**** BEGIN LOGGING AT Sat Jun 23 03:00:01 2018 Jun 23 03:03:43 build #40 of apm821xx/nand is complete: Success [build successful] Build details are at http://release-builds.openwrt.org/18.06/images/builders/apm821xx%2Fnand/builds/40 Jun 23 03:03:43 build #39 of malta/be is complete: Success [build successful] Build details are at http://release-builds.openwrt.org/18.06/images/builders/malta%2Fbe/builds/39 Jun 23 03:50:21 build #40 of ramips/rt3883 is complete: Success [build successful] Build details are at http://release-builds.openwrt.org/18.06/images/builders/ramips%2Frt3883/builds/40 Jun 23 04:22:01 build #38 of ar71xx/mikrotik is complete: Success [build successful] Build details are at http://release-builds.openwrt.org/18.06/images/builders/ar71xx%2Fmikrotik/builds/38 Jun 23 05:19:25 build #37 of ixp4xx/harddisk is complete: Success [build successful] Build details are at http://release-builds.openwrt.org/18.06/images/builders/ixp4xx%2Fharddisk/builds/37 Jun 23 05:42:52 build #37 of cns3xxx/generic is complete: Success [build successful] Build details are at http://release-builds.openwrt.org/18.06/images/builders/cns3xxx%2Fgeneric/builds/37 Jun 23 05:52:24 build #37 of ar71xx/tiny is complete: Success [build successful] Build details are at http://release-builds.openwrt.org/18.06/images/builders/ar71xx%2Ftiny/builds/37 Jun 23 06:30:56 build #39 of mvebu/cortexa9 is complete: Success [build successful] Build details are at http://release-builds.openwrt.org/18.06/images/builders/mvebu%2Fcortexa9/builds/39 Jun 23 06:42:25 morning Jun 23 06:43:26 jow: we need to restart the builders for rc2 Jun 23 06:43:40 hi guys Jun 23 06:43:51 i think we can live with either building oxnas/mt7623 manually or simply skipping them for rc1 Jun 23 06:43:55 Borromini: aloha Jun 23 06:44:18 so i dropped that mac80211 PS patch and received 4 mails from people saying thank you Jun 23 06:45:01 :) Jun 23 06:46:58 i may have been affected. cool. Jun 23 06:49:08 it looks like the "wifi drops out after 1-3 days" issue is caused by that patch Jun 23 06:49:24 i recently setup gramofon routers for spotify streaming Jun 23 06:49:36 they start to jitter the music on day 2 Jun 23 06:49:45 dropping that patch from my c7 and all works fine Jun 23 06:51:02 :) Jun 23 06:54:15 blogic: are you happy with gramofon? my brother has now a rpi3 with hifiberry. it's okay but a bit clumsy (not to mention raspbian is quite shoddy when it comes to software/security updates) Jun 23 06:57:01 love it Jun 23 06:57:09 i have 4 in total in the house Jun 23 07:22:51 Borromini: gninrom Jun 23 07:24:52 ultito: ? Jun 23 07:25:57 Borromini: read it backwards :-) Jun 23 07:28:23 it's to early for this :P Jun 23 07:37:48 opkg update Downloading http://downloads.lede-project.org/releases/17.01.4/targets/ramips/mt7620/packages/Packages.gz Failed to establish connection Jun 23 07:38:02 what's wrong with my router Jun 23 07:38:31 my pc can access to internet Jun 23 07:39:07 lk_: Hosed DNS maybe? Jun 23 07:39:53 it doesn't affect my pc to access internet ?/ Jun 23 07:40:21 It's quite possible that your PC is pointing at working DNS servers, while your router itself isn't... Jun 23 07:42:30 FOr example, my router is running dnsmasq that's glomming off Comcast's DNS infrastructure, while this laptop is running unbound that does its own resolution .. Jun 23 07:42:47 i see Jun 23 07:43:02 i used dnscrypt-proxy in my router Jun 23 07:43:50 Whuch can get fun when the other laptop is listening to dnsmasq on the router, so the two laptops end up with different views of DNS.. Jun 23 07:47:19 i can ping the dns in my router Jun 23 07:47:35 but root@LEDE:/# ping qq.com ping: bad address 'qq.com' Jun 23 07:47:53 lk_: Go look at what's in /etc/resolv.conf on both. Jun 23 07:48:56 search lan nameserver 127.0.0.1 Jun 23 07:49:42 OK.. So one box trusts some local process (probbably dnscrypt-proxy). Jun 23 07:50:12 i using the opendns Jun 23 07:50:44 Other possibility is that 'bad address' is because you only have IPv6 available, and qq.com just has 2 A records and no quad-A.s Jun 23 07:51:06 and/or you have busted IPv4 routing. Jun 23 07:52:05 * ldir curses his typing this morning - mind you *is* on a train. Jun 23 07:52:34 so what should i do i new to openwrt Jun 23 07:53:24 Hmm.. Interesting. Jun 23 07:53:34 ping qq.com Jun 23 07:53:34 PING qq.com (111.161.64.40) 56(84) bytes of data. Jun 23 07:53:34 64 bytes from dns40.online.tj.cn (111.161.64.40): icmp_seq=1 ttl=44 time=254 ms Jun 23 07:53:34 64 bytes from dns40.online.tj.cn (111.161.64.40): icmp_seq=2 ttl=44 time=259 ms Jun 23 07:53:42 Forward and reverse DNS don't match. Jun 23 07:54:43 ping 111.161.64.48 Jun 23 07:54:43 PING 111.161.64.48 (111.161.64.48) 56(84) bytes of data. Jun 23 07:54:43 64 bytes from 111.161.64.48: icmp_seq=1 ttl=44 time=296 ms Jun 23 07:54:43 And pinging the other A record listed for qq.com hits something that doesn't even have a PTR entry. Jun 23 07:55:31 yes i can ping the 111.161.64.48 Jun 23 07:56:13 i get the response Jun 23 07:57:23 may i get your telegram @valdisk Jun 23 07:58:08 telegram? Even Western Union got out of that business like 20 years ago. ;) Jun 23 07:59:08 me too but i using the shadowsocks so that i can use telegram Jun 23 07:59:59 :D Jun 23 08:02:50 jow: **** me! That was quick! :-) Jun 23 08:04:25 jow: consider for 18.06 ? Jun 23 08:17:02 Anyhow - there any known issues with IPv6 and neighbor discovery in recent lede-snapshot? Jun 23 08:26:06 can I build openwrt on a running openwrt system? Jun 23 08:26:45 apacar: Well.. only if it's a beefy enough system to handle the build and you installed all teh SDK pieces you'll need. ;) Jun 23 08:27:06 I hate to think how long a build would take on a WRT1200AC. ;) Jun 23 08:27:30 valdisk I'm on that, but seems ncurses, perl thread::queue module and "a file stat utility" Jun 23 08:27:58 libncurses nor libncurses-dev seem to help Jun 23 08:29:00 It's quite possible the SDK config is busted due to lack of use/testing... Jun 23 08:29:35 why would you run a build on a non-x86 system? :-( Jun 23 08:31:01 Borromini: Well, it's plausible to attempt it on a high-end ARM system, but yeah, even a low-end x86_64 laptop would be faster. Jun 23 08:32:05 Borromini I'm running in a x86 vm running openwrt Jun 23 08:32:47 aparcar: Why not spin up another VM running a proper full devel environment? :) Jun 23 08:32:47 oh, so stat is enabled in busybox/coreutils, that makes it a bit difficult Jun 23 08:33:07 luaraneda: as we talked yesterday, that may be true on your board, but the c2600 has an empty ubootenv partition Jun 23 08:35:50 valdisk well just testing, I though it be nice for CI to run a smaller OS Jun 23 09:58:27 hi what packages do i need to install an sd card under routerstation pro using lede 17.03 Jun 23 10:26:34 obinoob update to lede 17.4 or 18.6 first and then I dont know about what packages you will need sorry. try the wiki Jun 23 10:26:55 I have never used lede-openwrt on a sd card Jun 23 10:28:01 Tapper: I'm using lede-openwrt on a sd card I need just to add some storage and mount the sd card at /opt or whatever Jun 23 10:28:18 Tapper: I mean I'm not Jun 23 10:28:58 It mite be a bug in 17.3 Jun 23 10:29:19 you will nede block-mount Jun 23 10:30:05 and the package for what ever FS you are using Jun 23 10:34:35 Tapper: I've managed to mount the sd card, I need to format the card ext3 or ext4 what tools do I need any idea? Jun 23 10:38:34 obinoob: if you've managed to mount the card under owrt, then you're set Jun 23 10:40:24 movi: I've mounted the card but now I need to format to ext3 / 4 Jun 23 10:41:01 well yeah Jun 23 10:41:15 actually wait, no Jun 23 10:41:19 you said you've mounted it Jun 23 10:41:22 which fs? Jun 23 10:41:32 