**** BEGIN LOGGING AT Sat Jun 09 03:00:07 2018 Jun 09 03:06:08 the link won't help me. I'm not an advanced linux user, I can't build this properly without detailed instructions and me need router. and i ru Jun 09 03:07:22 https://translate.yandex.ru/?lang=ru-en&text=%D1%8E%D0%BC%D0%BE%D1%80 Jun 09 03:17:59 not bad pi2 lede + router 15$ l2tp as switch. but me not need two routers. Jun 09 03:37:35 not bad pi2 lede dhcp + router 15$ l2tp isp as switch . but me not need two routers. goodbye Jun 09 04:49:00 anybody know what the latest changes to mt7621.dtsi are about? i think i'm going to have to make some changes to my board's .dts file now, but just want to suss the situation Jun 09 04:49:38 because now my wifi isn't working again Jun 09 05:46:48 bah he's gone Jun 09 06:02:30 mangix: or not Jun 09 06:03:08 biangbiangmian: ping Jun 09 06:31:54 * ldir wanders in Jun 09 06:35:21 sup Jun 09 06:55:58 nbd: thanks for "kernel: backport patch to fix dst handling for offloaded connections", that was on my list Jun 09 07:29:30 hi guys Jun 09 07:31:21 * ldir waves Jun 09 07:32:11 :) Jun 09 07:32:50 ldir: your xiami has a gigabit switch right? Jun 09 07:33:31 yes, a 3 port gigabit switch Jun 09 07:34:03 ok. there seems to be a 100 Mbps varity as well, the naming scheme really is confusing. Jun 09 07:35:52 yes agreed. Jun 09 07:36:29 and mine still isn't in active use Jun 09 07:37:32 :) Jun 09 07:38:00 cannot get off the archers! Jun 09 07:38:11 :P Jun 09 07:38:21 you don't want to know my it backlog Jun 09 07:39:52 but i reckon your openwrt setup is a bit more complicated than mine Jun 09 08:44:42 I'm trying to add a new device: the armada-388-helios4 . Here's where I'm stuck: http://ix.io/1cO5 And this is my current openwrt tree: https://code.azarus.ch/openwrt/log/?h=helios4 Jun 09 08:44:54 Thanks for any clues I might get :) Jun 09 08:53:18 azarus: you have to patch arch/arm/boot/dts/Makefile to also build a dtb file from your dts file Jun 09 08:54:19 Hauke: ah yes! that makes sense. Jun 09 08:57:06 or could i just add the line to 105-build_ew_dtbs.patch ? Jun 09 08:57:50 azarus: I would suggest to just put this patch into the patches-4.14 directory: https://patchwork.kernel.org/patch/10449393/ Jun 09 08:58:33 Hauke: ah, perfect. Thank you. Jun 09 09:02:42 those kernel bumps keep coming :-/ Jun 09 09:27:39 Hauke: it builds, yay. any other things that might be wrong? https://code.azarus.ch/openwrt/commit/?h=helios4 Jun 09 09:48:33 * ldir apropos of nothing in this channel /me thinks that some people have a very bizarre way of asking for help. Jun 09 09:49:32 Well, some people have trouble aditting they need help with something or they're shy, or something else... Jun 09 09:51:59 true. Saying "ldir and others, this openwrt thing is a mess. Sort it out." is very odd. Jun 09 09:52:13 ;( Jun 09 09:52:52 "Your project sucks, but I need it and you need to improve it. Also, do it for free. Thanks." Jun 09 09:52:55 * r0bert is just me Jun 09 09:53:01 I was reminded why I didn't frequent the openwrt channel :-) Jun 09 09:54:45 * azarus scrolls through #openwrt backlog Jun 09 09:54:46 oh yup Jun 09 09:55:43 azarus: could you add the commit message to target/linux/mvebu/patches-4.14/106-armada_388_helios4_dtbs.patch Jun 09 09:55:54 so that people know who is the original author Jun 09 09:55:58 and the description Jun 09 09:56:25 Hauke: sure Jun 09 09:56:27 otherwise this looks good Jun 09 09:56:52 I haven't tested it at all yet tough o_O Jun 09 09:56:59 But thanks. Jun 09 09:57:33 testing as well??? *who are you?