**** BEGIN LOGGING AT Mon Jul 02 03:00:00 2018 Jul 02 06:51:30 JaMa: this week-end I was browing a bit the recipes that are current MARCHINE_ARCH, and I stumbled on luna-service2 ( https://github.com/webOS-ports/meta-webos-ports/blob/master/meta-luneui/recipes-webos/luna-service2/luna-service2.bb ) Jul 02 06:51:56 JaMa: would you happen to know what makes this recipe machaine_arch ? Jul 02 06:52:30 (ah, monday morning's typos...) Jul 02 06:54:55 oh, wait a bit... it's actually not in the ipks/mako directory... so it would be only the case for rosy ? mmmh... Jul 02 06:55:48 Herrie: we don't have any mido, onyx, athene folder in http://build.webos-ports.org/luneos-testing/ipk/, that looks like a mistake somewhere Jul 02 07:36:11 Morning! Jul 02 07:36:22 Tofe: That's because we didn't build anything on Jenkins yet for those Jul 02 07:36:30 As you can see there's not even a job for them ;) Jul 02 07:36:37 Only the Halium bits are build on Jenkins Jul 02 07:36:47 I didn't add the job yet for full image, since they're WIP a bit still Jul 02 07:39:25 It's easy to add them though Jul 02 08:46:16 Herrie|Laptop: ah, ok, yes Jul 02 08:46:51 Tofe: it's MACHINE_ARCH in webOS OSE, but shouldn't be in LuneOS Jul 02 08:50:25 JaMa: I'll double check my rosy build tonight, maybe I made a mistake Jul 02 12:17:35 Tofe: there was fancy wikipage somewhere about android symbols and gdb.. do you know link? Jul 02 12:21:21 found it Jul 02 12:27:13 bshah: I would welcome the link :p Jul 02 12:27:36 I'm using my own one-liner but it's quite a long line... Jul 02 12:29:17 (I'm using variations of this: https://bpaste.net/show/114ef0d0e59f ) Jul 02 12:30:45 (but it doesn't word that well) Jul 02 12:36:29 Tofe: https://wiki.ubuntu.com/Touch/Core/UbuntuDebugAndroid Jul 02 12:36:43 also libhybris have load_sym_files.py Jul 02 12:39:46 bshah: yes, but I always forget about it :/ Jul 02 12:45:53 Tofe: here is my backtrace : https://ptpb.pw/QtkL Jul 02 12:47:31 I'll do currentthreaddata thing soon Jul 02 12:49:38 oh, your bt doesn't point to the Qt GL thread apparently Jul 02 12:50:18 bshah: which thread triggers SIGBUS ? Jul 02 12:53:10 it might be I am missing some debug symbols although wait Jul 02 12:54:04 Tofe: can I see your backtrace? Jul 02 12:54:13 (for SIGBUS) Jul 02 12:55:49 it's Thread 1 for me Jul 02 12:57:23 bshah: it was https://bpaste.net/show/ca78b9ceac49 Jul 02 12:58:31 okay, lets see Jul 02 12:58:43 I am installing some more debugsyms Jul 02 13:00:17 ok Jul 02 13:01:28 hm for some reason thread 1 bt is crappy always Jul 02 13:15:24 I am failing at gdb Jul 02 13:16:01 Tofe: can you pastebin all commands you used? Jul 02 13:16:22 what I am doing is I set a breakpoint in set_thread_data, and then set watch on currentThreadData Jul 02 13:16:29 but I hit SIGTRAP Jul 02 13:33:27 ah you may have to set the watch on the address of currentThreadData, not the name Jul 02 13:33:58 so if ¤tThreadData == 0xbabababa, then watch *(0xbabababa) Jul 02 13:34:50 I don't have the history here, but I'll soon be able to replay this SIGBUS issue on my new rosy device, which has halium-7.1 Jul 02 13:38:56 Tofe: hm, still doing something wrong : https://paste.kde.org/ppmne1pf3 Jul 02 13:48:09 bshah: the previous watch is disabled? I had the issue once, and had to restart my gdb session Jul 02 13:48:36 yes, this is fresh session Jul 02 13:49:04 ah Jul 02 13:51:23 Tofe: some more progress, https://paste.kde.org/pfwwm3rvu Jul 02 13:51:27 but still SIGTRAP Jul 02 14:04:19 blargh Jul 02 14:04:21 I give up Jul 02 14:12:57 I have opened issue at libhybris issue tracker just for tracking : https://github.com/libhybris/libhybris/issues/391 Jul 02 14:17:52 I've inserted my 5.1 analysis; I'll give a try with 7.1 tonight Jul 02 14:51:18 Herrie: for tissot, what halium script did you use for initrd ? A patched one ? Jul 02 15:11:07 Tofe: For tissot I didn't patch anything in initrd Jul 02 15:12:48 For tissot used my modified repos because tissot-dev's were lazy/sloppy Jul 02 15:13:17 And just the changes that halium-devices setup script does Jul 02 16:37:27 Herrie: well that might be an issue; you'll need something like what is in this branch: https://github.com/Halium/initramfs-tools-halium/blob/a-b-partitioning/scripts/halium Jul 02 16:38:17 or https://github.com/Halium/initramfs-tools-halium/pull/14 might be even more mature Jul 02 16:51:23 bshah: do you have strace of your crash btw? Jul 02 16:51:54 nizovn, bshah: I'm push a non-patched qtbase to my halium-7.1 device right now Jul 02 16:52:02 pushing* Jul 02 16:52:15 strace doesn't cope well with libhybris, in general, for me Jul 02 17:08:33 oh no... can't reproduce the crash here... Jul 02 20:04:33 Interesting! My correction to init.rc replaces "write /dev/cpuset/foreground/cpus 0" with "write /dev/cpuset/foreground/cpuset.cpus 0", etc, on the grounds that cpuset.cpus is what actually exists. Jul 02 20:05:04 And now my container won't start again :-) **** ENDING LOGGING AT Tue Jul 03 03:00:00 2018