also, this is better suited for #openwrt Jun 23 10:45:37 movi: sorry for the confusion I can see the partition etc but not mounting since I need to format the card I've installed e2fsprogs but fsck.ext4 /dev/sdb1 is outputing some errors Jun 23 10:45:52 obinoob: switch to #openwrt Jun 23 10:56:32 hey i have openwisp setup and it works fine except wifi clients don't get an IP address, any caveats when just configuring two radios? i didn't change anything else Jun 23 10:56:47 two radios and two wifi networks, i just set up the SSID's and security Jun 23 11:19:56 How to disable usbip ? https://github.com/openwrt/packages/pull/6062 I tryied CONFIG_PACKAGE_kmod-usbip=n CONFIG_PACKAGE_kmod-usbip-server=n CONFIG_PACKAGE_kmod-usbip-client=n but without success Jun 23 11:47:38 jow: ping Jun 23 11:49:08 Can anoyne merge https://github.com/openwrt/luci/pull/1422 to lede-17.01 branch? Jun 23 12:45:07 i'm getting following error from PHP when trying to use some date function: Jun 23 12:45:09 Fatal error: DateTime::createFromFormat(): Timezone database is corrupt - this should *never* happen! Jun 23 12:46:04 this is with CONFIG_PHP7_SYSTEMTZDATA=y which made PHP compiled with --with-system-tzdata Jun 23 12:46:13 this didn't happen with 17.01 Jun 23 12:46:17 it's present in 18.06 Jun 23 12:46:37 does anyone know how PHP tries to get timezone when copiled with --with-system-tzdata? why it doesn't work with OpenWrt? Jun 23 13:29:28 this seems to be caused by a missing /usr/share/zoneinfo/ Jun 23 13:29:39 open("/usr/share/zoneinfo/UTC", O_RDONLY|O_LARGEFILE) = -1 ENOENT (No such file or directory) Jun 23 13:29:40 writev(1, [{iov_base="", iov_len=0}, {iov_base="\nFatal error: DateTime::__constr"..., iov_len=126}], 2 Jun 23 13:29:42 Fatal error: DateTime::__construct(): Timezone database is corrupt - this should *never* happen! in /root/test.php on line 3 Jun 23 13:29:43 ) = 126 Jun 23 13:59:49 Hi all, has anyone tried running OpenWRT on a UniFi nanoHD? Jun 23 14:06:37 someguy: no, and 5ghz might be impossible at the moment, as there's no open source driver yet Jun 23 14:06:46 it'S MT7615E Jun 23 14:24:20 Thanks rotanid. Jun 23 14:44:45 userip ausjke Jun 23 14:45:12 rmilecki: did you install tzdata? Jun 23 14:45:27 jow: i didn't Jun 23 14:45:36 shouldn't dependency handle that? Jun 23 14:45:55 apparently not Jun 23 14:45:58 I also think that with 17.01 it was working even without /usr/share/zoneinfo/ Jun 23 14:47:50 stupid question: are the RT5350F-based routers better or worse than rpi zero ws for openwrt? Jun 23 14:48:21 rmilecki: not sure Jun 23 14:48:30 rmilecki: in any case you need tzdata as php is patched to use it Jun 23 14:49:17 i can confirm in 17.01 it works without /usr/share/zoneinfo/ Jun 23 14:50:40 maybe php simply added additonal checks Jun 23 14:50:50 or its some uclibc vs. musl runtime behaviour thing Jun 23 14:51:03 17.01 uses musl, doesn't it? Jun 23 14:51:12 ah right Jun 23 14:51:15 i'm looking at strace from 17.01 and I can see open("/etc/TZ", O_RDONLY|O_NONBLOCK|O_LARGEFILE|O_CLOEXEC) = 3 Jun 23 14:51:21 so it's also checking timezone Jun 23 14:51:31 but it doesn't try to access /usr/share/zoneinfo Jun 23 14:57:06 the package is supposed to depend on zoneinfo-core when CONFIG_PHP7_SYSTEMTZDATA is set Jun 23 14:57:47 oh? not the opposite? Jun 23 14:58:02 but the dependency is not propagated to opkg Jun 23 14:58:10 so it only works when php is built into the image Jun 23 14:58:15 uh, weird... Jun 23 14:58:40 jow: oh, that explains much Jun 23 14:58:44 it has manual kconfig snippets that "select PACKAGE_zoneinfo-core" if CONFIG_PHP7_SYSTEMTZDATA Jun 23 14:59:00 but it has no DEPENDS:=+PHP7_SYSTEMTZDATA:zoneinfo-core Jun 23 14:59:07 do you know why/ Jun 23 14:59:21 oversight, I think mhei simply isn't aware of this Jun 23 14:59:48 i'll ping maintainer then Jun 23 14:59:53 thanks for looking at it