* :-) Jun 09 09:57:57 * ldir means to say "Welcome!" Jun 09 09:58:11 someone who doesn't want to embarrass himself with terrible patches >_< Jun 09 09:58:58 ah you will, at some point. We all do^w have :-) Jun 09 09:59:23 Hehe, I already have in some other projects ;) Jun 09 10:00:17 ldir: i need you to help me push 18.06 faster Jun 09 10:00:19 like stat. Jun 09 10:00:29 all those lazy devs sitting on their ass, i mean... Jun 09 10:01:04 if you could fix my roof leaking that would be nice as well. for free of course. Jun 09 10:01:56 * ldir puts borromini on the naughty step. Jun 09 10:02:07 :'( Jun 09 10:02:24 come back when you've thought about what you've said. ;-) Jun 09 10:02:44 i thought it very considerate ^_^ Jun 09 10:02:49 except that i forgot 'please' of course :D Jun 09 10:03:15 Hauke: edited the commit: https://code.azarus.ch/openwrt/commit/?h=helios4&id=8a6f7fc5682ba017a9d213f675f8fbb2d6d28746 Jun 09 10:04:13 azarus: looks good Jun 09 10:06:38 ldir: that's ironic, since you're almost always present Jun 09 10:08:02 my bouncer is mostly present (briefs periods of absence when I do a router firmware upgrade... which tells you how often I do it...quite a lot!). As for me be present, well, sometimes. Jun 09 10:08:29 * ldir has lack of sleep and cannot type English today. Jun 09 10:08:52 Hm -- I might actually buy the helios4, and/or kindly ask if I might get a discount on it for development :P Jun 09 10:09:39 ldir: yeah, I had a terrible night of sleep too. Just to be clear: it's a compliment that you're as active as you are =) Jun 09 10:11:04 * ldir flashes his hh5a with latest code in anticipation of the test ADSL line being installed on Monday. Jun 09 10:11:39 * Tapper hugs ldir Jun 09 10:11:53 ldir OK there matey? Jun 09 10:12:09 and ferretts around in a box of bits for an ADSL filter and phone lead.. ADSL filter... remember those! Jun 09 10:13:06 ldir it's not that bad any dude I just had my wife flip out on me because I wanted her to take the kids to kickboxing and she is pushed for time. lol Jun 09 10:13:46 Being on the wrong end of a northern lass is not a nice thing! Jun 09 10:14:36 * ldir is married to a Mancunian of the female persuasion :-) Jun 09 10:14:39 That is why I got so good at saying pleas hahahaha Jun 09 10:14:52 and thankyou Jun 09 10:15:38 So you know what i went through hahaha Jun 09 10:16:25 I will cook her somthing nice and she should forgive me I think Jun 09 10:16:30 I hope! Jun 09 10:16:32 haha Jun 09 10:16:39 jow: hi Jun 09 10:17:26 Tapper: lol Jun 09 10:17:28 jow: any fixes? :) Jun 09 10:22:42 * ldir emerges from the 'electronic mysteries & cables' box with an ADSL filter & cable. Jun 09 10:22:46 has anybody looked at getting the archer c7 v2 ported to ath79? Jun 09 10:25:44 huaracheguarache: blogic is aware of the desirability of such a port so it's on his radar. But it will be post his holiday & post 18.x release. He didn't think he had one until he realised that (like mine) it was serving a key point in his network infrastructure! Jun 09 10:26:22 unless someone does it in the meantime. Jun 09 10:26:46 ok, sounds good =) Jun 09 10:27:32 You know my c7-v2 is a grate AP I have it on a newesh snapshot and it is rock solad Jun 09 10:27:54 It has a lot of trfic go through it and it never falls over Jun 09 10:28:45 It's bin up for 15 days and has handled about 300 gb of dater Jun 09 10:29:33 I use one in client mode to facilitate a wireless link from my netgear r7800 to my iptv decoder. Jun 09 10:29:48 the decoder doesn't have wifi enabled in the firmware, lol Jun 09 10:35:41 Tapper: mine are lucky