**** BEGIN LOGGING AT Wed Jun 06 03:00:10 2018 Jun 06 06:25:42 so my apu2 rebooted 12 times since flashing it last night and enabling flow offload, definitely something wrong, disabled flow offload to verify it's actually flow offload causing it Jun 06 06:32:13 stintel: I've seen something like that on ipq8065 as well, but couldn't really put a finger on it (but it's 100% stable without flowoffloading) Jun 06 06:42:26 not as in 12 reboots over a night, but a reboot after 1-3 days (and similar on lantiq, although I can't really tell if that were just VDSL sync losses or reboots) Jun 06 06:43:31 uptime? Jun 06 07:22:37 build #13 of mediatek/mt7623 is complete: Failure [failed images] Build details are at http://phase1.builds.lede-project.org/builders/mediatek%2Fmt7623/builds/13 blamelist: Stijn Tintel , Daniel Golle , Ted Hess Jun 06 08:13:54 nbd: /win 4 Jun 06 08:13:58 oops Jun 06 08:14:01 ) Jun 06 08:14:05 :)* Jun 06 08:14:18 nbd: is there a reason why we enable wireless isolation by default ? Jun 06 08:14:43 there's a number of reports that wifi clients on 17.04.* are isolated from each other, despite "option isolate" no being set Jun 06 08:15:27 I've reviewed https://git.openwrt.org/?p=project/netifd.git;a=blob;f=scripts/netifd-wireless.sh;h=8816bff80abf235519309d2a1425a19e2cdaffd7;hb=a580028dae5b7bebcc14c240fe7dac31cbd89355 and noticed that "multicast_to_unicast" defaults to true if unset and forcibly enables isolate Jun 06 08:16:41 I vaguely remember something about that, indeed related to multicast_to_unicast requires isolate to be enabled Jun 06 08:17:03 I understand that, the problem is that it apparently breaks a lot of common use cases Jun 06 08:17:19 like, being able to reach a wireless host :) Jun 06 08:17:43 grepped my irclogs but didn't find anything Jun 06 08:18:07 but I don't disable multicast_to_unicast and my wireless clients can reach eachother just fine Jun 06 08:18:15 so there's probably something else going on Jun 06 08:18:19 the problem is also that luci does not expose "multicast_to_unicast" so users are not aware of it and cannot unset it Jun 06 08:18:31 which effectively means that all default installs enable wireless client isolation Jun 06 08:19:11 maybe this feature works fine with some wireless clients, maybe it requires additional config which happens to be present on your system Jun 06 08:19:20 maybe it conflicts with powersave stuff Jun 06 08:20:00 given that likely noone is going to debug mac80211 for lede 17.01* I'd propose to turn it off by default there adn to extend luci to allow enabling it if needed Jun 06 08:20:38 here's al ongish discussion on the issue: https://forum.lede-project.org/t/clients-in-same-wlan-cant-reach-each-other/2501 Jun 06 08:21:49 when isolation is enabled implicitly, it is replaced by hairpin mode on the bridge interface Jun 06 08:22:00 the bridge is supposed to handle forwarding between hosts in that case Jun 06 08:22:12 which apparently is not properly working Jun 06 08:22:25 maybe because passive hosts entries time out or something Jun 06 08:22:31 i thought we disabled multicast_to_unicast by default at some point Jun 06 08:23:18 multicast_to_unicast is great :) Jun 06 08:23:34 it's just to avoid sending multicast stuff at 1 mbps Jun 06 08:23:44 I know what it is for and what it is supposed to do Jun 06 08:24:01 the problem is that it causes regressions [at least on 17.01.4] Jun 06 08:24:34 I am not opposed to have it enabled by default on 18.x or master if it does not result in broken behaviour there Jun 06 08:42:01 ldir: ping, query? Jun 06 08:42:11 yo dude! Jun 06 08:42:18 yeah go ferrit Jun 06 08:55:16 stintel: disabled flow_offloading and now AC wifi seems to work fine with mt76. enabled hw_flow_offloading, still seems to work fine. so something's up with kernel flow offloading Jun 06 