Jun 23 15:00:39 yw Jun 23 15:02:22 pc lost connection to zbt1326 over 2.4G for 1-2 minutes once a while, had to manually reconnect to get going, cellphone shows RSSI always fine so wifi is on, connect via 5G to check, nothing unusual Jun 23 15:44:37 Which USB WLAN adapters does OpenWRT support? Jun 23 15:50:49 Hi all, what would be the best way to get feedback on ongoing work to port a device to another target (WNDR4300 from ar71xx to ath79)? Jun 23 16:00:16 Hertog_Jan hi is that TP-Link? Jun 23 16:00:28 if so then see here: Jun 23 16:00:30 https://forum.lede-project.org/t/ath79-fully-working-builds-with-all-kmod-packages-through-opkg-flow-offloading/15897 Jun 23 16:12:33 Hertog_Jan: forum or inside a PR Jun 23 16:13:04 Hertog_Jan: how far are you with porting ? i have a wndr4300 here on my desk but wont even start if someone is already working on the port Jun 23 16:16:01 * ldir has just submitted his most complicated ever PR to luci. Jun 23 16:21:12 192.168.254 Jun 23 16:22:53 ldir: it sound more G before Jun 23 16:24:04 * ldir chuckles Jun 23 16:24:35 * ldir and smiles & waves ;-) Jun 23 16:24:54 i've just prepared 28 patches for linux-mips, including all our current ath79 patch set Jun 23 16:25:13 need to add description to a few of the patches Jun 23 16:25:15 and then send it Jun 23 16:25:16 yikes Jun 23 16:25:21 but now the kids is nagging me .... Jun 23 16:25:33 good luck! With both. Jun 23 16:28:43 * ldir has just noticed firewall rule names are not appearing in luci firewall traffic rules... eyes 5142e40f9e70a09653e07a4dbc2abd6db3da36e4 with suspicion. jow. Jun 23 17:12:43 how can different vlan share one printer... Jun 23 17:13:15 by putting it on both vlans Jun 23 17:13:35 or... just make sure its reachable via ip Jun 23 17:15:11 normally the printer-search will search inside subnet and does not cross the boundary, need make sure the printer is transparent to all under the hood Jun 23 17:15:27 idiot proof that is Jun 23 17:15:58 modern ones use mdns etc Jun 23 17:16:12 but you can still just give an ip/hostnamw Jun 23 17:16:21 hostname Jun 23 17:16:25 maybe put printer, chromecast, kindle fire, etc all on one vlan, for all the public sharing devices Jun 23 17:16:37 pr maybe proxy the mdns if you want Jun 23 17:16:42 or Jun 23 17:16:48 yeah Jun 23 17:17:11 you can do ip and macvlams also Jun 23 17:17:22 which may be useful Jun 23 17:17:27 then proxy-arp to that public-device-vlan? Jun 23 17:17:48 dont need to do that Jun 23 17:18:01 only need discovery if you want auto config Jun 23 17:18:20 mdns is also within subnet only right? Jun 23 17:18:28 i personally just have printer.domain Jun 23 17:18:42 yeah but its multicast, you can relay it Jun 23 17:18:56 hmm Jun 23 17:19:16 igmpproxy or something ? Jun 23 17:19:27 never relayed multicast yet Jun 23 17:19:31 guess so Jun 23 17:19:36 theres a few tools Jun 23 17:19:49 probably an mdns specific util Jun 23 17:21:17 looks like theres a couple Jun 23 17:22:19 that might be the preferred approach, as zeroconf is enabled on nearly all devices these days, it seems Jun 23 17:31:09 * ldir realises that supporting Sweden may well delay any/all of his PRs :-) Jun 23 17:32:21 * ausjke is impressed rc1 was out in the middle of soccer frenzy Jun 23 18:32:57 oooops! Jun 23 18:35:21 still only rc1? :-D Jun 23 18:42:18 ldir: i'd laugh my arse off id germany looses Jun 23 18:42:33 i did not even realize they are playing, i am a football agnostic Jun 23 18:42:48 its not even that i dont like it, it simply does not exist in my reality Jun 23 18:43:08 and i dont own a TV so even if i wanted to, i could not watch Jun 23 18:43:18 the streets here are empty Jun 23 18:43:44 you could streak through towncenter and no one would even notice Jun 23 18:44:20 professional sports frenzy is great to reduce wars, so coders can code in peace Jun 23 18:46:34 coding in peace