to survive 24 hours between flashes ;-) Am always testing something. Jun 09 10:36:20 The c7 is the mane AP so it's the other ones that i flash all the time Jun 09 10:38:41 ldir: are you involved in the development of cake? Jun 09 10:39:54 I have a finger as opposed to a hand in it. bufferbloat/cake was actually the reason I even looked at openwrt. Jun 09 10:41:24 and then it became a love story Jun 09 10:41:30 ldir: how is the upstreaming effort proceeding? Jun 09 10:41:35 lol Jun 09 10:42:47 huaracheguarache: temporarily stalled is the best way of describing. Jun 09 10:44:18 huaracheguarache: Toke is the guy who's been driving the process of late and seems able to deal with the feedback from upstream. Unfortunately he's otherwise engaged for 3 weeks or so. Jun 09 10:45:37 from an openwrt perspective I try to keep the openwrt cake implementation up to date. Unfortunately there's a snag there at the moment too. Jun 09 10:46:03 what kind of snag? Jun 09 10:46:34 There was a recent change in cake to support setting rates using 64bit values. The kernel module and userspace (tc) talk to each other over netlink. Jun 09 10:47:26 for unknown reasons, these 64bit values don't seem to travel well across netlink on mips. Jun 09 10:48:45 Toke has been throwing code at me to try on my mips based archers but so far we've hit a brick wall as to why using 64 bit values over netlink just doesn't want to work for us. I posted on openwrt maillist for ideas help but so far no magic answer has appeared. Jun 09 10:49:06 strange =/ Jun 09 10:50:11 I'm carrying a local patch in my own tree to convert back to 32bit....so I can run 'latest-ish' cake without hitting the bug... but for hopefully obvious reasons I'm reluctant to throw that out in the wild. Jun 09 10:51:34 of course it all works fine on x86_64. Jun 09 10:52:36 ldir: do you have some more details? Jun 09 10:52:42 netlink is in natiuve endian Jun 09 10:53:36 ldir: have you added some debugging to check wrere the data gets currupted? Jun 09 10:55:00 ldir: does get_rate64() work correctly? Jun 09 10:55:20 https://github.com/dtaht/sch_cake/tree/cobalt is the 'out of tree' module that gets built. The upstreaming attempted code is in a different branch. Jun 09 10:55:27 anybody know what the latest changes to mt7621.dtsi are about? i think i'm going to have to make some changes to my board's .dts file now, but just want to suss the situation Jun 09 10:55:35 because now my wifi isn't working again Jun 09 10:56:10 biangbiangmian: Can you specificy 'which changes' beyond latest ? Jun 09 10:57:50 commit 66cc6dd Jun 09 10:58:47 oh sorry Jun 09 10:58:55 that's not in mt7621.dtsi Jun 09 10:59:01 Hauke: the tc code is https://github.com/dtaht/tc-adv but of course we don't run latest iproute in openwrt so the patch I produce for tc cake support in openwrt is https://github.com/ldir-EDB0/iproute2-cake-next/commits/cake. - the tc/q_cake.c that gets exported from that is the same as found in the tc-adv repo - it's just easier to produce a patch that way for me. Jun 09 11:02:36 commit 289b2f5 Jun 09 11:02:42 but could it also be the kernel bump? Jun 09 11:02:57 Hauke: as regards debugging. Toke did this: https://github.com/ldir-EDB0/openwrt/commit/c6e831f8a15a4fe03383ae3f78876ee1ac6f0d62 Jun 09 11:03:48 Hauke: and we simply don't see tca_xstats/tca_stats2 being passed. Jun 09 11:05:49 my hack is https://github.com/ldir-EDB0/openwrt/commit/88191edf8c5b09c1e11b3f2a40295dfe95b5b3db which is simply 'don't send 64bit values back' - no messaging structure is changed as such. Jun 09 11:06:33 Hauke: oh yeah..forgot... we send in a 64bit rate value from tc to cake and that gets sent correctly..... it's all the stats coming back that basically go missing. Jun 09 11:06:34 with this hack it works? Jun 09 11:06:38 yep Jun 09 11:07:18 so only the kernel to userspace direction is broken? Jun 09 11:07:25 and not the other direction? Jun 09 11:07:36 correct Jun 09 11:07:45 someone reported problems with 64 bit values in iproute2 oon MIPS BE+ Jun 09 11:08:08 ahh, that was me probably. Jun 09 11:08:27 ok Jun 09 11:08:58 they had a number of 'print value' functions that were invisibly promoting 32bit values to 64bit in the print function without checking/changing/using the correct printf format. Jun 09 11:09:15 yes I think that was the problem Jun 09 11:09:54 broke on big endian 'cos the print function was looking in the wrong place for the start of value. Jun 09 11:11:10 I guess that's one of the differences/'advantages' of LE... the Least significant byte is always first - you don't have to know how big a variable is. Jun 09 11:11:34 or more... the least significant byte doesn't move place Jun 09 11:13:16 but I don't think this issue is directly endian related anyway... checked on the MIR3G which is mips32le and I still get the same problem. Jun 09 11:15:43 ldir: how does this ATTR_PAD work? Jun 09 11:16:25 Am I allowed to say 'I have no idea'? Jun 09 11:19:49 I also do not see the problem there Jun 09 11:20:09 but it is also the first time I saw thia attribute padding for 64 bit values in netlink Jun 09 11:20:15 Hauke: you've got mail ;-) Jun 09 11:22:31 looking at the netlink packet would also be my next sugestion Jun 09 11:22:54 there is kmod-nlmon for it Jun 09 11:24:12 ok I'll look into kmod-nlmon Jun 09 11:24:39 it allows you to dump netlink like a network device Jun 09 11:24:49 then you can analyse it in a recent wireshark version Jun 09 11:24:57 2.6.0 or later Jun 09 11:25:32 sounds like configuration fun :-) Jun 09 11:27:59 * ldir upgrades wireshark version Jun 09 11:31:42 ldir bufferbloat/cake is why I started flashing openwrt to. Jun 09 11:32:17 i did it for the blinky lights Jun 09 11:32:46 I first flashed ddwrt because I had a e3000 that had a shit webinterface that would not work with my screenreader then I found the bufferbloat blog on the web and flashed over to OpenWRT Jun 09 11:33:13 blinky lights are no good for me hahahahah Jun 09 11:33:20 :P Jun 09 11:33:40 Can I get a router to play me a tune or somthing? Jun 09 11:33:57 pwm one of the inductors? Jun 09 11:34:31 I did get a wrt54gl to make me some smoke one time! that was a bit of a shok Jun 09 11:34:34 What could be the reason of a package, showing up in feeds/packages, but not in menuconfig? Jun 09 11:34:41 (currently trying to enable aria2( Jun 09 11:34:42 )* Jun 09 11:35:26 and this is what happens to me....I fall over some issue...someone says "have you look at x,y,z?" and off down another rabbit hole I go. Jun 09 11:41:33 ha! so I ponder how to set up this kmod-nlmon and I find myself on Julia Evans' blog. https://jvns.ca/blog/2017/09/03/debugging-netlink-requests/ Wonderfully geeky https://www.youtube.com/watch?v=0IQlpFWTFbM Jun 09 11:42:12 he found another mistress :'( Jun 09 11:44:08 she's antsy on the stage lol. almost a rapper with all the gestures Jun 09 11:44:45 if i install the file system kernel module does that improve performance? Jun 09 11:44:48 https://www.youtube.com/watch?v=HfD9IMZ9rKY Jun 09 11:44:50 such as kmod-squashfs Jun 09 11:45:56 kmod-fs-squashfs Jun 09 12:22:19 if procd is pid 1, why is /sbin/init still exist in the