**** BEGIN LOGGING AT Sat Jun 16 03:00:04 2018 Jun 16 07:25:56 khem, gm Jun 16 07:26:29 forgot the -S .. Jun 16 07:26:30 https://pastebin.com/91fsTn7g Jun 16 08:02:35 khem, moving on HEAD is now at a4c1ede... gcc: Upgrade 7.2 -> 7.3 Jun 16 11:08:52 khem, good news, this first gcc7.3 is ok Jun 16 11:09:01 readir.s: https://pastebin.com/Ew87xn8T Jun 16 11:09:36 no align traps runtime Jun 16 11:10:00 any idea about the next change to bisect? Jun 16 13:04:30 ok, jumping to HEAD is now at e176ab0... gcc: Do not use --with-linker-hash-style if LINKER_HASH_STYLE is empty Jun 16 13:07:27 2018-03-19 Jun 16 13:26:23 ''''''''''''''''' Jun 16 13:52:28 ant_home: can you check just before gcc8 got merged Jun 16 13:54:30 it was bad, I know Jun 16 13:54:40 last master was not good Jun 16 13:55:14 ok Jun 16 14:04:24 ant_home: may be try with older binutils Jun 16 14:12:11 this checkout is already 2.30 Jun 16 14:18:41 khem, is still building but I already see https://filebin.net/4f4cqr62c8qf1z2v Jun 16 14:20:21 not promising Jun 16 14:26:13 hmm Jun 16 14:26:34 khem, I am hunting through a sump pi3 build, and can't see where gfortran actually gets packaged Jun 16 14:33:55 khem, yes, align traps are there :/ Jun 16 14:34:20 so you mean, last oe-core checkout before binutils 2.30? Jun 16 14:34:36 it's before sumo Jun 16 14:35:34 2018-02-24 Jun 16 14:38:32 HEAD is now at 774e372... glibc: Update to tip of 2.26 Jun 16 14:50:42 Crofton|work: for target ? Jun 16 14:50:58 Crofton|work: include g77 g77-symlinks gfortran gfortran-symlinks Jun 16 14:51:21 ant_home: ok that means binutils are clean, Jun 16 14:51:53 ant_home: Can you check before last musl update Jun 16 14:52:15 there has been some armv5/thumb work in musl which we got into last update Jun 16 15:06:29 is anyone else seeing do_image_tar failures like tar: ./opt/ltp/testcases/open_posix_testsuite/conformance/interfaces/sigaction/sigaction_19-12.run-test: file changed as we read it ? It happens always with some ltp file Jun 16 15:13:26 seems like rburton was having this issue a year ago https://patchwork.openembedded.org/patch/141293/ Jun 16 15:52:11 khem, I am nor yet sure binutils2.30 is ok Jun 16 15:52:28 let me test it in a few mins Jun 16 15:52:53 if it is not tism then musl Jun 16 15:57:51 khem, right, they don't come with tools-sdk Jun 16 15:58:07 and just confirmed they are in buildhistory :) Jun 16 15:58:08 thanks Jun 16 16:03:43 ant_home: another option to try is to set GCCVERSION = "7.%" in local.conf on latest master and see if it work Jun 16 16:04:40 Crofton|work: yes, may be you can cook a patch to include them if fortan is enabled, although fortran is not a global feature so might not be east Jun 16 16:04:49 s/east/easy Jun 16 16:15:43 first lets see if this works Jun 16 16:15:55 trying to build scipy on target and see if it works Jun 16 16:27:30 khem, no align trap with binutils 2.29.1 Jun 16 16:27:56 it looks it is binutils 2.30 Jun 16 16:28:19 if you want, I move one commit further and see Jun 16 16:28:52 good: HEAD is now at 774e372... glibc: Update to tip of 2.26 Jun 16 16:29:18 ant_home: ok, is this using gcc8/binutils-2.29 ? Jun 16 16:29:33 no, gcc 7.3 Jun 16 16:31:36 Previous HEAD position was 774e372... glibc: Update to tip of 2.26 Jun 16 16:31:36 HEAD is now at c708506... binutils: Upgrade to 2.30 release Jun 16 16:35:20 ant_home: ok so when you use gcc 7.3/binutils-2.30 you see the problem right ? Jun 16 16:35:38 ant_home: and when you use gcc-7.3/binutils-2.30 the problem shows up Jun 16 17:44:16 khem, ah, there are files in those packages Jun 16 17:50:10 grr, I am an idiot Jun 16 18:13:06 khem, also good: HEAD is now at c708506... binutils: Upgrade to 2.30 release Jun 16 18:13:18 heh Jun 16 18:13:34 no align traps in dmesg Jun 16 18:14:38 something happened between 2018-02-24 and 2018-03-19 Jun 16 18:16:52 with gcc7.3.0 and binutils2.30 Jun 16 18:16:54 bbl **** ENDING LOGGING AT Sun Jun 17 03:00:02 2018