660 isofs" too?14:06
bjfapw, since we respun quantal, we need a new lowlatency spun14:18
bjfppisati, same for ti-omap414:18
ppisatibjf: ack14:18
zequenceapw: Just a reminder: new quantal kernel in ppa:ubuntustudio-kernel/linux-lowlatency-sru14:23
apwzequence, thanks14:24
apwtseliot, great thanks14:25
ppisatiogra_: both in userdata, right?14:29
ogra_ppisati, no, boot14:32
ogra_fastboot flash boot boot-bin-bash.img14:33
* ogasawara back in 2014:39
rtg_apw, Linus is pushing Al to include overlayfs14:49
apwrtg_, yeah i saw indeed.  al seemed to be saying bring them alll on it14:49
rtg_aufs includes :)14:50
rtg_though I don't think _that_ will happen soon14:50
apwno i suspect there was more than a little irony in there14:52
rtg_apw, ah, dentry_open is a new handler. That is why I couldn't figure it out yesterday, I thought it had morphed into atomic_open14:53
apwrtg_, they did add atomic_open but it has odd semantics14:55
ppisatiogra_: (WW) TEGRA(0): LVDS-1: Error querying display modes: No such device.15:15
ogra_no prob15:15
ppisatiogra_: it's the only warning i see in X15:15
ogra_thats normal 15:15
ppisatiX is up&running15:15
ppisatistill i got a blackscreem15:15
ogra_sudo service luxd stop15:15
ogra_(you still dont have a lamp in your house i suppose ?)15:15
ppisati[compiz] <defunct>15:16
ogra_aha !15:17
* ogra_ wonders if italians go to bed early then ... no lamps at all ... scary15:17
ppisatiogra_: not really, but i don't need lamps :)15:18
ogra_so try something like rm /etc/init/luxd.conf15:19
ogra_then reboot and see if that changes anything15:19
ogra_(so we can rule out the backlight from the issue)15:21
ppisatiogra_: guess what? :()15:22
ppisatiogra_: :)15:22
ogra_works ?15:23
ppisatiyes :)15:23
ogra_so that means after phablet was installed 0 is actually 015:24
ogra_while with our kernel 0 is actually 40% or so15:24
ogra_(for the brightness)15:24
ogra_so something redefines that value in a weird way15:24
* ogra_ will drop luxd for a start ... but thats a bug we need to inspect deeper15:25
ogra_definitely something between the two kernels15:25
ppisatimv /etc/init/luxd.conf /etc/init/foorbar.conf15:27
ppisatiand it worked15:27
ppisatii shall try to put it back15:27
ppisatiand see what happens15:27
ogra_that means the 0 value is different after you installed phablet15:27
ogra_or the ambient sensor doesnt work anymore 15:27
rtg_ogra_, is there a config option for the ambient sensor driver ? perhaps it isn't enabled.15:28
ogra_well, it definitely is, else luxd would never have worked15:28
ogra_luxd is a dumb script that just shortcuts the two devices15:28
ogra_ambient -> some computation  -> brightness15:29
jamonany BTRFS folks around? I have an impossible to rm file on a BTRFS volume15:30
ogra_rtg_, http://paste.ubuntu.com/5610989/ 15:31
jamonbtrfsck sees the file and references to others that were deleted, but this one won't go15:32
ogra_rtg_, if you never used the phablet image that works just fine 15:32
rtg_ogra_, yeah, I've yet to install it.15:32
ogra_i assume after you use the phablet image /sys/bus/i2c/drivers/al3010/2-001c/show_lux returns 0 15:32
ogra_constantly 15:32
ogra_(just a theroy though)15:33
ogra_i would suspect there is a firmware in android that triggers this 15:33
rtg_ogra_, lemme look at teh driver to see if it requests firmware15:34
ogra_since italians have no lamps its hard to get some values from ppisati though ... to see if the value changes if you apply some bright light 15:35
rtg_ogra_, maybe we don't pay him enough to own a lamp ?15:36
ogra_ah that might be it 15:37
ogra_should i mail leann to ask about giving him a $10 bonus to buy a desk lamp ?15:37
rtg_ogra_, no firmware, but there is a calibration file: "/data/lightsensor/AL3010_Config.ini"15:38
ogra_so it gets recalibrated then15:38
rtg_looks like it scans a single integer, 'sscanf(buf,"%d\n", &calibration_value);'15:39
rtg_ppisati, maybe you can disk around with that ? See calibration_regs in drivers/hwmon/al3010.c15:41
ppisatirtg_: i'll do15:43
jamonis there anywhere else i should go to report/debug issues with BTRFS?15:44
rtg_ppisati, though it looks like the defaults ought to work.15:44
ppisatiwell, i can reinstall phablet15:44
ppisatisee what it reportas15:44
ppisatireinstall desktop and compare15:45
rtg_ppisati, you can get a console under android ?15:45
ppisatirtg_: yes15:45
xnoxjamon: #btrfs ?15:45
ppisatirtg_: sshd and i cant access / from there15:45
ogra_rtg_, yup, adb :)15:45
ogra_or you can install ssh ... 15:46
xnoxjamon: their mailing list is very good for low level debugging and fixing stuff. See their homepage for details.15:46
jamonthat's an idea, hope they don't send me back here, but thanks, will take the slow and deliberate debugging approach15:46
jsalisburybjf, fyi, I'm chatting on #is about our broken bots/scripts on cranberry.  Seems they did some python updates this morning.15:53
* ppisati -> gym15:54
ppisatiback later15:54
apwjsalisbury, yay for updates16:41
jsalisburyapw, \o/16:41
* rtg_ -> lunch17:25
apwzequence, this package seems to have been built without the .orig.tar.gz17:32
apwis that intentional17:32
apwzequence, 25.25 seems have been with17:32
apwzequence, plus when you build the source one should use -v<version in -release/-updates> to get the full changelog17:37
