**** BEGIN LOGGING AT Sun Jun 17 03:00:02 2018 Jun 17 03:01:11 damo22: https://openwrt.org/docs/guide-user/network/wan/isp-configurations Jun 17 03:05:11 cool Jun 17 03:10:39 i will plug in my uart and dump the config, so i can post on that wiki Jun 17 03:16:07 which branch of openwrt is considered recommended to build Jun 17 03:16:19 master? or lede-17.01 Jun 17 03:17:28 like for a user to use Jun 17 03:19:46 are tagged commits considered stable? Jun 17 03:21:33 whatever you feel comfortable with, if you want to do development, the only branch that accepts new features is master Jun 17 03:21:43 right thanks Jun 17 03:24:33 as long as you don't expect help with long abandoned branches, like CC, BB or earlier Jun 17 03:25:22 :D Jun 17 03:25:43 v17.01.4 seems like a good choice Jun 17 03:29:23 I'm happy with master Jun 17 03:29:36 and given that 18.06 is merely weeks away... Jun 17 03:31:20 whats the difference between master and lede-17.01 Jun 17 03:31:29 which one is more up to date? Jun 17 03:31:32 one and a half year of development Jun 17 03:31:45 :O Jun 17 03:48:38 * ldir wobbles in - I've been up since 01:00 and it's just way too flippin early. Jun 17 04:08:42 * mangix wants Travis to die in a fire Jun 17 04:09:43 help me Jun 17 04:10:11 pkgadd help me Jun 17 04:12:33 gotta love it when Travis failures are not reproduceable localy Jun 17 04:17:56 Why does this error appear in the make? /userdev/openwrt/include/toplevel.mk:216: failure in the instructions for the 'world' objective Jun 17 05:00:00 https://openwrt.org/docs/guide-user/network/wan/isp-configurations <--- i added my config :D Jun 17 07:48:57 Ola, what is difference between quilt patch and patch -p0 < my.patch ? Jun 17 08:14:39 muhaha: quilt helps you with larger patch series and is quite good at refreshing/ rebasing patch stacks. while quilt can support any patch level, the usual patch level is -p1 Jun 17 08:15:13 ah thanks. Jun 17 08:17:39 And one more question. I am trying to patch file, but it fails to Hunk FAILED ... This leads to changed source code, but I dont know from where the code is changed. I just want to apply two patches :] Jun 17 08:19:42 that means you'll have to rebase your patch, usually with quilt push -f, find . -name \*\\.rej fixing the conflicts manually via quilt edit and finally quilt refresh to commit your changes (rebasing) to the patch Jun 17 08:28:15 pkgadd: And from what source is target file changed ? wrt build system downloads patches from internet, or what? Jun 17 08:31:59 I am trying apply https://forum.lede-project.org/t/pc-engines-apu2-leds-not-working-in-lede-17-01/12722/9 to 18.06 branch. But patches are created everytime from master branch ? And I should rebase them agains 18.06, 17.01 branch?\ Jun 17 08:37:02 the wiki has a detailed description about how to work with patches in the build system, under the developer documentation Jun 17 08:40:15 i read https://wiki.openwrt.org/doc/devel/patches manytimes... Jun 17 08:40:39 no mention about branches Jun 17 08:45:19 quilt doesn't care about branches, it's independent from git and works on whatever branch you have currently checked out Jun 17 08:46:52 then i do not get error Hunk FAILED ... Jun 17 08:47:09 normally patches/updates are done to the master branch and are then cherry picked to release branches if deemed relevant. If the change won't cherry-pick because of conflicts then a patch for that branch needs to be produced. Jun 17 08:50:11 I did git clone openwrt; git checkout 18.06 ; ...prepare.... ; cp mydir/302-fix-apu2-nct5104d-chipID.patch package/kernel/gpio-nct5104d/patches/302-fix-apu2-nct5104d-chipID.patch ; make ------> but this is not working, fails about Hunk FAILED Jun 17 08:50:56 right, so the patch doesn't apply cleanly and hence also wouldn't cherry pick cleanly. Jun 17 08:51:54 you need to look at the reject file, manually apply the required change, refresh the patch and update back into the openwrt tree. Jun 17 08:53:00 One more question, is 18.06 prod ready? Or should I use 17.01 branch ? Jun 17 08:53:41 the last stage updates the patch file in the tree, and the updated patch is what you would then git commit and produce