**** BEGIN LOGGING AT Tue Nov 10 02:59:59 2015 Nov 10 05:27:01 So I submitted a patch for a new board to the mailing list a while ago. What else do I need to do to get that into the trunk and then into 15.05 CC? Nov 10 10:49:51 argh, what's the magic for getting a new profile to show up in menuconfig? Nov 10 10:57:10 * russell-- is experimenting with a target/linux/x86/64/profiles/pcengines.mk Nov 10 10:57:24 not seeing it in menuconfig Nov 10 11:15:17 ahh touch target/linux/x86/Makefile did the trick Nov 10 11:49:32 build #111 of omap is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/omap/builds/111 Nov 10 11:57:43 build #139 of cobalt is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/139 Nov 10 12:00:28 build #139 of cns21xx is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/cns21xx/builds/139 Nov 10 12:03:04 build #139 of orion is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/139 Nov 10 12:05:38 nbd r47437 trunk/package/firmware/ath10k-firmware/Makefile * ath10k-firmware: add qca6174 firmware (mostly untested) Nov 10 12:05:53 nbd r47438 trunk/ (32 files in 11 dirs) * toolchain: add support of ARC architecture Nov 10 12:05:58 nbd r47439 trunk/package/network/services/openvpn/patches/200-small_build_enable_occ.patch * openvpn: enable options consistency check even in the small build Nov 10 12:06:04 nbd r47440 trunk/rules.mk * build: add -Wno-error=unused-result to target cflags Nov 10 12:10:40 stintel: did you ever test glibc + gcc 4.8 on octeon? Nov 10 12:11:07 nbd: currently running it Nov 10 12:11:20 ok, nice Nov 10 12:11:23 ehr, no Nov 10 12:11:24 sorry Nov 10 12:11:30 tried that yesterday but kernel panic Nov 10 12:11:38 then made a build with gcc 4.6 Nov 10 12:11:42 but gcc 5.2 works? Nov 10 12:12:02 5.2 worked. I had some weird issue with jffs that took ~8min to "rebuild" the overlay Nov 10 12:12:16 but ran into that with 4.6 as well Nov 10 12:12:22 ok Nov 10 12:12:27 and now I made a new build and it's not doing that anymore Nov 10 12:13:19 nbd: sysupgrade is still broken though. I'm not sure I created a ticket for that Nov 10 12:13:25 so i can finally delete gcc 4.6 Nov 10 12:13:50 nbd: would you mind letting me test again with 5.2 once more ? Nov 10 12:14:23 sure Nov 10 12:14:25 go ahead Nov 10 12:14:32 i'll hold off on deleting 4.6 then Nov 10 12:14:34 I'll fire up a build Nov 10 12:14:41 thanks Nov 10 12:16:46 nbd: currently I am using this to fix sysupgrade, in target/linux/octeon/base-files/lib/preinit/11_sysinfo: https://bpaste.net/show/f30ad3631312 Nov 10 12:17:20 i'm running the combo gcc-5.2 / glibc on my erlite, too - no problem, yet Nov 10 12:17:30 cool Nov 10 12:18:49 irclogs/2015/04/Freenode/#openwrt-devel.log:30|21:01:32< stintel> nbd: I suspect r45496 broke sysupgrade on erlite. it fills /tmp/sysinfo/board_name with "cavium,octeon-3860" while sysupgrade expects it to be "erlite" Nov 10 12:19:30 i'll take a look Nov 10 12:19:33 ok my bad, never made a ticket for it Nov 10 12:22:19 I actually suspect the rebuild of jffs taking ~8min is related to that Nov 10 12:22:38 I built on a different machine where I did not have that file target/linux/octeon/base-files/lib/preinit/11_sysinfo yet Nov 10 12:22:49 after adding that again, the slow rebuild went away Nov 10 12:34:02 hmmm, no. the slow rebuild doesn't go away. I just don't really notice it, as now my config is being restored after sysupgrade and the router works just fine Nov 10 12:34:24 [ 580.885565] jffs2: notice: (1998) jffs2_build_xattr_subsystem: complete building xattr subsystem, 0 of xdatum (0 unchecked, 0 orphan) and 0 of xref (0 dead, 0 orphan) found. Nov 10 12:35:30 distemper: are you running ext4 or squashfs + jffs overlay on your ERL ? Nov 10 12:35:47 ext4 Nov 10 12:36:01 ok, so you wouldn't notice the slow rebuild I am experiencing Nov 10 12:36:07 does sysupgrade work for you ? Nov 10 12:37:04 i didn't try it for a long time, i always took out the usb stick, because it never worked in the beginning Nov 10 12:37:09 ah Nov 10 12:38:35 atm i'm trying to get a newer kernel running, but it's always crashing Nov 10 12:39:11 I'm on 4.1.11 Nov 10 12:42:48 intresting, the last really stale one (for me) was 4.0.9 Nov 10 12:42:55 s/stale/stable/ Nov 10 12:43:07 hold on, I'll send it to the ml Nov 10 12:59:39 just send patch with support for 4.1, or make a series that switches to 4.1 and then removes 3.18 ? Nov 10 13:02:05 going for the latter, blogic can choose what to with it :) Nov 10 13:04:46 i don't think blogic currently cares about octeon Nov 10 13:05:04 so i will take care of the patches Nov 10 13:05:07 ok :) Nov 10 13:05:13 your opinion on above question ? Nov 10 13:05:31 thanks, nbd Nov 10 13:05:31 and, while refrshing patches, 2 patches in generic are changed to. include that in the patch ? Nov 10 13:05:36 and stintel, too ;) Nov 10 13:05:48 s/to/too/* Nov 10 13:05:57 btw. i don't think the next release is going to be based on 4.1 Nov 10 13:06:02 4.4 was already announced as LTS Nov 10 13:06:11 so if you want, you might as well go for 4.3 Nov 10 13:06:22 as a preparation for switching to 4.4 when it's out Nov 10 13:06:28 k Nov 10 13:07:09 distemper: you want my current patch for going to 4.1 to test? or wait for 4.3 ? Nov 10 13:08:54 don't know yet when i find the time... but if you send it to the devel list, i'l find it Nov 10 13:09:04 ok then it'll be 4.3 to ML :) Nov 10 13:09:12 nice Nov 10 13:09:17 if it works that is Nov 10 13:09:59 I want to have a stable setup by the end of the week here in my apartment in Bulgaria, I am leaving for Belgium on friday and tuesday I'm going to florida until beginning of december Nov 10 13:11:33 I guess the slow jffs rebuild is because it is 1.6GB Nov 10 13:15:49 build #137 of pxa is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/pxa/builds/137 Nov 10 13:29:21 build #128 of ar71xx.nand is complete: Failure [failed shell_14] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx.nand/builds/128 Nov 10 13:36:57 build #119 of ixp4xx is complete: Failure [failed shell_14] Build details are at http://buildbot.openwrt.org:8010/builders/ixp4xx/builds/119 Nov 10 13:40:45 stintel: btw. if you manage to get 4.3 working, you might see a netfilter crash on reloading the firewall Nov 10 13:40:49 i'm currently fixing that one Nov 10 13:42:52 build #137 of ppc40x is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/ppc40x/builds/137 Nov 10 14:09:49 nbd: ok Nov 10 14:27:30 nbd: got a compile failure with xtables-addons, that's not what you are talking about I suppose Nov 10 15:01:55 4.3 running on my ERL, but my config isn't restoring again Nov 10 15:23:08 nbd: http://pastebin.com/2tt9ThXZ Nov 10 15:23:14 nbd: that's the bug you were talking about ? Nov 10 15:30:08 nbd: and I think you can close https://dev.openwrt.org/ticket/20885 and https://dev.openwrt.org/ticket/20884 after 47440 ? Nov 10 16:03:31 tip, anyone can tell me what happens with test_feature_macros and _DEFAULT_SOURCE ? Nov 10 