apwzequence, yeah -26.26 also used and orig.  so perhaps we should rebuild a -26.28 which is built using the orig and the -v3.5.0-17.1817:38
zequenceapw: Ok. My bad. Going to rebuild it then.17:41
apwzequence, it will upload a heck of a lot faster with an .orig as well17:46
apwzequence, let me know when the source is in as i do not need the binaries for our purposes17:47
zequenceapw: I'm wondering about how I should do with the changelog. Should I just change the upload number on the current last entry,  make a git commit and the rebuild, or do I need to create a new entry in the changelog completely?17:56
zequence..since I'm not really changing anything in the source other than the changelog17:56
apwzequence, i would normally make a new section none the less, and just put like17:57
apw * no change rebuild with .orig.17:57
zequenceok, will do17:57
apwzequence, don't forget the -v<version> on the dpkg-buildpackage -S17:59
zequenceapw: Haven't used that one before. Changes before version? Last version, or the last orig version?18:08
rtg_zequence, the last version in -updates18:08
zequenceLearning debian packaging fully is on my todo list. This is really awful18:10
apwzequence, last version in -updates indeed18:16
zequenceOk, so I need to use versione each time we do an non ABI update18:22
zequenceapw: It's in the ppa now18:23
rtg_zequence, you should do it every time you upload18:23
zequencertg_: But, if the previous upload ended up in -updates?18:24
rtg_zequence, the rule is simple. every time you package for an upload you specify the version of the package in -updates. that is what creates the diff in launchpad.18:26
manjoapw, patch upstreamed .. thanks for pointing that out 18:29
manjoapw, patch upstreamed .. and you are cced too 18:29
apwmanjo, cool18:32
manjortg_, trying to speak but mumble is not working 18:37
manjomumble sucks18:41
rtg_manjo, its more likely to be pulseaudio18:41
manjoI need to try my laptop ... might be the mic on the webcam 18:42
ohsixthankfully the desktop mixer shows sources and levels!18:46
antarusinfinity: ping19:01
infinityantarus: Sup?19:02
antarusso apt-get autoremove19:02
antarusis that supopsed to work...?19:02
antarusit wants to remove weird stuff on my system19:02
antarusand my debian developer co-workers are all like 'oh that thing is totally unsafe'19:03
infinityIt's supposed to work, yes.  Define "weird stuff".19:03
antarusit wants to remove rake19:05
antarusbut I hav epuppet and rails installed, which depend on rake19:05
antarusalso my precious linux-tools-package (that installs the kernel 'perf' binary)19:06
antarusI'm fairly sure is just a 'recommends'19:06
antarusinfinity: basically for the kernel pruner, they inquire as to why we do not just prune the packages directly, instead of marking them for autoremove19:06
infinityErm, I'd want to see the output of that.  autoremove can't remove anything that's depended on, unless it also removes the other package.19:07
infinityAnd, indeed, I see rails -> ruby-rails-2.3 -> rake here as a dependency chain.19:08
infinityNow, if rails itself was never marked manually installed (was it a dep of something else that you've since removed?), then autoremove is doing its job if it pulls out rails.19:08
antarusit is a dep of puppet19:09
infinityNo it's not.19:09
infinityAnd there's no reason it should be.19:09
antarusour puppet is extra special, probably19:09
antarusi   puppetmaster        Depends  puppetmaster-common (= 2.7.18-2gg21)19:09
antarusi A puppetmaster-common Suggests rails (>= 1.2.3-2) 19:09
antarussee ;p19:09
antarusI'm guessing we do not always trigger 'manual installs' properly19:10
antarusso we just cannot trust apt-get autoremove at all19:10
infinityPossibly not.  Shame.19:10
* antarus has no idea how manual isntalls work19:10
infinityWell, you could duplicate the same logic in the autoremove-kernels snippet to just do it yourself.19:10
antaruswe probably will19:11
infinityThings are marked manually installed if you either manually install them (say, apt-get install foo) or if they're a direct dependency of something in the "metapackages" section.19:11
infinityAny dep pulled in lower than that is marked auto.19:11
antarusso dpkg debs are not manual?19:13
antarusbecause dpkg and apt don't chat with each other?19:13
=== chiluk` is now known as chiluk
MoPacHello; I'm looking for some help with configuration of hooks or modules when creating initrd images for my kernel.  After repairing a kernel from a live USB, the boot process "forgot" that it needed to open my LUKS container to find the boot partition.  I have restored an old kernel image that works, but I'm worried that this is going to happen again on next kernel upgrade / image build.  What do?20:05
rtg_MoPac, I think everything in /etc/modules gets bundled into the initrd.20:08
MoPacSo do I just add lines that say "dmcrypt" or "cryptsetup" or somesuch?  No need to actually direct it to a particular drive or make a script to open a given container?20:09
* rtg_ -> EOD20:09
=== kentb is now known as kentb-afk
electronplusplusHi. I'm learning unix and kernel programming. One thing the I want to do is to code a complete unix command. my question is: What kind of command, that are useful, are missing in linux?23:06

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!