rootfs? will 'preinit+procd' suffice? Jun 09 12:47:44 Hi all, what happened to lede-dev? Why is it invite only? Are we all using OpenWRT-dev again? Jun 09 12:48:09 and one quick question what execatly does scripts/config/conf do? Jun 09 12:52:35 because the merger has progessed Neighbor11111112 Jun 09 12:53:15 Cool thanks Borromini, is there a place to read the details of the merger? Jun 09 12:54:47 the mailing list, there's nothing centralised afaik Jun 09 13:03:19 Ok great thanks Borromini. I need to keep more of an eye on that Jun 09 14:13:47 Hauke: ping Jun 09 14:23:34 Hi, I having some issue with my OpenWRT configs... The DNS resolver is very very slow in one of my wireless networks...Anyone is available to help me please..(my openwrt configs https://forum.lede-project.org/t/very-slow-dns-resolver-wireless-network-2-ssid/15447) Jun 09 14:23:43 I see that for the board I'm trying to add, the makers of that board have forked uboot; https://github.com/helios-4/u-boot-marvell , would it be a good idea, or would it even be possible, to somehow port the changes to OpenWRT? Jun 09 14:35:22 ldir: pong Jun 09 14:36:29 Hauke: ok, so after nearly drowning under a sea of packets... think I've a capture of a working v broken netlink response from cake. Jun 09 14:37:42 and I'm sort of at 'now what' 'cos it doesn't decode in wireshark 'cos wireshark doesn't speak cake attributes... fair enough. Jun 09 14:38:51 they're pretty similar, other than the 64 bit one having an extra 4 bytes here & there for 64 bit values. Jun 09 14:40:04 ok Jun 09 14:41:17 I personally guess/think this is a bug in iproute2 parsing of nested netlink attributes Jun 09 14:42:52 this sounds at least more linkly than a bug in the kernel Jun 09 14:47:46 anyway I've had enough for today. Jun 09 15:22:54 Hey Jun 09 15:23:21 Does anyone here have experience with the soldering part on a HH5A? Jun 09 15:24:24 Because I knocked the resistor at r77 off and I was wondering if it was required Jun 09 15:33:25 As a general rule if they put it there, it needs to be there. Jun 09 15:33:52 Damn Jun 09 15:33:59 Well I guess rip this board Jun 09 15:34:35 You should really get one preflashed unless you're proficient at this sort of soldering. Jun 09 15:45:27 Have to say I soldered to the nearby pads when I did mine instead of to the resistors - like this https://equk.co.uk/2016/10/04/unlocking-bt-homehub-5a/ Jun 09 15:46:46 They're actually vias and are quite delicate themselves. Jun 09 15:48:13 That's what I was trying to do Jun 09 15:48:20 But I knocked the resistor off Jun 09 15:49:23 Also how do I find out what com port it's on? Jun 09 15:51:15 When I turn the board on all the lights flash and then it dies Jun 09 15:52:09 But it lights up when I'm not pulling r45 to ground Jun 09 15:52:17 So I guess the board isn't dead Jun 09 15:52:32 It's not dead but it's also not going to work. Jun 09 15:53:58 I'm back, the router lives somehow, but I'm not getting a serial output Jun 09 15:54:11 Because you removed a needed component. Jun 09 15:54:23 the serial port isn't going to work - I don't know if r77 is a pull up or pull down Jun 09 15:54:49 That sucks Jun 09 15:55:05 Ah well, it was fun to try Jun 09 15:59:34 hi every one Jun 09 16:00:18 i just extract last brcmfmac4366c-pcie.bin Jun 09 16:00:35 some one is interesed to have it Jun 09 16:03:11 https://pastebin.com/XBCxYDKH Jun 09 16:05:44 So in conclusion I killed it? Jun 09 16:05:53 unless I can resolder R77 Jun 09 16:07:08 Simber2: no, I'm sure the device is fine... but you won't be able to put openwrt on it without