sounds like a plan Jun 23 18:50:31 blogic: "if germany loses" ... looks like that is going to happen. and then they're out. Jun 23 18:51:34 drmr: that would be rather funny Jun 23 18:51:57 i've not heard any fireworks go off so i am guessing no goals yet :-D Jun 23 18:52:09 it would definitely make my days quieter. my office is straight above a football bar. Jun 23 18:52:50 blogic: right now? halftime, sweden is 1:0 in the lead. they probably don't have disappointment fireworks. Jun 23 18:54:05 (and no, not particularly interested in football either.) Jun 23 18:55:36 I don't watch it, even though I frequently work on it. Later tonight I'm bringing an American Major League Soccer match to the world. Jun 23 19:21:18 https://github.com/openwrt/luci/commit/5142e40f9e70a09653e07a4dbc2abd6db3da36e4#diff-a943bdadbc124679501897ac9081f4b3R75 has b0rked display of rule names in traffic rules & similarly in port forwards. Jun 23 19:21:51 looks like https://github.com/openwrt/luci/commit/5142e40f9e70a09653e07a4dbc2abd6db3da36e4#diff-a943bdadbc124679501897ac9081f4b3R75 has b0rked display of rule names in traffic rules & similarly in port forwards. Jun 23 19:22:25 ooops - didn't mean to repeat myself. scrollback fail. Jun 23 19:41:16 * ldir appears to have an archer c7 v2 running 4.14 in ath79 - trying that PR out. Jun 23 19:44:23 blogic: Re: wndr4300 port, I just started essentially. I have created a very basic dts. The device boots with that, but doesn't yet reach a prompt. On the forum someone mentioned that the NAND driver would need to be ported as well Jun 23 19:46:06 Also I have only started today, so I have spent some time on the dt wiki pages to see how that kind of works Jun 23 19:52:40 Hertog_Jan: well good luck, keep us posted on the progress Jun 23 19:54:08 Will do! Jun 23 20:04:51 jow: what values is nlbwmon looking at exactly? the stuff it prints out here are kinda out whack Jun 23 20:14:00 jow: i'm downloading multigig files, but nlbwmon is only seeing a couple of kilobytes? :/ Jun 23 20:21:25 ah i figured it out, it doesn't work with ipt_offload :/ Jun 23 20:21:38 movi: it will soon Jun 23 20:21:52 i have a patch in the making that does flow accounting for flow offloading Jun 23 20:21:57 blogic: oh btw what's the difference between ipt_oflload and nft_oflload? Jun 23 20:22:10 *offload Jun 23 20:22:16 hello fellas i followed WR741ND unbrick guide with my FTDI and successfully re-wrote the firmware image but after i used the 'boot' cmd on bootloader it never went back.. not even to bootloader shell! what do you suggest? thanks Jun 23 20:22:42 xtrWrithe: throw it away and buy a new one Jun 23 20:22:44 blogic: also ooh please share :P Jun 23 20:22:54 if you fragged the bootloader there is no return Jun 23 20:23:02 blogic: why you are sure i did that? Jun 23 20:23:05 movi: will do once functional Jun 23 20:23:16 one if using nftables, the other iptables - upstream flow offload code only supports nftables, OpenWrt's fw3 currently can only deal with iptables - so ipt_offload is all you need right now Jun 23 20:23:31 because i have been doing embedded evelopement for 15+ years and if the uboot prompt does not show you have a real brick Jun 23 20:23:39 pkgadd: exactly Jun 23 20:23:51 i just did the boot command exactly as shown in wiki and then it cameback to bootloader normal loop , when i taked off the uart bridge it didnt came back Jun 23 20:23:51 pkgadd: thanks for that, got kinda confused Jun 23 20:24:16 pkgadd: will fw3 switch to nft at some point, stay with ipt or wait for that newfangled berkeley thingie? Jun 23 20:24:25 blogic: ok bro but its like i wanna know the problem of the dead , ofc i will get a new one soon if this keeps like this Jun 23 20:24:29 movi: good question Jun 23 20:24:53 xtrWrithe: if you see no boot prompt even after power cycle you most likely fragged uboot somehow Jun 23 