**** BEGIN LOGGING AT Fri Jun 06 03:00:00 2014 Jun 06 08:44:57 good morning Jun 06 08:47:24 hi mckoan Jun 06 08:52:25 hi kroon Jun 06 08:56:11 hi woglinde Jun 06 09:38:06 hi woglinde, all Jun 06 12:02:56 XorA, can you add meta-arndale to the layer index? Jun 06 12:18:28 hi crofton Jun 06 12:27:10 gm Jun 06 13:20:17 I need expert Jun 06 13:20:36 something broke my image creation Jun 06 13:20:59 I realize now that image creation is unique to each development... or mostly Jun 06 13:21:03 mine Jun 06 13:21:08 worked for 4 years Jun 06 13:21:14 broke on yocto 1.6 Jun 06 13:21:36 but when I look at the commands that are used to make the image, I cannot see how a version bump to 1.6 could have caused the issue Jun 06 13:21:58 check the logs Jun 06 13:22:15 the only outside program that is used is, mtools-native Jun 06 13:22:20 the logs say nothing Jun 06 13:22:26 because the image creation worked Jun 06 13:22:28 in the image workdir? Jun 06 13:22:34 yes Jun 06 13:22:41 it as 3 partitions Jun 06 13:22:48 part 1 vfat boot for uboot Jun 06 13:22:54 part 2 rootfs for all the files Jun 06 13:23:02 part 3 another vfat for something... not sure what Jun 06 13:23:08 part 1 and part 3 works Jun 06 13:23:12 part 2 wont mount Jun 06 13:23:15 not on the device Jun 06 13:23:23 not when I plug it directly into linux Jun 06 13:23:27 further more Jun 06 13:23:31 in my image folder Jun 06 13:23:35 I have bunch of files Jun 06 13:23:40 .ext4 mounts fine Jun 06 13:23:52 so bitbake and oe is doing its thing correctly Jun 06 13:23:59 so Jun 06 13:24:11 I have the rootfs in the form of a ext4 file Jun 06 13:24:21 trying to bake it into an image using dd and mtools Jun 06 13:24:26 this somehow now fails Jun 06 13:24:28 so Jun 06 13:24:39 I usually extract the tar.gz in place, on the specific partition Jun 06 13:24:48 if I do a fsdisk -d image.dsimg I get.... Jun 06 13:24:52 so you're 100% sure Jun 06 13:24:54 if I do a fsdisk -d image.sdimg I get.... Jun 06 13:25:14 sfdisk -d dev-yoco1.5-impedo-duo.sdimg Jun 06 13:25:14 # partition table of dev-yoco1.5-impedo-duo.sdimg Jun 06 13:25:14 unit: sectors Jun 06 13:25:14 dev-yoco1.5-impedo-duo.sdimg1 : start= 63, size= 128457, Id= c, bootable Jun 06 13:25:14 dev-yoco1.5-impedo-duo.sdimg2 : start= 128520, size= 4209030, Id=83 Jun 06 13:25:14 dev-yoco1.5-impedo-duo.sdimg3 : start= 4337550, size= 1028160, Id= c Jun 06 13:25:14 dev-yoco1.5-impedo-duo.sdimg4 : start= 0, size= 0, Id= 0 Jun 06 13:25:27 that file Jun 06 13:25:39 so the file is there Jun 06 13:25:46 it looks correctly prepared Jun 06 13:25:53 part1 works Jun 06 13:25:58 part3 works Jun 06 13:26:05 part2 which is my ext4 wont mount Jun 06 13:26:10 bad fstype Jun 06 13:26:17 whne I do a Jun 06 13:26:28 maybe dd those sectors out from the image to a separate file and see if you can loop mount that? Jun 06 13:26:44 losetup -o 63 /dev/loop2 image.sdimage Jun 06 13:26:48 it wont mount Jun 06 13:27:02 dmesg says: [3219134.495680] FAT-fs (loop2): bogus number of reserved sectors Jun 06 13:27:05 dmesg says: [3219134.495680] FAT-fs (loop2): bogus number of reserved sectors Jun 06 13:27:30 ok let me try paul's move Jun 06 13:29:37 there is one command that bothers me: Jun 06 13:29:41 dd if=${ROOTFS} of=${SDIMG} conv=notrunc seek=${PART2_OFFSET} bs=${SECTOR_SIZE} Jun 06 13:29:53 this is the part where it copies the ext4 into the image Jun 06 13:30:06 there is no size specified... so it takes the size that ext4 is? Jun 06 13:33:32 yes, dd will continue until the input file finishes Jun 06 13:34:26 kk Jun 06 13:34:38 pompomJuice: I would try a fresh SD Jun 06 13:34:46 omg Jun 06 13:34:52 that could be it Jun 06 13:35:10 my SD got stolen Jun 06 13:35:16 so this new one got dropped on my table Jun 06 13:35:24 let me get one Jun 06 13:35:35 this one looks old and used Jun 06 13:37:40 I already faced to 1st stage bootloader partitioning that broke the SD Jun 06 13:40:27 waitaminute Jun 06 13:40:32 why does it say yocto1.5 Jun 06 13:40:34 im on 1.6 Jun 06 14:03:39 ok... build finished let run some tests.. Jun 06 14:04:49 the yocto1.6 branch still has yocto1.5 in it's image name Jun 06 14:04:54 thats gotta be an oversight Jun 06 14:05:27 or Jun 06 14:05:35 nm I put that in Jun 06 14:05:41 * pompomJuice ducks Jun 06 14:29:49 ok Jun 06 14:29:54 that took a while Jun 06 14:30:19 dd if=${ROOTFS} of=${SDIMG} conv=notrunc seek=${PART2_OFFSET} bs=${SECTOR_SIZE} Jun 06 14:30:39 that is the command that copies the image.ext4 into my sdcard image Jun 06 14:31:14 then I did what paul said and extract it with dd if=broken.sdimg of=rootfs.ext4 skip=128520 bs=512 count=4194304 Jun 06 14:31:32 cannot mount rootfs.ext4 Jun 06 14:31:48 when I do a diff they also differ Jun 06 14:32:07 when I do hexdump they are way different Jun 06 14:32:32 makes no sense because the creation dd is basically the last command that runs. Jun 06 14:32:44 nothing afterwards can screw it up Jun 06 18:12:00 Crofton|work: done, get reviewing :-) Jun 06 18:12:52 XorA, thanks Jun 06 18:13:10 * XorA has been travelling all day Jun 06 18:19:05 Crofton|work: I expect loads of patches from you now :) Jun 06 18:21:05 I've only got that board fro mteh hard float guys Jun 06 18:21:27 but now when we drive people to the index and they search for arndale, something comes up Jun 06 23:23:09 did anyone add the brokensep thing to alsa-lib? it fixes the build problem with meta-toolchain Jun 06 23:23:59 seems like a good idea until a better option is identified Jun 06 23:28:57 and, on another note, any interesting images for wandboard besides weston? Jun 06 23:29:20 that one has video problems and a slightly old kernel Jun 06 23:29:27 *on master **** ENDING LOGGING AT Sat Jun 07 02:59:59 2014