=== zz_chihchun is now known as chihchun === heathkid|2 is now known as heathkid [03:51] hi all! My Nexus 7 has been flashing for the past hour. [03:51] not a good sign? [03:51] 16GB model. [04:02] i see i'm not the only one experiencing this - http://askubuntu.com/questions/206984/installation-on-a-nexus-7-gets-stuck-while-flashing-the-root-filesystem [04:08] jrgifford: yeah, we've seen that occasionally. only advice we have is to try again [04:08] achiang: should i try with the 8GB image? [04:08] jrgifford: you won't brick the device as long as you can reboot back into fastboot mode [04:09] jrgifford: i would try 16G one more time, then try 8G [04:09] i'll reopen that question on AU and then drop an answer saying that you need to try again... [04:09] achiang: ok, thanks! [04:10] jrgifford: good luck [04:21] achiang: tried again, it's "Preparing the root filesystem, please wait, this will take a few minutes..." :D [04:21] jrgifford: great, now you wait ~15m and it will be all good [04:21] :) [04:22] achiang: yay! :) [04:28] achiang: it works! [04:28] thanks [04:29] jrgifford: cool, have fun [04:29] will do. :) [09:24] xnox, we'll also need the simg2img (to turn sparse back into normal loop mountable images in cse you want to take a look inside) [09:26] xnox, also your clean target removes the fastboot binary (or tries at least) [09:26] *the simg2img binary === doko_ is now known as doko [10:05] ogra_: ok, will check. === chihchun is now known as zz_chihchun [10:19] ogra_: is there something wrong with remove fastboot binary? [10:19] * xnox thought we compile it. [10:22] well, it wont exist in the extras dir [10:22] ah =) [10:23] ogra-nx7: ogra_: can't seem to find simg2img [10:24] hmm, should be in the same branch in the same dir upstream [10:24] (extras) [10:24] there is "-S don't use sparse output format" in the ext2simg.c [10:25] thats not the same [10:25] ok. [10:25] -S is used to create non sparse images [10:25] i.e. only at creation time [10:26] https://code.google.com/p/usefulshellscript/source/browse/trunk/simg2img.py ? [10:26] simg2img actually turns a existing img into a mountable nonsparse one [10:27] hmm... found an android one as well. [10:27] https://gitorious.org/0xdroid/system_extras/blobs/9c842adc177c1bcd22c2038d8d237bfb70654dca/ext4_utils/simg2img.c [10:27] gimme a few minutes, i got a branch on my PC upstairs [10:27] but that's not android official repo? [10:27] ah, yeah, that looks good [10:28] there must be an android one as well [10:32] ogra-nx7: so the jelly-bean branch has img2simg and simg2img but not the master branch. [10:34] aha [10:35] and the existing package uses master ? [10:37] ogra-nx7: it was moved from extras ext4_utils to core/libsparse.... [10:37] found it, will compile. [10:37] yay, hide and seek with code [10:38] * xnox ponders about continue to monkey compile or actually compile libsparse into a separate lib. [11:11] ogra android_reboot.c ?! =) [11:12] but that wants linux/reboot.h and I am not sure if I need the android kernel header for that, or just any local one. [11:12] ugh [11:13] xnox, where is that needed ? [11:14] ogra_: well, QA might like that... won't they? [11:14] fastboot can reboot devices [11:15] does android_reboot.c buy us anything extra ? [11:15] not really. [11:15] unless they want to set up autotests with android for comparison reasons i doubt QA will use it [11:16] * ogra_ already feared that was some kind of dep of simg2img [11:16] setup_fs wants it for some reason. [11:16] but I'm not sure we need setup_fs. [11:17] i dont think so [11:17] if it pops up we can compile it =) [11:17] :) [11:17] yeah [11:20] janimo, do we have a bug against linux-firmware to include our wlan/BT blobs ? [11:44] ogra_: there are now following binaries packaged: ext2simg ext4fixup img2simg make_ext4fs mkuserimg simg2img simg2simg simg_dump test_ext4fixup [11:44] xnox, awesome, i think thats more