20:25:19 :( Jun 23 20:26:00 hey somebody managed to load some BSD with uboot on tplink hardware? Jun 23 20:51:12 blogic: quick question about QCA8K: as far as I can see it doesn't have support for multiple CPU ports yet - can you confirm that? Jun 23 20:51:41 (use-case: one of my boards has a QCA8337 with 4x LAN and 1x WAN ports, LAN is using a different CPU port than WAN) Jun 23 20:51:58 xdarklight: upstream DSA does not support multiple cpu ports Jun 23 20:52:08 mediatek target has a patch to make it work Jun 23 20:52:26 that would need to be added and then the qca8k needs a minor update aswell Jun 23 20:52:41 i already have that included in my ipq806x qca8k branch Jun 23 20:54:23 I see, you started a discussion about multiple cpu ports last year... Jun 23 20:54:31 correct Jun 23 20:54:37 but netdev is soooo complicated Jun 23 20:54:57 its near impossible to work with them unless you are in the inner circle or working on something they are not working on Jun 23 20:55:10 but trying to patch something they are actively working on does not work Jun 23 20:55:15 its like a sandpit Jun 23 21:01:20 blogic: https://patchwork.kernel.org/patch/9754271/ in case I missed it (netdev archives are big...): do you know if [florian] (second-last reply in that thread) ever followed up with a patch? Jun 23 21:02:55 xdarklight: he did not Jun 23 21:03:14 so i had v1, got nak'ed with suggestion of how v2 should look Jun 23 21:03:43 then new release, some time went by, by the time the window opened for V2 they changed so much that their idea of how to solve it was stale and a v3 was required Jun 23 21:03:55 a v3 was as you saw announced and never happened Jun 23 21:04:14 yeah, I've read through the whole v2 thread and I (think I) understand their point Jun 23 21:04:29 i then asked via private mail, got a 3rd suggestion of how to implement and then for some reason never followed up and decided to just carry my own patch in owrt Jun 23 21:04:38 xdarklight: it is valid Jun 23 21:04:48 the problem is that the validity of the point changes Jun 23 21:14:11 xdarklight: which board if i may ask? i'm compiling a build with this https://git.openwrt.org/?p=openwrt/staging/blogic.git;a=commitdiff;h=9b2f86fdbd59c0b54a958cc5d0097d785bea3bae;hp=51740777fb37cb7bdc250d74b366840269439cf3 Jun 23 21:14:19 and i'm not quite sure if it'll actually work tbh :F Jun 23 21:15:40 blogic: ok, [florian] is ping'ed too - maybe he still has plans to send that patch (and just forgot about it last year). let's wait for his reply too :) Jun 23 21:16:20 movi: it's an Belkin f9k1115v2 (trying to port it to the ath79 target) Jun 23 21:17:02 there's a ath79 with a dual core? :O Jun 23 21:17:45 it's single core, where does it say dual core? :o Jun 23 21:18:20 ah sorry, i got confused because of the dual port thing Jun 23 21:19:18 ah, that just means that the SoC has to ethernet controllers and that board connects both ethernet controllers to the switch Jun 23 21:19:28 *two ethernet controllers Jun 23 21:19:56 to the same switch, right? Jun 23 21:20:05 yep Jun 23 21:20:08 why? Jun 23 21:20:13 :D Jun 23 21:20:34 money Jun 23 21:20:39 sorry, performance obviously ;) Jun 23 21:21:15 no really, money (the switch has 5 Gbit PHYs built in - they could have added a separate PHY instead but that would have meant higher bill-of-material costs) Jun 23 21:21:17 in what way? you only have a single cpu? i'm sorry if this sounds dumb Jun 23 21:22:46 performance in a sense where they could (assuming that the CPU has enough cycles to do it...) run 1Gbit/s of data on the WAN port from/to the CPU and another 1Gbit/s of data from/to the CPU from the LAN ports (= total of 2Gbit/s) Jun 23 21:23:24 which wouldn't be possible with just one link to the switch (1 link = 1Gbit/s on QCA8337) Jun 23 21:23:52 assuming the cpu is might enough, you sure you can sum these up? Jun 23 21:24:14 