=== JanC_ is now known as JanC [00:15] Are porting questions also to be asked in this room. If yes. Does Ubuntu Touch also use init.rc? [00:17] Ubuntu Touch uses upstart (at the moment - switching to systemd is a work in progress, IIUC) [00:19] So you *can* stick things in /etc/rc.* and the right thing will happen, but you're probably better off dealing with native upstart. [00:20] So if I have to load / insert a kernel module at an early stage? [00:22] It doesn't get autoloaded? [00:22] But you can certainly write an upstart job for that. [00:22] (Or in init.d script, of course) [00:23] Well, I have a kernel module which creates block devices so I can mount my NAND filesystem. [00:23] Ah, so you'll need it to be in the initramfs. [00:23] Which I *believe* we use on Touch. [00:24] Is it possible to output kernel log to the screen if that's the case? [00:25] Hm. This must actually be in the android half of Touch, because that's where the kernel is loaded from. [00:26] Well, should be also fine. [00:26] I'm not familiar with the details of porting, but the kernel comes from Android... [00:29] I have a working Android Kernel built for my device. I tried to port CM. But init.rc (for the main system) seems to cause problems, which is why I consider porting to Ubuntu Touch first. [00:31] If I have adb at an early stage of the boot sequence I'm more than satisfied. [00:32] You've presumably seen https://developer.ubuntu.com/en/phone/devices/porting-new-device/ ? [00:32] Yes I've seen it. [00:33] Under the Debugging header: “This will open an adb shell and pause execution of the initrd,” [00:34] So, yes - you'll need to ensure your kernel module ends up in the initrd and you should be able to use a script in hooks/modules to load it. [00:34] Sounds great! [00:36] Thank you. I'll definetly try to hopefully get something to run tomorrow! [00:39] Fighter19: why is that driver not just built into the kernel? [00:43] * RAOF expects some totally appalling reason, like it's a binary blob. [01:05] dobey, because it's a propietary blob. [01:08] * Fighter19 overlooked the "me" line my RAOF . [01:08] HAH! Called it. [01:09] What manufacturer should we be pointing and laughing at? [01:11] At Rockchip ;) [01:13] Same goes for the mali modules. [01:13] Those at least you don't need in order to mount your root filesysem. [01:13] And GPUs are *somewhat* more complex than NAND interfaces. [01:14] That's true. [01:17] I remember some strange stuff had to be done to get Mali chips running under Linux. I hope it doesn't get too complicated again. And I hope they're powerful enough [01:18] Fortunately you don't need to get the Mali chips running under Ubuntu userspace; you only need to get them running under Android userspace :) [01:19] Does ubuntu access the framebuffer? :O [01:20] Only though the Android HAL. [01:21] (This is the point of mir-graphics-platform-android - given an Android graphics stack, provide an Ubuntu one on top of it) [01:21] Wait, if no OpenGLES and no FB, how is it supposed to show something? Am I missing something? xD [01:21] The magic of libhybris - Mir accesses the usual Android graphics stack to do this. [01:21] So you need working Android drivers, but once you've got them it's done*. [01:22] Nice. [01:22] So it should automatically work [01:22] *: Modulo the fact that graphics drivers are terrible and you might need to quirk things. [01:22] I'd still like to activate GLES if possible at some time. [01:22] I think Ubuntu Touch supports it? [01:23] You'll get it for free. [01:23] If you can bring Mir up on your android drivers, any client connecting to that Mir server will get GLES support. [01:24] Again, modulo the fact that graphics drivers are terrible. [01:25] I guess I'll just have to read through some documents first :D [01:28] I'm kinda surprised. Because normally the libMali.so contains the symbols for GLES. So it can be linked with the normal GLES headers. But in the end effect it accesses /dev/mali which again is created by the kernel module. [01:29] That's how the userspace component of graphics drivers generally works. [01:30] Provide the GL/GLES/whatever ABI at one end, talk to the kernel module at the other. [01:38] When you think that your Linux knowledge helps you to understand Android xD. The HAL thing was new to me, but as I understand it, it's inside the kernel and provides a unified way to access devices of the same type. Sounds easy enough, de facto. So my kernel will know what to do once the modules are loaded. [01:45] Ah no, that was BS what I just said. [01:50] Your device has a piece of code which provides the interface for the communication with the hardware (often through device nodes) for HAL. I think THIS is right. (sorry to spam the chat, sometimes I have to say out thinks "loud" :D ) [01:51] Just wondered why there was no config (putting the so inside the proper directory is enough :O ) [01:54] Thank you! Have a good night / or day, wherever you are :) [03:56] @duflu - There is a video on OMGUbuntu from April showing a Nexus 4 and a Nexdock with correct