than enough :) [11:45] ogra_: I pushed the branch. I'll try to repack an image and flash it, to see if it works. [11:45] * ogra_ pulls [11:46] builds fine here [11:50] ogra_ hi [11:50] hey victorp [11:50] nice blogpost ! [11:50] ogra_ thanks [11:50] I am working in something else at the moment - that is why I asked you about the battery yesterday [11:50] i fear firefox simply doesnt use GLES while webkit in chromium does [11:50] ogra_ could well be [11:50] am I right that cat /sys/class/backlight/pwm-backlight/brightness [11:51] will give me the screen brightness level? [11:51] i talked to chris about it but it seems the GLES support in FF would only get us WebGL, not faster rendering [11:51] should, yes [11:51] ok.. I think I found a bug [11:51] ma [11:51] maybe [11:52] do you know what the "Dim screen to save power" is suppose to do? [11:53] dimming after a defined time [11:53] ogra_^ [11:53] running on battery? [11:53] you can define it in the power settings for battery and AC modes [11:53] there is another bug with the slider in there [11:54] it allows you to actually actively set the current brightness ... and doesnt catch if you are setting it to 0 [11:54] cant find the power settings you mention.. [11:54] so you can end up with a black screen ... havent reported that one yet [11:54] nice [11:54] iirc it is called brioghtness or so [11:55] ah, brightness and lock [11:55] yes, that one only gives me a tick box [11:56] oh, yeah, seems there were options dropped [11:56] there used to be a monitor or display tab in the power settings that allowed finer grained management [11:56] I run a test overnight and brightness was dimmed from 99 to 70 something [11:57] almost immediately [11:57] but I kept printing the file above and every hour it seems that it bounce back to 99 [11:57] and then back down [11:58] how even it never when lower than 70+ even when the battery was at 10% [11:58] yeah,. definitely some bug [11:58] I think it would make sense to make the dimming a function of the battery level dont you think? [11:58] we also dont have a working ambient light sensor yet, that might influence the behavior [11:59] that might get annoying if it dims more and more based on the battery level [11:59] well, assuming that the room is constantly at the same light [11:59] I doubt you would notice it [11:59] :) [11:59] if you only do that for the nexus you can easily pick the level thats set ... but if yxou want to make that a generic function you end up with funny stuff [11:59] you should be able to turn it off too [12:00] i.e. on my ac100 the battery icon turns red at 20% [12:00] ah [12:00] I see [12:00] that still is like 3h of worktime [12:00] on my x86 lappie 20% means find a wallplug in the next hour [12:00] ack [12:01] it seems that battery sensors are not thought for ARM [12:01] the prob here is that percentages are used [12:01] only x86 high consuming beasts [12:01] there are finer grained values that could be used [12:01] (see what upower -d gives you) [12:02] well that could be easily tuned if systems are pre-installed. I still think that reducing brightness with battery is not a bad idea... I wonder to whom I should suggest it. raise a bug? [12:02] yep [12:03] oooh [12:03] start with filing against gnome-power-manager [12:03] upower -d is cool [12:03] :) [12:03] only with janimo's latest kernel though :) [12:03] seriously , I wish there was a wiki for this sort of tips [12:04] the original one we had didnt have all the "energy" values filled [12:04] btw - I notice that my systems was prompting me to install updates [12:04] yeah, there were PPA updates [12:04] is that part of the newark ppa or should I not install them [12:04] check in sw-sources [12:05] ok, running update now [12:05] make sure ypou didnt accidentially enable updates/security/proposed [12:05] * victorp loves to ssh to it [12:05] if thats the case, updating is safe [12:05] * ogra_ got pretty good with onboard already [12:05] btw I just got this... works well with the n7 http://www.amazon.co.uk/PadStand-Black-Portable-integrated-charging/dp/B003QGPAHS/ref=sr_1_1?ie=UTF8&qid=1352203517&sr=8-1 [12:05] forcing yourself to only use the device this way for a day or two really helps [12:06] neat [12:06] ogra_ I am waiting to get my usb keyboard for that [12:06] * ogra_ has http://www.amazon.de/Twelve-South-Compass-Aluminium-St%C3%A4nder-Apple/dp/B003XQMUVA [12:07] and a really cool BT keyboard i still cant use :) [12:07] * victorp does not believe on BT [12:07] very nice! [12:07] well, i like to charge the device while its on the stand [12:08] I was going to ask about that. I thought achiang said that janimo had a patch to charge whilst in usb OTG mode [12:08] yeah, havent tested it yet [12:09] i currently run the latest though with some self added test modifications, i probably should try it [12:09] hmm, compiling the composite gadget into the kernel wasnt a clever idea it seems [12:10] makes both network managers freak out if i plug it into my PC [12:10] they both try to automatically establish a usbnet connection [12:10] sings of modern days.. I dont have a USB to USB cable [12:10] * ogra_ makes a note to make that a module next time [12:10] I have millions of USB to micro USB [12:10] you have the charger cable [12:10] so I can test it either [12:11] which goes to micro [12:11] which is fine for the usbnetwork stuff :) [12:11] if you ten enable port forwarding on your PC you have actual wired network on the nexus ;) [12:12] *then [12:12] ogra_ what I wanted to do is to plug keyboard+mouse+charger to the same USB hub [12:12] right, or use a powered hub [12:13] there should be male-male cables though [12:13] indeed [12:13] I only have one but is hooked to my pc [12:13] i know there are female-female ones [12:13] I suppose I could swap them around [12:14] lemme play with my hbus [12:14] hubs [12:15] naah, is not charging [12:16] how do I know if I have the right kernel? [12:16] janimo, ^^^ should that work already ? [12:16] victorp, i think it should show a -7 in uname -r [12:17] (my homebrewed one shows 3.1.10-7-nexus7 ... should be the same for the PPA one since they come from the same tree) [12:17] it's a 6 [12:17] maybe needs a reboot? [12:18] yeah, if t was updated it should tell you about it though [12:18] i.e. turn the gear icon in the panel red [12:18] yes, it doesnt [12:18] might be because jain uses the linaro packaging [12:18] for the kernel package [12:18] checking that the ppa is in the sources [12:19] look in /boot [12:19] if there is a vmlinuz-3.1.10-7-nexus7 [12:19] nope [12:19] ppa is in the sources [12:20] and update manager says my system is up to date [12:20] :( [12:20] aha [12:20] the uploads happened to the private PPA [12:20] I guess we should get janimo to confirm that 7 has been pushed out [12:21] ?? [12:21] it is on the new image but someone needs to copy it over to the public ppa [12:21] shouldnt they go to the normal ppa [12:21] the image is buiult from the private one [12:21] seems like we should make that part of the process [12:21] so uploading there first and then copying is required i think... not sure [12:22] nah, we should immediately start uploading to raring instead :) [12:22] once there is a raring image :) [12:22] look forward to that [12:23] in the meantime, If you can make the honours I can test the patch [12:26] victorp, we need to clarify where to ship firmware before working raring images I guess [12:26] janimo, just file a bug and leave that bit to the kernel team [12:27] janimo, any objections if i pocket copy the 7.10 kernel to the public PPA ? [12:27] i think its on the images already [12:27] and indeed we should start using the public PPA for image builds while we still do 12.10 based spins [12:27] so should be safe to do [12:27] ogra_, no objections, thanks [12:27] * ogra_ does it then [12:28] along weith the tegra driver fix [12:28] ogra_, I was waiting for people with OTG cables to see if it works [12:28] janimo, that is fine. The discussion came because the public ppas had not been update..and we got to raring images :) [12:28] and whether it causes other USB related regressions [12:30] victorp, apt-get update should give you a new tegra driver and a new kernel now [12:30] * ogra_ goes for coffee [12:41] ogra_, done - still nothing in /boot [12:42] you ran update-manager afterwards ? [12:42] apt-get update only updated the lists of available packages [12:43] apt-get dist-upgrade is the equivalent of what update-manager does then [12:44] * ogra_ definitely sees the new packages in https://launchpad.net/~ubuntu-nexus7/+archive/ppa [12:44] 16.0-0ubuntu3 of the tegra driver and 3.1.10.7.10 of the kernel [12:50] ogra_, do we need an externally powered USB hub with the nexus? [12:50] I plugged a keyboard in and nothing happens [12:50] just plugging in the kbd should work [12:50] directly i mean [12:50] i only have used the device with powered hubs yet [12:50] when i used a hub [12:55] janimo, why iso9660 ? [12:55] ogra_, ? [12:55] [ Jani Monoses ] [12:55] * [Config] Enable ISO_9660_FS [12:55] * [Config] Various options needed for LXC support. [12:55] * [Config] Enable SND_USB_AUDIO, disable SND_HDA_INTEL [12:56] i understand the last two ... [12:56] but am wondering why we enable iso9660 [12:56] ogra_, ah. There was a bug about some USB stick not mounting [12:56] it can be used on non-CD media too [12:57] oh,. indeed [12:57] maybe even our ISO hybrids are like that [12:57] yeah, you can just dd them to USB [12:57] not that that makes isolinux usable on arm though :) [12:57] I should have added bug numbers to the changelog, but at that point they were private still [12:57] but yeah, i get the point === Ursinha_ is now known as Ursinha [13:00] ogra_ running update manager now [13:02] yep, it just popped up here as well, offering me the new driver [13:02] seems propagation takes a while in the PPA [13:02] even though the copying is instantly done [13:03] ogra_ is it normal that I dont find the update-manager but instead I have a software-updater ? [13:03] yeah, i'm just stuck with old naming in my head [13:04] the binary is still called update-manager though [13:06] well the kernel is still 6 [13:07] eek [13:08] janimo, looks like we dont have linux-nexus7 in the image [13:08] i only see the linux-image files installed, not the meta [13:08] ogra_, right, the metapackage is new and did not get the chance to tell vanhoof to update the seed [13:08] ah., k [13:09] victorp, so for now that means manual updating ... actually for everyone who used the old (and todays) image [13:09] it will only start to work once the meta package is preinstalled [13:09] ol [13:09] ok [13:09] so do I need to download the .deb [13:09] and install them manually? [13:09] you can work around it by installing linux-nexus7 i think [13:10] manually [13:10] that will pull in the metapackage alongside [13:10] yeah, that works [13:13] janimo, was the wlan firmware file fw_bcmdhd.bin ? [13:13] * ogra_ cant remember and tries to find it [13:13] ogra-nx7, yes [13:19] janimo, bug 1075549 [13:19] Launchpad bug 1075549 in linux-firmware (Ubuntu) "please include fw_bcmdhd.bin and bcm4330.hcd in linux-firmware for support of the nexus7" [Undecided,New] https://launchpad.net/bugs/1075549 [13:20] ogra_ doing that now.. can we add it to the wiki pls [13:20] lets leave the rest to the kernel team, i'm sure they have a process for checking the legal stuff etc [13:20] janimo, might be helpful if someone else than me could set it to confirmed [13:20] ogra_, ok. I am fine whichever way it is solved :) [13:20] ogra_ the raring image will have to have the same disclaimer that you get from the installer [13:21] * janimo wonders what broke his chromium that