**** BEGIN LOGGING AT Sun Mar 26 03:00:04 2017 Mar 26 05:39:37 Has anyone successfully built an image for Gateworks Ventana SBCs using Morty? Mar 26 13:33:51 is it possible to create a directory image? Mar 26 13:35:06 I am tried to extract the rootfs every time rebuild to debug Mar 26 15:20:24 ayaka: have a look at one of the directories in the image WORKDIR? Mar 26 15:23:24 RP, you mean tmp/deploy/images// ? Mar 26 15:24:11 ayaka: no, I mean something like tmp/work/qemux86_64-poky-linux/core-image-sato/1.0-r0/rootfs/ Mar 26 15:25:04 RP, it is empty for me Mar 26 15:25:13 tmp/work/cortexa17hf-neon-poky-linux-gnueabi/rockchip-media-image/ Mar 26 15:25:32 ayaka: using rm_work? Mar 26 15:26:17 ayaka: although that path doesn't look quite right Mar 26 15:26:27 ayaka: is your machine called cortexa17hf-neon ? Mar 26 15:26:37 RP, I forgot, but I think I have not clean a package Mar 26 15:26:44 RP, it is the target machine Mar 26 15:27:06 but I don't think I have clean up the image recipes Mar 26 15:27:06 ayaka: well, assuming you haven't enabled rm_work, the rootfs should appear there Mar 26 15:28:16 I didn't do that in my local.conf Mar 26 15:36:15 RP, it seems not work for me https://paste.fedoraproject.org/paste/XW1VnJar7n2pBLVOac15BV5M1UNdIGYhyRLivL9gydE= Mar 26 15:37:14 ayaka: what are you setting MACHINE to? Mar 26 15:37:43 RP, full local.conf https://da.gd/M12Hs Mar 26 15:39:47 ayaka: so your target machine is MACHINE = "firefly-rk3288" Mar 26 15:39:55 yes it is Mar 26 15:40:35 ayaka: so look in tmp/work/refly_rk3288-poky-linux-gnueabi/rockchip-media-image/ assuming you're building the image rockchip-media-image Mar 26 15:42:54 RP, ok I found them, but they are still two directory https://paste.fedoraproject.org/paste/HdhdTkFZftPs-P3ZrSBce15M1UNdIGYhyRLivL9gydE= Mar 26 15:43:25 ayaka: one looks like the debug symbols version of the other Mar 26 15:43:58 RP, so with the gdb, I would set sysroot to the rootfs-dbg but load binary from rootfs? Mar 26 15:44:16 ayaka: probably, I don't often do that with gdb Mar 26 15:44:19 I would like to make them combine Mar 26 15:44:27 RP, I see thank you Mar 26 15:45:06 ayaka: combined is the default, you triggered separate with IMAGE_GEN_DEBUGFS = "1" Mar 26 15:45:18 ayaka: you probably want to add dbg-pkgs to IMAGE_FEATURES instead Mar 26 15:45:53 RP, but I don't want to install so much dbg packages to target Mar 26 15:46:13 so what I should do is remove that line Mar 26 15:46:18 ayaka: well, you can't have both a rootfs with everything and one without Mar 26 15:46:34 but some other build system would do that Mar 26 15:46:46 generating a debug rootfs Mar 26 15:47:01 ayaka: roofs-dbg *is* a debug rootfs Mar 26 15:47:38 RP, I know, just a little different to the other Mar 26 15:48:10 for example, the buildroot would generate a staging directory which are the unstripped files Mar 26 15:48:37 but the target directory for the stripped, they don't have a .debug directory Mar 26 15:48:54 I am ok with both of them Mar 26 15:49:55 ayaka: OE puts the debug symbols into a separate object and links it to the original. The debug symbols go into a separate package. When you enabled IMAGE_GEN_DEBUGFS is created a shadow image containing all the debug information for the main image Mar 26 15:50:14 yes I know Mar 26 15:53:17 RP, thank you anyway Mar 26 15:53:45 I also have a problem with cross gdb, I offer meet Python Exception Installation error: gdb.execute_unwinders function is missing: Mar 26 15:54:12 ayaka: this did get discussed a while ago: http://lists.openembedded.org/pipermail/openembedded-core/2016-March/118511.html - unfortunately no conclusion was reached about what the correct gdb magic is to make things work :( Mar 26 16:25:55 RP, I see thank you Mar 26 16:28:10 I found my gdb meet python problem at starting up time Python Exception No module named 'imp': Mar 26 16:28:28 but I can't find the module imp from a python package Mar 26 18:52:38 ayaka: include python-misc in your image see if that helps **** ENDING LOGGING AT Mon Mar 27 03:00:02 2017