**** BEGIN LOGGING AT Thu Jun 26 02:59:58 2014 Jun 26 07:09:16 hi all (omg, him again :-)) Jun 26 07:10:24 anyone know why '-c populate_sdk' builds qemu? Jun 26 07:10:28 * rink_ only wants the SDK ... Jun 26 07:10:41 and I want that because I want the cross-compiler and sysroot Jun 26 07:34:05 Crofton: still around? Jun 26 09:01:04 is there a difference in what kernel modules RPROVIDE between rpm and ipk? seems to be -kernel-module for ipk and kernel-module- for rpm? Jun 26 09:03:12 or maybe not what they rprovide, but I have recipes for kernel modules that has rprovides like -kernel-module which works until I change from ipk to rpm when I instead need to rprovide kernel-module- Jun 26 09:03:56 feels a bit odd Jun 26 11:01:22 hi. i'm trying to understand the booting and image construction of yocto with raspberrypi bsp. Jun 26 11:02:13 firstly, what boot loader is used by default on rpi? it seems to be a proprietary boot loader...? is it possible to use u-boot, and if so how do i specify what boot loader to use? Jun 26 11:03:25 secondly, i need to create at least two more partitions other than boot and root. is there any way to do that or will i have to modify sdcard_image-rpi.bbclass? Jun 26 11:06:51 ionte: rpi needs to use a custom boot loader because - iirc - the rpi boots through the GPU Jun 26 11:08:04 rburton: ok, but still there is support for rpi in u-boot and there is a u-boot recipe in the rpi bsp ... but perhaps if first boots with the custom boot loader, and then u-boot...? Jun 26 11:08:14 guess so Jun 26 11:08:24 you've reached the limit of my rpi booting knowledge :) Jun 26 11:08:31 ;) Jun 26 11:08:36 google can probably fill in the gaps, this is well documented Jun 26 11:09:21 not easily ... :( Jun 26 11:11:37 http://raspberrypi.stackexchange.com/questions/10489/how-does-raspberry-pi-boot Jun 26 11:12:17 great! thanks! Jun 26 11:29:41 regarding u-boot on rpi: i just had to specify KERNEL_IMAGETYPE="uImage". i thought that would just change the image type, but it also replaces the boot loader ... Jun 26 11:33:17 but it fails. Unrecognized filesystem type. Jun 26 11:40:13 hello, what is the procedure to be able to create a branch on the poky-contrib git repo? Jun 26 13:27:57 ionte: It's possible to boot uboot with RPi - instead of pointing to a file that contains the kernel, point to a file that contains uboot. Instead of loading and executing the kernel, it'll load and execute uboot, which then has the responsibility of finding and loading the kernel. http://elinux.org/RPi_U-Boot Jun 26 14:48:12 Hi. Is there a way I can easily get /usr/include/linux on my target? Jun 26 14:48:23 On sysroots i mean Jun 26 14:51:30 the target and the sysroots are completely different things, can you clarify your question? Jun 26 15:01:01 zman97211: yes i imagined so, and i also got u-boot to boot (see above). however it fails to load the kernel. there's some discussion about that on the mailing list. Jun 26 15:05:00 kergoth, I meant sysroots, I'm sorry. I found the linux-libc-headers package, I think this is what I need. The problem is now that I have need a 2.6 kernel for this device, for which I have my own recipe, should I also add my own linux-libc-headers_2.6.bb to my layer and att PREFERRED_VERSION_linux-libc-headres = "2.6" ? Jun 26 15:05:31 linux-libc-headers is only used to build the toolchain, specifically the c library. Jun 26 15:09:25 Oh ok, so is there another way to make /usr/include/linux available in my MACHINE sysroot so it can be used by another package later on? Jun 26 15:13:01 kergoth, I see that the kernel sources get staged by the kernel.bbclass: https://github.com/openembedded/openembedded-core/blob/master/meta/classes/kernel.bbclass#L312 Jun 26 15:13:37 The question is if I need to write something like this in my kernel recipe or is there another way that I don't see that already exists. Jun 26 15:13:53 If I want the same thing for linux-libc-headers Jun 26 15:30:39 moin Jun 26 15:34:49 Techqua69 Jun 26 15:34:56 bah, wrong window Jun 26 16:03:59 kergoth: frantically changing passwords? Jun 26 16:04:54 heh :) glad i switched to a password manager years ago, not many sites sharing any Jun 26 16:34:02 :) Jun 26 17:03:06 ok, must updatge bugzilla Jun 26 17:09:15 is someone interested in combo-layer improvement which will first sort commits from all layers (based on