a patch from that to send upstream. Jun 17 08:57:16 oh :( Jun 17 08:57:29 what I have to do to sumbit patches https://forum.lede-project.org/t/pc-engines-apu2-leds-not-working-in-lede-17-01/12722/10 to 18.06 branch? Jun 17 09:00:36 Are lede snapshots builded from 17.01 or 18.06 branch ? Jun 17 09:29:44 can anyone help how to compile just one package package/kernel/gpio-nct5104d/ ? I dont want to wait for whole build .. I just want to know if patch is ok. Thanks Jun 17 09:38:07 muhaha: 'make package/kernel/gpio-nct5104d/clean; make -j1 package/kernel/gpio-nct5104d/ V=s' Jun 17 09:38:19 thatll let you see any errors more easily Jun 17 09:38:39 actually Jun 17 09:38:54 it might just be package/gpio-nct5104d/ not the full package/kernel/gpio-nct5104d/, i dont remember Jun 17 09:39:29 that assumes you've already built the toolchain and kernel once before already Jun 17 09:42:10 m4t: Thanks! Is package/kernel/gpio-nct5104d/patches directory also considered in this phase? Jun 17 09:42:27 yes if you clean it it will reapply the patches Jun 17 09:43:55 also if you do make package/gpio-nct5104d/prepare instead of compile that will just extract and apply the patches Jun 17 09:47:14 Ok. Thank you :) Jun 17 10:05:22 dangole ping? Jun 17 10:51:00 nbd could you please explain this commit? I don't understand why wan and lan shoudl use the same mac address https://github.com/openwrt/openwrt/commit/975e1357e6136ab5126e8d079e80bdcde338f8a6#diff-d0f217ef055f40184d0c2f1641a1cf47R357 Jun 17 10:56:09 morning Jun 17 10:56:39 hi ultito Jun 17 10:58:27 18.06 cannot be compilated ... Applying ./patches-2.0/100-musl-compat.patch using plaintext: Jun 17 10:58:27 patching file src/usbipd.c Jun 17 10:58:28 Hunk #1 FAILED at 453. ... Jun 17 10:58:36 any idea? Jun 17 11:05:17 muhaha: What package did the build die in? Jun 17 11:05:47 usbip Jun 17 11:09:08 Hmm.. I don't have usbip in my build. But it's possible that the patch has since been merged into the base usbip source. Jun 17 11:09:35 Try ./package/feeds/packages/usbip/patches-2.0/100-musl-compat.patch to someplace save, and re-try the build. Jun 17 11:09:56 If it builds cleanly without the patch, you're golden. Jun 17 11:11:11 (I hit something like that with m4 - glibc changed stuff, I did a local patch for it, then the next build failed because a new dot release of m4 already included the fix.. Jun 17 11:14:04 CC usbip_list.o Jun 17 11:14:05 usbip_list.c: In function 'list_devices': Jun 17 11:14:05 usbip_list.c:214:11: error: assignment makes pointer from integer without a cast [-Werror=int-conversion] Jun 17 11:14:05 devpath = udev_device_get_devpath(dev); Jun 17 11:14:05 ^ Jun 17 11:14:07 cc1: all warnings being treated as errors Jun 17 11:14:08 Makefile:418: recipe for target 'usbip_list.o' failed Jun 17 11:14:20 without patch Jun 17 11:14:31 so broken ,lol Jun 17 11:15:18 Hmm.. that's different from what the patch actually does. ;) Jun 17 11:15:47 which is change usbipd.c line 460.. Jun 17 11:17:02 Looks like that 100-musl-compat.patch isn't needed anymore, but some other new patch *is* needed.. Jun 17 11:23:16 i guess that snapshots are not builded from 18.06 i guess Jun 17 11:23:47 snapshots are build from master Jun 17 11:32:23 This is about trying to get 2 patches into gpio-nct5104d.c right? Or have I missed something ? Jun 17 11:35:21 ldir: no ... Jun 17 11:35:31 I mean yes Jun 17 11:35:46 that fills me with confidence! Jun 17 11:36:28 ldir: https://github.com/lukasmrtvy/docker-lede-apu2c4/tree/test Look at this, in Dockerfile there is command step by step.... Jun 17 11:37:14 What branch are you trying to do this to ? Jun 17 11:37:41 Hunk FAILED, everytime.... 