a working serial port. Jun 09 16:08:12 the fact the component was 'knocked off' suggests too much heat applied for too long. Jun 09 16:09:39 I was meaning killed the serial, it still boots but when I pull R45 to ground to boot into uBoot (or whatever its called) it just dies straight away Jun 09 16:09:39 so before you do anything else, if you're going to try to repair, get a more suitably size iron and tip. Jun 09 16:10:01 Yeah, the iron I was using was bigger than the resistor Jun 09 16:10:13 and older than me and I can drink legally Jun 09 16:10:29 and the solder was older than me too Jun 09 16:10:50 it doesn't die....it has gone into it's 'boot from serial port' boot mode. Jun 09 16:11:03 Oh Jun 09 16:11:15 It just flashes the lights for a split second Jun 09 16:11:20 yep Jun 09 16:11:48 if you had a working serial port......... you get the picture Jun 09 16:12:10 oh ok Jun 09 16:16:57 Well time to get someone better than me at soldering to resolder r77 Jun 09 16:19:16 well a more suitably sized iron & tip for this sort of work at least ;-) Jun 09 16:29:17 It might not be entirely dead Jun 09 16:29:25 I got this when I started it http://prettyawful.net/u/2018-06-09_17-29-02.png Jun 09 16:29:48 Im having to hold the ground in place as the solder broke off Jun 09 16:29:52 Yes, that's called RFI. Jun 09 16:38:03 That took a while Jun 09 16:48:57 The most I got out of it was a few random chars http://prettyawful.net/u/2018-06-09_17-48-38.png Jun 09 16:52:36 is that all because of RFI? Jun 09 16:52:49 i've seen that stuff before myself... not a hero with serial either Jun 09 17:12:20 ok, zbt we1326 deployed live at last :) Jun 09 17:12:35 still have good wifi in the corner of our bedroom, and no longer have to pay monthly router rental Jun 09 17:12:36 Fizzle and pop in 3, 2, 1.. Jun 09 17:14:48 upload speed is slower than last time, but... that could just be the speed server testing, nothing provable Jun 09 17:34:47 Hello! :D Jun 09 17:36:31 I have my modem set up in bridge mode to my tplink 7c with LEDE, the weird thing is that another router (asus rt68u) is setup as access point and connected to the LEDE router. When I connect with the access point i'm online but not with and without cable when im connected to the LEDE router Jun 09 17:37:01 Is there anything i should setup in the interface settings according to the LAN WAN or WAN6 Jun 09 17:37:32 Kind of odd since my access point is also connected to the tplink same as my laptop but does get internet Jun 09 17:46:11 sounds like a routing issue almost Jun 09 17:46:16 but i'm no expert on it Jun 09 17:54:21 morning Jun 09 17:55:48 morning :-) Jun 09 17:56:07 evening Jun 09 18:07:07 ldir: Just thinking would it work with only rx Jun 09 18:07:54 you need to upload a new bootloader which is a tx operation. Jun 09 20:05:28 goodness it's quiet in here Jun 10 00:20:38 build #697 of armvirt/32 is complete: Failure [failed pkgbuild] Build details are at http://phase1.builds.lede-project.org/builders/armvirt%2F32/builds/697 blamelist: Florian Eckert , David Bauer , Sven Eckelmann , Jeremiah McConnell , Kevin Darbyshire-Bryant @darbyshire-bryant.me.uk>, Matthias Badaire , Martin Schiller Jun 10 02:48:07 build #888 of brcm2708/bcm2708 is complete: Failure [failed pkgbuild] Build details are at http://phase1.builds.lede-project.org/builders/brcm2708%2Fbcm2708/builds/888 blamelist: Florian Eckert , David Bauer , Sven Eckelmann , Jeremiah McConnell , Kevin Darbyshire-Bryant Jun 10 02:48:07 , Matthias Badaire , Martin Schiller **** ENDING LOGGING AT Sun Jun 10 03:00:04 2018