CommitDate:) and then apply them in "chronological" order instead of all-from-layer1, all-from-layer2, .. ? Jun 26 17:09:53 it's not ideal for layers with merge commits, but better than nothing Jun 26 17:11:06 Totally OT, but the budgeting software You Need A Budget is on sale on steam flash sale for the next 8 hours for 75% off, $15 instead of $60. I highly recommend it Jun 26 17:18:38 interesting, looks much more powerful than my localc table :) (more than I need) Jun 26 17:20:41 in other news there's also some awesome games ;) Jun 26 17:20:55 if i had the time, a faster mac, and a controller, i'd have got batman yesterday Jun 26 17:21:06 * JaMa highly recommends Witcher2 :) Jun 26 17:21:15 sadly only one of those is fixable with £20 on amazon Jun 26 17:21:31 i have so many games that don't run on my macbook air, and i'm determined to avoid resurrecting a desktop, so they'll have to wait till i get a new mbp this fall Jun 26 17:21:33 heh Jun 26 17:22:00 I'm a little crazy about getting games running on the mac Jun 26 17:22:08 kergoth: everything here has intel gpu, so gaming is limited. my old core2 macbook pro just about managed portal though. Jun 26 17:22:12 https://gist.github.com/kergoth/f553f205704adfb6f98f Jun 26 17:22:38 populous! Jun 26 17:22:53 every game but those in broken run fine, and those in 'too slow' are stuck waiting on a faster machine Jun 26 17:22:56 but all the others are playable Jun 26 17:23:11 took forever :) Jun 26 17:23:29 rburton, my primary wic problem is I do not know what I am doing Jun 26 17:23:37 kergoth: i seem to remember populous 2 being awesome Jun 26 17:23:43 good way to find bad errors paths Jun 26 17:24:07 Crofton: my first day of using wic was mostly exceptions that I sent to tom and he added more error paths for idiots Jun 26 17:24:14 hah Jun 26 17:24:38 but now it's time to see the kiddies i think (frozen has finished) Jun 26 17:24:44 rburton: gog.com's summer sale is pretty good too, if you haven't seen it. Jun 26 17:24:58 course that doens't help the available time problem Jun 26 17:25:15 when i do find the time, there's too many to choose from, so i don't play any anyway :) Jun 26 17:25:16 kergoth: heh did you buy that Guiness Book game collection sold for $750000? :) I undestand that budgetting software now :) Jun 26 17:25:23 JaMa: hah Jun 26 17:26:54 256 bought through gog, 330 through steam, then random humble bundles and whatnot Jun 26 17:26:59 * kergoth may have a problem Jun 26 17:27:49 * JaMa has <50 and ever played maybe 25 from them, but would like to find time to play more Jun 26 17:28:27 i've been playing through the blackwell series of point and click adventures lately, pretty good Jun 26 17:28:40 most of them from steam when it was extra cheap Jun 26 17:29:02 * kergoth nods Jun 26 17:31:35 its the big sales that get you(me) Jun 26 17:31:38 summer, christmas Jun 26 17:31:42 bastards Jun 26 17:33:54 I need to find a way to change the wic error to be more helpful :) Jun 26 21:21:21 tomz, do you know anyone that has used wic to boot an arm system using u-boot? Jun 26 21:42:43 kergoth, regarding our wic talk yesterday, there is also a patch from enea around Jun 26 21:44:14 ah, whats it do? Jun 26 21:49:02 not sure yet :) Jun 26 21:49:13 http://comments.gmane.org/gmane.comp.handhelds.openembedded.core/52172 Jun 26 21:49:18 gmane is acting goofy Jun 26 21:49:44 3) Build rootfs image only from a package list (on wks file) similar with IMAGE_INSTALL from bitbake. Jun 26 21:49:44 Package dependencies will be installed by default by the package manager. Jun 26 21:50:50 that's a little weird, in that it's changing wic's functionality a fair bit, adding an entirely new feature, since in the past it's relied on bitbake doing the rootfs creation Jun 26 21:50:52 interesting, though Jun 26 21:53:49 yep Jun 26 21:54:11 I'd like to see the rotfs creation split from the u-boot stuff Jun 26 21:54:15 I need to read moew Jun 26 21:54:17 course i was just saying yesterday that i might like that functionality split out, though i meant a standalone tool :) Jun 26 21:54:33 yeah, its a little odd that they're combined that way, i'm guessing there's a reason for it, but the email doesn't discuss the details or motivation Jun 26 22:33:10 also seems like the "u-bbot patch" is more about making a ramdisk **** ENDING LOGGING AT Fri Jun 27 02:59:59 2014