**** BEGIN LOGGING AT Fri Aug 02 03:01:07 2019 Aug 02 03:16:29 ugh. this wireguard crap is so frustrating Aug 02 03:16:58 i get simplicity and everything but it's impossible to debug. Aug 02 03:17:17 Is perfect, no debug need. Aug 02 03:17:29 Everyone knows if you just write it properly the first time it's all good, right? Aug 02 03:17:36 lol not for me Aug 02 03:18:01 ifconfig shows two packets under RX but no further traffic Aug 02 03:18:07 Why have error handling when your code doesn't have bugs? ;) Aug 02 03:18:16 hahaha Aug 02 03:19:05 anyway, i give up. this crap refuses to work. Aug 02 04:01:22 leitec: please fix your network, that join flood has been going on for about a week now. Aug 02 06:41:09 jow: what's the reason behind using 'IGNORE_ERRORS=n m' on images builders in `make -j8 package/compile V=s 'IGNORE_ERRORS=n m' BUILD_LOG=1` ? Aug 02 06:43:24 jow: does it make a sense to continue with the build if some of the packages fails? I mean, it's very likely, that the package might be needed later in the image, like it's the case for broken adb in ar71xx/mt7620 images Aug 02 06:46:11 I've seen something similar with layerscape and tfa week ago, but I've thought, that it's the problem of their convoluted makemess Aug 02 07:07:56 Tapper: hello Aug 02 07:13:01 and good morning everyone! Aug 02 07:13:03 hello guys :D Aug 02 08:03:25 ynezz: i don't know about 'IGNORE_ERRORS=n m', but IGNORE_ERRORS=m means that if the package is a module (that is, not installed in the image), the build will continue, which is probably what you want. you can figure out later what didn't build and investigate, while building all that's possible. Aug 02 08:04:17 i've never seen the 'n' in there, and i don't know what it means. Aug 02 08:38:45 ynezz: ping - will try to resume my effort on firmware-utils. did you merge in my tweaks to your branch, thinking about rebasing on top of your latest tools branch Aug 02 08:42:09 jow: did we have a fix for the bu introduced by Aug 02 08:42:10 https://github.com/torvalds/linux/commit/10f4e765879e514e1ce7f52ed26603047af196e2#diff-0c4c7b6b5fb1b3801c25d1dc3e2fab97 Aug 02 08:56:55 blogic: https://github.com/openwrt/openwrt/pull/2266 Aug 02 08:59:24 russell--: http://buildbot.openwrt.org/master/images/builders/ramips%2Fmt7620/builds/25 images fails, because some device needs adb-enablemodem which depends on adb and adb fails to build Aug 02 09:00:20 that adb build breakage is likely caused by the kernel bumps Aug 02 09:01:13 same here http://buildbot.openwrt.org/openwrt-19.07/images/builders/ar71xx%2Fgeneric/builds/9 Aug 02 09:02:56 so probably `DEVICES_PACKAGES += adb` would make the build dependency handling work Aug 02 09:06:47 ldir: not yet, it's a long term effort :p Aug 02 09:09:50 russell--: the 'n' is to prevent failed builds of packages that are neither selected as m or y, but still build as a build dependency (because e.g. static linking), from stopping the build Aug 02 09:25:34 ynezz: thx Aug 02 09:26:04 kristrev: i pushed the mediatek update, if you fond the time to test on the unielec that'd be great Aug 02 09:26:35 Rene__: there is a patch from MTK inside the mediatek v4.19 patches folder. has lots of fixes and features, might want to have a look at them Aug 02 09:37:38 blogic: Sure, will let you know Aug 02 09:39:41 blogic: Is that mt753x driver suitable for mt7530? If so I'd like to port it to ramips for mt7621. Aug 02 09:41:59 blogic: nm. I'll ask the driver author :) Aug 02 09:55:48 gch981213: ;) Aug 02 09:56:02 it is, i have a bananapi 64 here with mt7622 and mt7530 Aug 02 09:56:06 mtk mailed it to me Aug 02 09:56:22 gch981213: the DSA driver will be upstream in Q4 I am told Aug 02 09:57:48 KanjiMonster: thanks for the explanation Aug 02 10:14:19 blogic: I was told that MTK's internal testing shows ethernet performance is slightly worse when using DSA, and that's why they still wrote a swconfig driver for their SDK. Aug 02 10:14:33 could be Aug 02 10:14:52 for$corporate marketing benchmarks that is fine Aug 02 10:14:57 i'd prefer a DSA driver Aug 02 10:15:19 gch981213: i have no problem with