resolution - http://www.omgubuntu.co.uk/2016/04/ubuntu-touch-nexdock-video [03:58] JasonD: Cool, thanks. Maybe I/we/mir-team will end up having to order one. [03:58] Because I don't have any other way to test the problem right now [03:59] Hi, yeah I thought I had seen it working properly somewhere, could it be a regression or my slimport cable? [03:59] Happy to help if I can [03:59] JasonD: More likely a regression. Faulty cables will result in no image (which is VERY common). Do you get any image? [03:59] I've found 50% of Slimport cables are unusably faulty.... [04:00] Yes, just very difficult to read as the font is so small/fuzzy [04:00] Cool, that's something at least [04:00] This is my second one, they are not very well made it seems [04:01] If the Nexdock does advertise other resolutions I don't think mir's Android backend will report them. Maybe it's just Mir defaulting to the wrong resolution [04:04] Dunno, but if you watch the video the resolution looks correct, on mine the icons and address bar in the web browser are way smaller === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [06:08] hi [06:09] Hi [06:10] JasonD: hi [06:10] how you get on with your nexux 4 and the nexdock [06:10] etc [06:11] Still wrong resolution atm, tried it with RPi2 ok [06:11] JasonD: it's just a tv screen tough or kind of [06:12] I thik meant to adust resolutions on the devices that are dconnected to it reallhy ? [06:12] JasonD: or like with a desktop. you adjust the reesolutino in windows :d, and then it goes on the screen [06:13] A bug has been raised, I did find a video from earlier in the year showing a N4 working correctly so may be a regression [06:13] JasonD: or you have a kid across the roadcoming to your house, since was sort of freindly with your sons at the time so me and my brother [06:13] It looks like the N4 is outputting FHD regardless [06:13] JasonD: the id changes the resolutiohn to something my Dad doesn't like [06:13] and he gets all uh about it etc [06:14] all annoyed about it, and I think may have even gone across the road to get cahngedback cn't remeber [06:14] oh 90's with Windows h eh [06:14] 1990 's [06:14] good times? [06:14] h eh [06:14] Should work like this: http://www.omgubuntu.co.uk/2016/04/ubuntu-touch-nexdock-video [06:15] JasonD: the Nexus 4 isn't FHD though I think ? [06:15] Rebooting after every change? lol [06:15] what in Windows? [06:15] It outputs FHD via the slimport [06:15] yes [06:15] yes inded re booting after lots of things that' Windws [06:16] brb, of to lunch [06:16] JasonD: ok [08:28] First of all, my original aosp tree builds and runs. My ubuntu tree instead, fails with CAPEWrapper.cpp:16: error: undefined reference to '__xlog_buf_printf' [08:28] that function is declared inside cutils/xlog.h, but the only file where it could be defined is a binary inside vendor/$something/libs/* [08:28] I added vendor/$something/libs to main.mk, but my build still complains about that definition missing [10:01] dobey: did you ask about phablet-tools yesterday because it broke during update? [10:01] (it did here) [11:21] popey, seems he's off this whole week [11:36] jgdx: ta === dpm_ is now known as dpm [11:50] popey: what broke? [11:52] davmor2: nvm :) === _salem is now known as salem_ [13:03] “Device krillin not found on server https://system-image.ubuntu.com channel ubuntu-touch/stable” [13:09] Should I be using bq-aquaris.en instead? [13:10] But that was last updated in July… [13:12] mpt: that sounds about right you are looking at stable [13:12] mpt: ubuntu-touch/rc/bq-aquaris.en contains the ota we are looking to release if that is what you are after [13:14] thanks davmor2 [14:34] Can I use a CM tree as reference to initialize the phablet tree? [14:34] Or does this only work with an AOSP tree? [14:41] Which branch is considered stable? :O [14:41] phablet-5.1.1? === dandrader is now known as dandrader|afk [15:34] Fighter19: ubports tree is based on CM === chihchun is now known as chihchun_afk === dandrader|afk is now known as dandrader === JanC is now known as Guest38270 === JanC_ is now known as JanC === mcphail_ is now known as mcphail [17:01] Got an error building the recovery executable. Inside roots.cpp cryptfs has not been found [17:03] (I'm on the phablet-5.1.1_r36 branch) === raj` is now known as raj === dandrader is now known as dandrader|afk [19:23] My system partition is only 100 MB big, can this be right? :O [19:23] Regarding what has been built. [19:33] hello. Is it safe, (easy?) to buy a meizu mx4 with flyme os and install ubuntu touch myself ? thank [19:39] jonathan___, it's possible to install it the same way as on ported devices [19:54] Does any one here have an idea how I debug the startup on a new device. I can boot the created recovery partition without problems, but if the same kernel tries to boot the ramdisk from boot it fails. [19:55] I even tried to set break=top in the CMDLINE. It even shows up in the recovery so that can't be the problem. === dandrader|afk is now known as dandrader [23:06] Looks like my device only likes binary inits.....