**** BEGIN LOGGING AT Wed Dec 12 03:00:01 2018 Dec 12 06:46:04 ant_home: you should add #include Dec 12 06:53:23 ant_home: also see this patch in meta-openembedded meta-networking/recipes-support/libldb/libldb/libldb-fix-musl-libc-conflict-type-error.patch Dec 12 06:53:32 that might be interesting for your issue Dec 12 07:07:51 khem, thanks, I'll test Dec 12 12:36:02 rburton, sigh.. qemumips64(el) kernel modules with wrong endianness..confirmed Dec 12 12:36:24 linux-yocto-4.18.14+gitAUTOINC+bf98e195a4_6d91dc6281 Dec 12 12:37:13 seems a regression Dec 12 12:40:39 ant_home: bug please Dec 12 12:40:51 bonus points if you can bisect :) Dec 12 12:41:37 yes, that's the idea Dec 12 12:41:53 I might have some traces about the good checkouts Dec 12 12:42:27 btw what's the mips bsp for? Dec 12 12:43:50 https://github.com/MIPS/meta-img Dec 12 12:43:54 did never use that Dec 12 12:44:35 ah, here all the variants Dec 12 12:45:54 with standard oe-core trying DEFAULTTUNE = "mips64r6el" gives Dec 12 12:45:56 -mips64r6el} qemumips64) for DEFAULTTUNE (mips64r6el) does not contain TUNE_PKGARCH (${MIPSPKGSFX_VARIANT_tune-mips64r6el}-nf).Toolchain tunings invalid: Dec 12 12:45:56 Tuning 'mips64r6el' has no defined features, and cannot be used. Dec 12 12:46:29 I am using generic "mips64el" Dec 12 12:48:21 note that I am absolutely not mips-expert :) Dec 12 13:44:16 rburton, hm, why is there sed -n -e "/ac_cv_c_bigendian/p in siteconfig.bbclass? Dec 12 13:44:51 ant_home: i dread to think Dec 12 13:45:06 been wondering about actually benchmarking the impact that code has Dec 12 13:45:42 its been responsible for a few complex issues in the past Dec 12 13:46:37 I am trying to grasp that CONFIG_SITE multiline... Dec 12 13:47:28 run a generated configure scripts, write a cache, and filter values we want to keep Dec 12 13:47:31 its horrible and i hate it Dec 12 13:47:42 and i'm still not convinced its a huge win these days Dec 12 13:49:20 let see if extending it works Dec 12 13:50:27 you see, I am starting to wonder whether I have really built the modules or not, the compilation did break for other reasons before and I was fixing that Dec 12 13:50:44 still open is samba for 64b Dec 12 13:50:55 (almost done) Dec 12 13:59:31 running a benchmark on with/without siteconfig now, curious if its makes a real difference to the buildstats Dec 12 14:18:54 rburton, it's a linux-yocto problem Dec 12 14:18:56 CONFIG_CPU_BIG_ENDIAN=y Dec 12 14:18:56 # CONFIG_CPU_LITTLE_ENDIAN is not set Dec 12 14:20:45 good work Dec 12 14:20:54 so we blame zedd right Dec 12 14:20:58 this is the normal qemumips defconfig, there isn't any qemumipsel machine Dec 12 14:21:09 I have tweaked DEFAULTTUNE Dec 12 14:21:36 so linux-yocto points to qemumips defconfig (be) Dec 12 14:22:51 well, now happens with qemumips64 actually Dec 12 14:23:22 * ant_home waiting for Bruce to appear Dec 12 14:27:01 pinged in the other channel for you Dec 12 14:27:25 debugging samba/cmocka.h in the mwanwhile Dec 12 14:45:04 khem, there is apparently a race in samba, it failed but succeeded on second run, to stop as expected wrt uintptr_t Dec 12 14:45:42 Unable to determine origin of type `struct cli_credentials Dec 12 14:46:20 there is a ton of warnings..hard to read the logs... Dec 12 14:46:57 it probably fails with clang Dec 12 14:53:19 khem, libldb-fix-musl-libc-conflict-type-error.patch seems malformed Dec 12 14:53:41 I applied it by hand and compilations goes on Dec 12 14:54:58 it's a miracle if it runs with so many warnings :p Dec 12 15:35:02 khem, patch sent, please test with llvm/clang Dec 12 22:02:20 khem: i did get a nice musl build image on morty Dec 12 22:02:34 i still need to push my changes Dec 12 22:03:13 * using an "extra" poky-musl distro config Dec 12 22:04:27 any recommendations for newer branches that could use some musl testing? Dec 13 00:47:37 hmm poky-tiny uses musl if you are poky user **** ENDING LOGGING AT Thu Dec 13 03:00:01 2018