the quality/performance trade-off :-) Aug 02 10:16:31 I have no deep experience, but all I have read from DSA is that it is very nice concept of how to implement networking layer(s) Aug 02 10:17:38 blogic: I am experiencing problems with MTD layer - it will not read completely blocks when they are not divisible by 512-byres, leaving out some data. See the mac partition on the Archer C60 V2 DTS in ath79. Aug 02 10:17:46 Slight perf drop I could accept, especially for other gains it ought to offer, optimizations ofcourse allways welcome Aug 02 10:18:15 blogic: I looked a little bit into it and saw it's the mtd layer not reading. Is this correct and so we need to fix DTS files or should I talk to mtd-layer guys or look at it? Aug 02 11:01:32 ynezz: ok, np :-) Yes it's a bit of an elephant task. Aug 02 11:08:05 blogic: thanks Aug 02 11:08:27 blogic: Still working on the 2nd version of the upstream patches Aug 02 11:10:45 gch981213: My upstream patches for MTK and mt7530 already support mt7621 and also have PHYLINK support. PHYLINK is most useful for devices with SFP/Fiber port. Aug 02 11:14:15 Phylink patches for ethernet: https://patchwork.ozlabs.org/patch/1136551/ Aug 02 11:14:18 Rene__: I was hoping that vendor's driver would have more features and I just found out that I was wrong. Aug 02 11:14:46 phylink patches for mt7530 with port 5 support https://patchwork.ozlabs.org/patch/1136551/ Aug 02 12:45:38 ynezz: have done a bit more from a macOS / clang perspective Aug 02 12:45:52 I think 4 more utils. Aug 02 16:33:53 sorry, here's a classic re procd and why I'm so ranty. banip init script: Aug 02 16:33:56 stop_service() Aug 02 16:33:57 { Aug 02 16:33:57 rc_procd "${ban_script}" stop Aug 02 16:33:57 rc_procd start_service Aug 02 16:33:57 } Aug 02 16:34:35 the last thing a STOP_SERVICE does is..... rc_proc ***START**** SERVICE Aug 02 16:36:26 This is the equivalent of "Ken Thompson has an automobile which he helped design. Unlike most automobiles, it has neither speedometer, nor gas gauge, nor any of the numerous idiot lights which plague the modern driver. Rather, if the driver makes any mistake, a giant "?" lights up in the center of the dashboard. "The experienced driver", he says, "will usually know what's wrong."" Aug 02 16:43:48 ldir: I love that Ken Thompson joke. Aug 02 16:55:58 * ldir is very cross Aug 02 16:57:41 We fuss about commit messages & minutia of lines with spaces/tabs whatever. Well in my world f**k signed off by and all that s**t - DID YOU UPDATE THE F*****G DOCUMENTATION? Aug 02 17:02:01 There should be canonical set of 'this is how to do procd' scripts. Aug 02 17:03:36 * ldir logs out of irc before I murder someone :madeyes: Aug 02 18:02:33 ldir: i was busy hacking procd Aug 02 18:02:37 ldir: so no docs Aug 02 18:02:41 but agreed we need them Aug 02 18:03:01 i worked out a proposal for prpl that they must put an asciidoc into the dri of each package Aug 02 18:03:16 once the format proposal is final we should consider to also adopt this for owrt Aug 02 18:03:31 if anyone knows a technical writer for the job i could organize funding Aug 02 19:31:00 ldir: are you related to this? https://www.bbc.com/news/uk-england-derbyshire-49189955 Aug 02 19:32:48 sorry. you are darbyshire, not derbyshire. Aug 02 20:53:11 russell--: If you're suggesting I'm cracking up.... I couldn't agree more! Aug 02 20:54:11 mostly it was reading the landslide blog and seeing a name that looked familiar Aug 02 21:08:28 blogic: did you see the procd patch? Aug 02 21:09:08 * russell-- renews appeal to apply my scons patch to openwrt tree, worry about moving it to packages later Aug 02 21:11:03 https://patchwork.ozlabs.org/patch/1137962/ Aug 02 23:29:05 looks like that musl update is causing long range build failures Aug 03 01:33:29 hello Aug 03 01:35:17 need some help figuring out how to flash openwrt on beaglebone black. I've built the code, but haven't been able to flash it. is there someone here that can help? Aug 03 02:10:30 prasanna: https://openwrt.org/toh/texas.instruments/beagleblack **** ENDING LOGGING AT Sat Aug 03 02:59:57 2019