17.01, 17.01.4 release, 18.06 seems its ok, but it fails with usbip patching... GRRR:( Jun 17 11:38:55 Neither of those patches touches anything usbip related. Jun 17 11:39:22 I know, like I said, its ok on 18.06 but there comes another error with usbip Jun 17 11:39:35 even with these patches is impossible to compile 18.06 Jun 17 11:39:49 *without Jun 17 11:47:53 ldir: so ..? Any idea? Jun 17 11:49:47 not really - I'm waiting for a toolchain to build so I can try to build an image as well. Jun 17 11:50:22 A thought occurs that the packages feed should really be on the 18.06 branch if you're trying to build an 18.06 release Jun 17 12:03:52 How can I install a proprietary Wi-Fi driver for Broadcom devices on latest OpenWRT? Jun 17 12:04:31 The device I have only has 802.11g support on open-source drivers and no no 802.11n or ac support. Jun 17 12:09:39 muhaha: silly question time... do you *need* the usbip package ? Jun 17 12:18:51 ldir: No, I am just wondering why its not working... Jun 17 12:20:10 make package/usbip/compile -j1 V=s will give you the error/s Jun 17 12:23:47 so how to compile without usbip? Jun 17 12:25:33 make menuconfig - go and de-select the kmod-usbip,client & server Jun 17 14:12:51 I would like to ask, but others will probably kill me. So I will not ask, but if you know answer, please tell me :-) Jun 17 14:28:28 lol Jun 17 14:29:01 * ldir has packets of gravel, false beards & otters noses for sale :-) Jun 17 14:32:55 * ldir realises he won't get any sense out of the German openwrt contingent for at least another 2 hours + Jun 17 14:59:59 I love the world cup Jun 17 15:00:24 dlir you got the game on dude Jun 17 15:00:58 ldir Jun 17 15:01:04 you got the game on? Jun 17 15:05:56 yes I have :-) Not even a football fan, and I've been up since 0100 so sleep depreved and slightly delirious. Jun 17 15:13:10 ldir as a blind dude my new google home mini is proper cool! Jun 17 15:13:35 I just ask for the game on five live and boom! it plays. Jun 17 15:14:16 for me radio is much better. Jun 17 15:16:16 I was thinking about getting some smart lights so i can make shure my kids don't leve the lights on but I cant find any that are not a risk to my network. Jun 17 15:16:44 lots of the venders seem a bit dodgy Jun 17 15:27:16 I'd like to get Ikea ones, but what bugs me is that they seem to be something-zigbee, but you still need their proprietary "hub" Jun 17 15:28:39 tp-link does proper wifi ones you can hack with own firmware, but ... eh, it's all shoddy. Jun 17 15:40:00 what "smart home" stuff? Jun 17 15:59:18 hah, i'm trying to solve that problem right now- but just in the form of trying to get lirc packages updated so that i can use the dumb cheap infrared-control LED lights everywhere Jun 17 15:59:59 because screw dealing with all the infrastructure just for expensive, complicated lights that are gonna need their firmware replaced after getting taken over by a serbian botnet Jun 17 16:01:58 Putin on the lights - /me fetches coat Jun 17 16:10:36 Putin has nothing to do with Serbia :p Jun 17 16:11:22 ldir: is there created any issue about usbip package in 18.06 ? Jun 17 17:18:28 gzc__: oh goody, the last time I have seen a working LIRC package was ... 12.09? Jun 17 17:22:32 we'll see how close it can get to working :D using https://github.com/danitool/openwrt-pkgs as a template, but i'm still getting used to the build system and having dumb issues with trying to build it against python3 (which, yep, is now a lirc dep) Jun 17 17:49:37 speaking of which, i've apparently mixed something up cross-compiler-wise if these LONG_BIT errors are any indication. https://pastebin.com/07middaV . Putting the makefile up too just in case anyone feels inspired :) https://pastebin.com/9ft7yWEa Jun 17 18:08:11 gzc__: ugh, full-on python? not sure if it's worth the effort then. Jun 17 18:09:39 well, there's a 'lite' package :/ using this on a modern device with crazy amounts of space (comparatively) Jun 17 18:31:50 hello Jun 17 18:31:59 jow: could you take a look at my regression report? Jun 17 18:32:11 i was wondering if there is some easy fix for that we could commit before rc1 Jun 17 18:32:37 "Regression in interface names & /etc/config/network (ethX order)" Jun 17 19:16:05 <_lore_> quit Jun 17 20:38:16 oh great, I've another not-booting device in 18.06 :/ Jun 17 20:38:22 DIR-885L Jun 17 21:44:22 Anybody know of any known issues with IPv6 neighbor discovery in very recent lede-snapshot? Jun 17 22:19:55 build #30 of mediatek/mt7623 is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/mediatek%2Fmt7623/builds/30 Jun 17 23:57:26 updated openwrt/upstream, https://sdwalker.github.io/uscan/index.html Jun 