16:12:42 txomon: man feature_test_macros ;) Nov 10 16:17:29 cyrusff: yeah, I understand that _DEFAULT_SOURCE enables all of them, what I don't understand is why aren't they included by default Nov 10 16:17:51 I mean, why are they being disabled? Nov 10 16:18:26 well not all of them are available on all platforms Nov 10 16:18:41 and on some platforms some functions even behave differently Nov 10 16:20:38 so you can e.g. say, i only want C99 standard functions and not all the gnu stuff, because i also want to compile my project on say osx Nov 10 16:22:45 cyrusff: ahhh I understand now... so the idea is to limit to the minimum macros you need? Nov 10 16:23:06 anyway when you say platforms, within openwrt, won't they all work? Nov 10 16:23:19 with platforms i mean, different c-libraries mainly Nov 10 16:23:26 oh ok Nov 10 16:23:34 build #118 of ramips.mt7620 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ramips.mt7620/builds/118 Nov 10 16:23:35 e.g. glibc, musl, android's bionic, whatever osx has Nov 10 16:24:06 hmmm ok Nov 10 16:24:19 and also different versions of c-libraries Nov 10 16:27:03 cyrusff: but which is the flag that is enforcing that? Nov 10 16:27:25 I mean, I am looking at CFLAGS and I don't see anything like -ansi Nov 10 16:42:58 there is a default set of functions exposed Nov 10 16:43:24 depending on your compiler settings Nov 10 16:43:36 you usually set a baseline with -std=c99 or so Nov 10 16:43:54 i don't know what's the default for current gcc though Nov 10 16:47:30 it changed from gnu89 to gnu11 with 5 Nov 10 17:03:08 currently trying to get ath10k working for qca6174 for 3.18/4.1 but issues with the fw/board.bin Nov 10 17:03:46 3.18.23 not working because potentially pulling ath git commit breaking something w/6174 Nov 10 17:04:08 3.18.21 not working because fw/board.bin not present for 6174 Nov 10 17:04:10 KanjiMonster: about time Nov 10 17:04:52 can i set-up a specific commit hash for ath git pulled to build kmod-ath10k ? Nov 10 17:05:25 massoud: kernel version should have no influence on how well ath10k works, as wireless drivers are built out of tree (-> package/kernel/mac80211/) Nov 10 17:05:48 unless you try to use the in-tree one, then: don't do that. Nov 10 17:07:57 KanjiMonster: but can i know which commit hash is used to build ath10k from my config ? Nov 10 17:13:18 massoud: no idea; but you have a date that should give you a rough estimate from when the sources were taken (from linux-next, not ath) Nov 10 17:20:42 massoud: so quite likely this one -> https://git.kernel.org/cgit/linux/kernel/git/next/linux-next-history.git/commit/?h=next-20151022&id=6dcf94ff0c9e28e5790799e53641dd256745f425 Nov 10 17:31:13 build #126 of arm64 is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/arm64/builds/126 Nov 10 18:05:44 stintel: here's the untested fix for the netfilter kernel bug: http://nbd.name/netfilter-fix.patch Nov 10 18:05:48 bbl Nov 10 18:54:35 build #109 of octeon is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/octeon/builds/109 Nov 10 19:06:12 using inet_ntoa doesn't seem common in openwrt code... What do you use to convert from IPs to char? Nov 10 19:56:35 nbd: that netfilter fix seems to work Nov 10 21:12:35 nbd r47441 trunk/toolchain/gcc/Config.in * gcc: use version 5.x for octeon instead of 4.6 - it has been tested with glibc Nov 10 21:12:41 nbd r47442 trunk/toolchain/gcc/final/Makefile * gcc: enable LTO support unconditionally Nov 10 21:12:48 nbd r47443 trunk/ (6 files in 5 