**** BEGIN LOGGING AT Wed Jan 28 02:59:59 2015 Jan 28 04:07:57 * oldtopman recompiles openembedded from scratch Jan 28 08:17:33 good morning Jan 28 08:36:17 morning Jan 28 08:40:43 morning all Jan 28 09:02:27 hi bluelightning, mago__, all Jan 28 09:03:45 gm Jan 28 09:12:50 hmm, do we have fix for this ghost business in dizzy yet? Jan 28 09:13:45 i thought it was fixed in glibc 2.18? dizzy has 2.20 i think Jan 28 09:14:07 I suspect I'll get asked :) Jan 28 09:14:22 thanks Jan 28 09:15:16 (don't quote me on that ;-D) Jan 28 09:15:34 Crofton: you're into xilinx zynq, right? Jan 28 09:15:38 yeah Jan 28 09:16:01 nrossi, is also :) Jan 28 09:18:20 have you ever tried integrating a baremetal application (built with the Xilinx SDK) into a OE build? say you want to run the zynq in AMP configuration, and want the OE tree to automatically compile the baremetal app in addition to the normal linux stuff. I found the xilinx "embeddedsw" repository that seems to be the master copy of libxil (which seems appear in Xilinx SDK in various subsets, depending on your Jan 28 09:18:21 project settings). Do you know if anyone at xilinx ever tried to make recipes for these things? e.g libxil, compiling baremetal apps in an OE tree, preferably using the ARM toolchain that OE provides for arm targets? Jan 28 09:23:17 maybe it was a depressing question ;-) Jan 28 09:26:32 I think he'll be back in a bit Jan 28 10:40:07 Crofton: did you see my message before, or were you too busy rebooting after patching your desktop glibc? ;-) Jan 28 10:41:40 rebooting Jan 28 10:41:49 have you ever tried integrating a baremetal application (built with the Xilinx SDK) into a OE build? say you want to run the zynq in AMP configuration, and want the OE tree to automatically compile the baremetal app in addition to the normal linux stuff. I found the xilinx "embeddedsw" repository that seems to be the master copy of libxil (which seems appear in Xilinx SDK in various subsets, depending on your Jan 28 10:41:50 project settings). Do you know if anyone at xilinx ever tried to make recipes for these things? e.g libxil, compiling baremetal apps in an OE tree, preferably using the ARM toolchain that OE provides for arm targets? Jan 28 10:42:15 I ahven't done any bare metal stuff, but nrossi might know something Jan 28 10:42:38 he is in oz though Jan 28 10:45:31 okay Jan 28 10:47:46 Was there discussion about this on the meta-xilinx list? Jan 28 10:48:34 Crofton: not really, you and nathan have been banging heads over this embeddedsw before when you were discussing how to provide ps7_init.[c,h], but so far i've seen no discussions on actual baremetal apps. There has been a lot of AMP discussion though Jan 28 10:50:07 Crofton: i guess the easiest way to integrate a baremetal app (or amp binary) is to simply have the OE tree check out a binary from somewhere Jan 28 10:50:31 then you can use whatever tools you want (such as xilinx sdk) to build the elfs, and you don't have to deal with the mess of putting that into OE Jan 28 10:56:56 yeah Jan 28 10:56:59 not optimal Jan 28 18:41:57 Is it possible to run a command in a recipe as if it were to be run on the live system? I'd like to setup gpg by importing keys so it's all setup in the image. Jan 28 19:27:58 FileNotFound: the only real way to run a target binary on the host is either to do so at first boot (which you could arrange — ship your keys and import on first boot via the postinst), or to leverage user mode qemu to do so at rootfs construction time, but the latter is generally only used in postinst-intercepts at the moment. there's a bbclass for that Jan 28 20:01:30 kergoth: thanks! Forgot you could specify pkg_postinst_${PN}. Simply added gpg --import my_key.gpg in there. Jan 28 20:02:36 FileNotFound: you probably dont want that running at do_rootfs time unless you depend on gnupg-native, so you'll want to check for D being set and exit nonzero if set Jan 28 20:04:39 kergoth: Ahh, thanks. I noticed that in some other packages. Jan 28 20:48:39 How do I start matchbox on openembedded? Jan 28 20:51:36 There was no startx, but it could have been the corrupted image I built at the time. Jan 28 22:11:22 hrms why does git-send-email sends out two emails Jan 28 22:12:53 hms used -1 again Jan 29 02:03:54 nrossi: You around? Jan 29 02:04:02 oldtopman: i am Jan 29 02:06:31 nrossi: I rebuilt the whole build environment today. Jan 29 02:07:01 I got rid of a media player and sdl and now the whole thing compiles with only one issue. Jan 29 02:07:07 I'm now out of space :P Jan 29 02:07:14 * nerdboy does the old-guy secret wave Jan 29 02:07:23 oldtopman: thats always the issue ;) Jan 29 02:07:45 [ 3.831889] EXT4-fs (mmcblk0p8): bad geometry: block count 162137 exceeds size of device (131072 blocks) Jan 29 02:07:49 * oldtopman waves back at nerdboy Jan 29 02:08:45 oldtopman: are you just writing the .ext3 file directly to your sd card? Jan 29 02:10:31 tape maybe? Jan 29 02:11:22 * nerdboy scrounges for the 9-track tape scavenging utility Jan 29 02:14:54 nrossi: No, there's some kind of flashall script that I use. Jan 29 02:15:25 http://hastebin.com/ocafudadoj Jan 29 02:18:23 oldtopman: oh right the edison has emmc not sd ;) Jan 29 02:18:59 Say, while you're around, what's the command to start x on oe? startx didn't work yesterday, but it may have had to do with my fubar'ed build system. Jan 29 02:19:00 oldtopman: i am guessing this file 'pft-config-edison.xml' has the partition config Jan 29 02:19:48 Actually, it's in edison-image.bb and edison.env (I think) Jan 29 02:20:06 (as per https://communities.intel.com/thread/55612 ) Jan 29 02:20:45 oldtopman: regarding startx, you are probably after the session, so https://github.com/openembedded/openembedded-core/blob/master/meta/recipes-sato/matchbox-sato/matchbox-session-sato/session Jan 29 02:21:16 Ah, there we go. Thanks :D Jan 29 02:21:39 Break for dinner, repartition, reflash, ???, profit Jan 29 02:21:43 oldtopman: and xinit is the thing that normally starts it **** ENDING LOGGING AT Thu Jan 29 02:59:59 2015