18 00:39:46 it would be really nice if someone made a gsm radio that had partial sdr, so that they are useful in more than one region Jun 18 00:40:18 domt need to open up the baseband just allow software to set bands and transport Jun 18 00:41:12 i mean, some of the new bts stuff is almost like that Jun 18 00:42:32 as especially with lte, the cross section of region and carrier is basically zero unless you have a radio for each one Jun 18 00:50:59 the problem just is that the FCC, and ETSI and friends doesn't like that at all Jun 18 00:51:56 they want vendors to cut the feature set, well, the regdom stuff as deeply into hardware as possible Jun 18 00:52:20 yeah Jun 18 00:52:38 but they could implement restrictions in hardware Jun 18 00:53:04 either with passive listening to figure out region Jun 18 00:53:08 or crda type thing Jun 18 00:54:27 dont transmit until it hears something and gets region, kinda like dfs Jun 18 00:59:28 idiots removing (valid) dfs restrictions are probably the biggest reason for the latest 'crackdown' on wireless vendors from FCC/ ETSI - maybe not a whole lot of cases in numbers, but rather high impact ones (as it spoils weather radar over pretty long distances) Jun 18 01:00:13 heh Jun 18 01:00:18 yeah Jun 18 01:01:09 they should have used a better frequency range obviously heh Jun 18 01:03:04 they probably see it in exactly the reverse way, as in 'we should have never allowed wlan in the upper 5 GHz band at all' Jun 18 01:04:42 yeah thats what i mean Jun 18 01:04:59 or at least not have bands that overlap with radar Jun 18 01:06:54 the fun of europe and the US having exactly contrasting uses of the frequency band, which made sense, but is hurting us now Jun 18 01:06:59 but still, hardware enforced passive listenimg before transmit seems like a reasonable compromise Jun 18 01:07:05 heh yup Jun 18 01:13:30 see now the older iphones do have sort of configurable radios Jun 18 01:13:47 but its secure, so i guess thats enough for the fcc/etsi Jun 18 01:15:28 the question is if they'd still consider it secure enough for a new device entering the market (and Apple may be viewed more favourably than the adnroid eco system, 'cause no one has ever rooted iOS... as long as the general public believes that) Jun 18 01:15:39 or.. it isnt anymore and thats why they have limited radios Jun 18 01:16:07 yeah Jun 18 01:16:24 theres only been one incident of modifyimg baseband though iirc Jun 18 01:16:35 and it couldnt configure the radio Jun 18 01:17:17 and that was on the 3gs, silly intel crap :D Jun 18 01:17:41 its pretty insane really Jun 18 01:18:20 to get usable cell service with any modern device (apple, android, whatever) i would have to buy at least 3 variants Jun 18 01:18:38 na, eu, apac Jun 18 01:19:38 i was kinda lucky with euro/japan card in the usa as t mobile operated hspa+ on a common band in most locations Jun 18 01:20:17 pretty silly though Jun 18 01:21:20 my na device shares exactly 0 common bands with one of the networks here, its insane Jun 18 01:21:40 but thats largely uk using new 800mhz bands Jun 18 01:22:28 it's definately annoying as hell Jun 18 01:23:59 yup Jun 18 01:24:07 its a ridiculoys problem Jun 18 01:24:12 ridiculous Jun 18 01:24:34 one that seems at least to me, to be easily solvable Jun 18 01:24:53 qualcomm already make wideband radios with configurable rf Jun 18 01:25:49 the problem is that the legal bodies really don't want it to be solved, and have insisted on that view for at least the last 80 years - for good reasons, even if we'd disagree. Jun 18 01:25:54 yup Jun 18 01:26:28 but policy isnt going to stop people doing dumb shit, so they already lost Jun 18 01:27:39 unfortunately they consider it as just more reason to come down hard on the manufacturers and insisting on even more drakonian countermeasures Jun 18 01:27:51 yup Jun 18 01:29:37 question is why hasny some shady chinese firm made such a radio heh Jun 18 01:29:42 hasnt Jun 18 01:29:56 they dont normally care about regs Jun 18 01:34:22 actually, looks like the chip i was looking at is already in phones, samsung so probably locked in bootloader **** ENDING LOGGING AT Mon Jun 18 03:00:06 2018