dirs) * gcc: remove version 4.6, it is no longer needed Nov 10 21:12:52 nbd r47444 trunk/package/system/opkg/Makefile * opkg: remove obsolete gcc version check Nov 10 21:12:59 nbd r47445 trunk/package/libs/toolchain/Makefile * toolchain: remove obsolete relinking code Nov 10 21:13:15 nbd r47446 trunk/toolchain/gdb/Makefile * toolchain/gdb: disable python support Nov 10 21:13:21 nbd r47447 trunk/target/linux/generic/patches-4.3/611-netfilter_match_bypass_default_table.patch Nov 10 21:13:22 kernel: fix breakage in the 4.3 version of 611-netfilter_match_bypass_default_table.patch Nov 10 21:13:23 stintel: nice, thanks for testing Nov 10 21:18:15 nbd: np :) Nov 10 21:18:34 I'll try to send patches for 4.3 for octeon soon Nov 10 21:19:41 nbd: had to update xtables-addons though, send that as part of the series ? Nov 10 21:19:45 or separate Nov 10 21:20:09 you can send it now separately or later as part of the series Nov 10 21:20:11 doesn't matter Nov 10 21:33:54 going to do a new build after your latest changes Nov 10 21:33:59 and try ar71xx too Nov 10 22:30:40 luka r47448 trunk/package/network/services/uhttpd/files/uhttpd.init * uhttpd: add support for configuration option ubus_cors Nov 10 22:33:41 nbd: hmm getting assembler errors now Nov 10 22:33:48 like ../src/mips/n32.S:116: Error: opcode not supported on this processor: mips4 (mips4) `l.s $f12,0*8($25)' Nov 10 22:34:20 ugh, why the fsck is that even enabled Nov 10 22:34:26 it's trying to build python now Nov 10 22:35:09 noticed that last night as well, it builds python for no reason Nov 10 22:35:26 ./scripts/feeds uninstall python and everything still builds fine Nov 10 22:35:35 going to try that now Nov 10 23:12:07 hauke r47449 trunk/target/linux/ar71xx/patches-4.1/820-MIPS-ath79-add_gpio_function2_setup.patch * ar71xx: add function to access the GPIO function2 register Nov 10 23:13:05 hauke r47450 trunk/target/linux/ar71xx/patches-4.1/821-serial-core-add-support-for-boot-console-with-arbitr.patch * ar71xx: serial: core: add support for boot console with arbitrary baud rates Nov 10 23:15:04 hauke r47451 trunk/target/linux/ (9 files in 8 dirs) * ar71xx: add Arduino Yun Nov 10 23:19:05 nbd, distemper: patches for octeon/4.3 sent Nov 10 23:20:22 nbd: saw you closed #17753, maybe mips64 should be switch to glibc by default for now ? Nov 10 23:20:38 (and remove uClibc completely?) Nov 10 23:23:05 hauke r47452 trunk/target/linux/ generic/config-4.1 generic/config-3.18 generic/config-4.3 * kernel: add missing config options Nov 10 23:26:25 stintel: the new arc target uses uclibc-ng Nov 10 23:34:47 Hauke: yeah I know, I was just talking about the non -ng variant Nov 10 23:36:44 stintel: mips64 is already switched to glibc by default, and the non-ng variant is already removed Nov 10 23:38:19 oh Nov 10 23:38:47 guess I got confused by diffconfig output Nov 10 23:38:57 brb Nov 10 23:38:59 flashing Nov 10 23:40:30 that went well :) Nov 10 23:47:40 meh, did break my ipsec config somehow Nov 10 23:48:21 http://pastebin.com/cdzVjsYR Nov 10 23:49:21 iirc that's a missing kernel module Nov 10 23:54:49 think I am missing the aes module Nov 11 00:00:00 ah no, that's in-kernel Nov 11 00:05:36 yeah I messed up by enabling CONFIG_CRYPTO_SHA{1,256,512}_OCTEON Nov 11 00:07:22 anyone here who manages openwrt-devel ? Nov 11 00:07:40 looks like my from: is messed up Nov 11 00:37:22 grr, still not werking Nov 11 00:37:32 goddamn useless log entries **** ENDING LOGGING AT Wed Nov 11 02:59:58 2015