I need help choosing between these routers. I would prefer the one with higher CPU frequency, at least 2 cores and MIMO. I also need it to work stable with OpenWRT. https://usercontent.irccloud-cdn.com/file/2Df5x8gV/Untitled-1.png Jun 23 21:26:27 movi: that's why CPU cycles are required: it's not about sending data from WAN to LAN (or vice versa) with the help of the CPU - it's (for example) about sending 0xaa from the CPU to WAN while receiving 0xbb from LAN simultaneously Jun 23 21:27:29 you will also find that the gmac linking to the lan ports is not gbit but faster allowing you to saturate gbit on more than 1 port Jun 23 21:27:59 if you had a single gbit mac doing rx and tx for wan and lan ports then the max throughput lan->wan would be 500 mbit Jun 23 21:36:21 mangix: temperature: 499 on mwlwifi Jun 23 21:50:12 blogic: do you have the same datasheet as "we" have for mt7621? Ones from different sites or do you have the full package? Jun 23 21:50:50 i have one for the gsw part Jun 23 21:50:58 and another for the core registers i think Jun 23 21:51:08 there is no full package afaik Jun 23 21:51:26 i downloaded mine directly from the MTK support site Jun 23 22:00:45 blogic: Is that open for everyone? Jun 23 22:01:23 nope Jun 23 22:02:56 what info are you after ? Jun 23 22:03:02 maybe i can look it up for you Jun 23 22:06:53 For me it is unclear how the gsw is connected/interacts with an external phy like AR8033. Jun 23 22:07:58 port 6 can be switched between internal and external phy Jun 23 22:08:57 or was it 5 ? Jun 23 22:09:07 it is 5 Jun 23 22:09:29 gmac 5 can switch between port 0, 4 of external phy. Jun 23 22:09:38 correct Jun 23 22:10:58 To send mdio command to the external phy I have to do that via gsw registers? Jun 23 22:11:29 no Jun 23 22:12:15 target/linux/ramips/dts/Y1S.dts Jun 23 22:12:20 look at the gsw node Jun 23 22:12:29 it tells the gsw to use gmac mode Jun 23 22:13:00 and sets up ethernet to use normal phy4 for wan Jun 23 22:13:16 and once you have done so the kernel can talk to the ar8033 directly on phyaddr 4 Jun 23 22:14:03 Ok Jun 23 22:15:52 Y1S.dts is a reference for your board Jun 23 22:16:03 simply copying the ethernet and gsw nodes should work Jun 23 22:16:11 and remember to add the ar8033 driver Jun 23 22:16:18 AND important, the reset gpio Jun 23 22:16:29 ar8033 silicon is broken and required a reset gpio Jun 23 22:17:53 thanks for the information. I hope I can use the SFP port on my Ubiquity ER-X-SFP. Jun 23 22:18:02 ah Jun 23 22:18:08 there is a wiki post about that Jun 23 22:18:14 its been on my todo for a while Jun 23 22:18:28 you need to use i2c to trigger some gpio iirc Jun 23 22:21:29 ok Jun 23 22:27:30 what is phy address 0x1f? That is addres first internal phy? Jun 23 22:28:36 the switches control registers are mapped to that addr Jun 23 22:30:14 ok Jun 23 22:31:53 ok i have a 2M hosts file that i want to dnsmasq-ipset, is there a way to generate a binary(e.g. b-tree) of some sort and dnsmasq can use that instead of doing everything in raw txt? Jun 23 22:33:42 https://github.com/infinet/dnsmasq this fork seems is trying Jun 23 23:20:13 argh, i tried dsa, and it kinda works, except for my lan ports not working for some reason :/ Jun 23 23:25:24 movi_noznc: just to make sure: you did reset your config, right? Jun 23 23:25:31 xdarklight: indeed i did Jun 23 23:25:41 (I believe the network config is not auto-migrated from non-DSA to DSA builds or vice versa) Jun 23 23:25:42 xdarklight: it's weird cause on wifi everything is fine Jun 23 23:25:50 xdarklight: no no, i knew of that, nuked everything Jun 23 23:26:32 xdarklight: i think it's something about the default firewall, since they're getting dhcp leases :O Jun 23 23:30:56 xdarklight: any hints? Jun 23 23:34:32 vv Jun 24 01:53:17 dissent1: alright that sounds bogus. **** ENDING LOGGING AT Sun Jun 24 03:00:00 2018