most pages are black backgrounds and not much else [13:21] victorp, then we cant do it in the distro [13:22] victorp, an EULA dialog on first boot or something less intrusive? [13:22] first boot wont help [13:22] ogra_, you're right [13:22] can't access the net [13:22] it will have to happen before you even download anything [13:23] and thats something we have a policy against in the distro ... but we ship a ton of other broadcom firmware, i'm sure there is a way [13:23] (to include it properly) [13:23] I'd also want to see if we can have it legally shipped before spending too much time on technical workarounds [13:23] lets leave it to the experts dealing with firmwares every day ;) [13:24] right. if kernel team says no, we can think about workarounds like stealing it from the /system partition [13:24] ogra_ -7 now running [13:24] but i'm confident we can put it into linux-firmware [13:24] about to plug the usb hub, hold your breath [13:24] victorp, great [13:25] janimo, how about a kernel upload to raring ? [13:26] ogra_, I guess I could. [13:26] do it :) [13:26] If 7.10 is ok for all, I could upload this as a start [13:26] no USB regressions then? [13:26] NEW will probably take a while so we can be sure the binary is there once we start image builds [13:26] someone tested it last week and said he could not see the kbd [13:27] not so important for raring i guess [13:27] ogra_, does it count as working if OTG is now completly broken? [13:27] i need smoe package for the builds [13:27] could even be empty ;) [13:27] victorp, not really [13:27] i thought so [13:27] :) [13:27] victorp, how did OTG regress? [13:27] mouse not working [13:27] probably the same issue I got reported before [13:27] neither kbd? [13:27] does not detect anything [13:27] and they used to work in 6 ? [13:27] nope [13:28] or my usb stick [13:28] yeap [13:28] all of them but charing [13:28] charging [13:28] and it is still not charging [13:28] ok, seems like a serious issue. I need to rever that charging patch, I think that broke it [13:28] well I am going to un install this kernel; [13:28] do we do testing?? problably not achiang :) [13:29] janimo, looking in dmesg i see it enmables the HUB [13:30] (tegra-otg) [13:30] but it doesnt seem to see any devices [13:30] so better revert that one :) [13:30] I am plugging the same hub that with -6 [13:30] with -6 the keyboard, mouse and usb stick where all recognised [13:31] i plugged in various things here [13:31] nothing is recognized ... but the cable properly loads/unloads tegra-otg when plugging it [13:31] with -7, not even the mouse plugin is working [13:31] yep [13:31] same here [13:35] ogra_, how do I get rid of it [13:35] uninstalling linux-nexus7 ? [13:35] well, that and the linux-image-*-7-* package [13:36] * ogra_ does his first upload to raring [13:36] * janimo did his 5 min ago [13:36] *sniff* starting my release with a binary blob [13:43] there we go, the bug is "inprogress" :) [13:48] ogra_ I removed that package too but still running -7 [13:49] victorp, oh, flash-kernel cant downgrade kernels (debian removed that feature in the latest version and i didnt get around to write a fix for that yet) [13:50] victorp, sudo abootimg -u /dev/mmcblk0p2 -k /boot/vmlinuz-3.1.10-6-nexus7 -r /boot/initrd.img-3.1.10-6-nexus7 && sudo reboot [13:50] ogra_ soooo [13:50] do I need to re-flash the image? [13:51] that will get you back to -6 [13:51] * victorp swearing [13:52] well, its just a bit longer than sudo flash-kernel, dont swear [13:53] too late :) [13:53] (thats actually the line f-k would call *if* there wouldnt be a version check before that) [13:54] flasing also loses anything I set up [13:54] ?? [13:54] what do you talk about ? [13:54] packages? [13:55] you should only write the old kernel to your boot partition with the command i gave you above [13:55] victorp, sudo abootimg -u /dev/mmcblk0p2 -k /boot/vmlinuz-3.1.10-6-nexus7 -r /boot/initrd.img-3.1.10-6-nexus7 && sudo reboot [13:55] from your running nexus [13:55] no need to re-flash the device or anything [13:55] ogra_, in flash-kernel mmcblk0 is equivalent to mmcblk0p1 in our case? [13:56] ogra_ I thought that is what you were telling me [13:56] nahh+ [13:56] you never need to install a machine more than once with an ubuntu system :) there is always a way to repaitr it [13:56] janimo, no, mmcblk0p1 would [13:56] let me try that command [13:57] victorp, dont get the device name wrong :) [13:57] the mouse is alive!!! [13:57] great [13:57] ogra_ I copy and paste it [13:58] see, dont get desparate :) re-installing is for QA junkies :P [13:58] * victorp cries [13:59] * ogra_ actually has machines around that were installed on warty and upgraded regulary ... they still run without issues [13:59] ogra_, hmm, flash-kernel would need quite some changes to support alternate partitions [13:59] ogra_, would specifying the boot part directly without having to scan for it break things? Is it not fixed for various devices? [13:59] janimo, well, my original ac100 code was capable .... sadly that wasnt what debian implemented in the end [14:00] as it is now we cannot easily boot from recovery and have it upgraded there [14:00] you need to edit the db [14:01] oh, wait, there is that autodetection that debian also trashed [14:02] janimo, i would say lets add a check for "Boot-Device:", if thats set we should override the autodetection with it [14:02] currently android devices only use Android-Boot-Device and do a pointless autodetection that debian even broke [14:03] janimo, but that will always be a hack until f-k actually supports overrides for the DB [14:03] whicjh it totally doesnt atm [14:04] f-k 3.0 is still in its infancy [14:07] infinity, oh, fyi, bug 1056206 [14:07] Launchpad bug 1056206 in flash-kernel (Ubuntu) "3.0~rc.4ubuntu23: cannot flash an old (previous Q) kernel" [Wishlist,Confirmed] https://launchpad.net/bugs/1056206 [14:25] janimo, oh, look, low hanging fruit coming your way in bug 1072657 [14:25] Launchpad bug 1072657 in ubuntu-nexus7 "Can't mount a nfs disk" [Undecided,New] https://launchpad.net/bugs/1072657 [14:26] ogra_, I think I added NFS at one point prompted by that bug [14:26] in 6.9 maybe [14:26] ah, its not closed though [14:26] still, I'd wait with any extra config changes will we sync up with ubuntu. No point in adding them piecemeal [14:27] ogra_, right, since only today did the new kernel become released [14:28] yep [14:35] hmm, bug 1075415 is intresting [14:35] Launchpad bug 1075415 in ubuntu-nexus7 "Touch is too sensitive: touch/tap is interpreted as a drag/swipe" [Undecided,New] https://launchpad.net/bugs/1075415 [14:35] i think in android the touchscreen ships a calibration file [14:37] yeah, there it is [14:37] https://android.googlesource.com/device/asus/grouper/+/505d795dfcfd473d2a5a2f20177d355bcd254492/elan-touchscreen.idc [14:43] ogra_, you might like this one : http://victorpalau.net/2012/11/06/ubuntu-nexus7-7-hours-battery-life-whilst-browsing/ [14:45] sweet [14:45] 7h is great for the fact that we havent touched power mgmt at all [14:46] ogra_: so reading scrollback... did we upload a broken kernel to our PPA? [14:46] achiang, well, not having the meta nobody will get it offered anyway [14:46] achiang, so no worries [14:47] but yes, seems we did [14:47] well, i'm a bit worried about what happened to our qa process [14:47] as in, we didn't have one [14:47] ogra_ pretty good I reckon [14:47] imagine if screen brightness was adjusted :) [14:48] achiang, well, all i saw was that the new kernel ended up in the new image, i should probably not have done the pocket copying but it felt inconsistent to have it in the new image but not in the PPA [14:48] achiang, I am not sure the problem was the pocket copy [14:48] victorp, right and CPUs get powered down if not needed etc [14:49]