08:55:40 (dir860l b1) Jun 06 08:56:31 paulius: pretty sure that's the case, yes Jun 06 08:56:40 I should hook up a serial cable and get a crash dump Jun 06 08:56:53 this is with nbd's branch from ~23hrs ago Jun 06 08:57:34 shout out to nbd :) Jun 06 08:58:27 although I'm getting only ~290mbits on 5ghz, it rarely goes over 585mbit link speed when downloading Jun 06 08:58:32 seems kind of low Jun 06 08:59:09 3 meter line of sight from the router Jun 06 09:09:18 hooked up laptop to the serial port of my apu2 and enabled flow_offloading again Jun 06 09:27:45 hmmm, apu2 rebooted but no crashlog in the serial console :/ Jun 06 10:05:12 second time, no crashlog :( Jun 06 10:05:39 disabling flow offloading againg for now Jun 06 10:12:00 Is it possible to separate make compilation outputs to dedicated package log files (for example openssl-compile.log, openvpn-compile.log) ? Jun 06 10:14:09 build #483 of at91/sama5d3 is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/at91%2Fsama5d3/builds/483 Jun 06 10:19:01 edgasm1: make ... BUILD_LOG=1 Jun 06 10:19:12 output will end up in the logs/ directory Jun 06 10:23:37 jow: Thank You very much Jun 06 10:33:07 ooh has github gone awol? Jun 06 10:33:28 * ldir thinks it's probably installing an update Jun 06 10:34:36 MS bought it, expect to have to buy the upgrade... Jun 06 10:35:42 Have You considered using GitLab for hosting OpenWRT project? Jun 06 10:43:31 no Jun 06 10:44:23 we host our own git, github is just a mirror Jun 06 10:52:51 schm0: pong Jun 06 10:57:52 and everyone who has a clone of the repos and keeps them up to date has a backup copy too. Jun 06 11:18:40 ping Hauke Jun 06 11:18:51 Homehub 5a - once I've flashed openwrt onto it, is there any need to keep the boot_sel wires hanging around or can I unsolder them? I don't ever anticipate going back to stock firmware. Jun 06 11:20:05 ldir: shouldn't be required as long as you don't mess with the u.boot Jun 06 11:21:43 mkresin: so if I destroy u.boot (unlikely) then bootsel is a way of transfering a good uboot over serial...and then on to full recovery? Jun 06 11:27:43 jow: for ath79, i've got a bunch of 'old' atheros devboards, i'll try to add the dtses for that Jun 06 11:30:56 ldir: correct Jun 06 11:31:02 wigyori: that sounds great :) need to look into my hardware zoo as well Jun 06 11:31:29 mkresin: how 'low level' is that bootsel loader then? Where does it live? Jun 06 11:32:49 ldir: as low as it is possible. it is in the SoCs bootrom. the first code that runs after pwr on Jun 06 11:33:40 mkresin: oh wow - so if that doesn't come up...the device really is a brick! Thank you for this fantastic info. Jun 06 11:45:15 jow: zoo is the perfect word :) Jun 06 12:52:16 jow:ping Jun 06 12:53:57 dedeckeh: pong Jun 06 12:54:48 jow:I had a look at PR https://github.com/openwrt/openwrt/pull/805 Jun 06 12:55:20 jow:after the latest coments of guidosarducci I would be in favor of merging this Jun 06 12:55:39 jow:what's your opinion ? Jun 06 12:56:23 no objections, just merge it Jun 06 12:56:39 but please cherry pick into lede-17.01 and openwrt-18.06 if possible instead of merging the separate PRs Jun 06 12:57:53 jow: will do Jun 06 13:48:59 Hi. I'm using LuCI lede-17.01 on a Netgear X6. I have installed all (I think) the usb and samba packages and have an NTFS HDD plugged into the USB3 socket. I have successfully mounted the drive in a folder but it is empty and no files are showing. I have run an error scan on a PC and it's in good health. Can anyone assist please? Jun 06 13:53:44 schlimig: if you will be using this HDD solely for the router, then perhaps formatting it as ext4 is better Jun 06 13:53:49 did you install ntfs-3g ? Jun 06 13:55:50 I cannot format as ext4 because it will occasionally (particularly if router fails) need to be connected to Windows machine Jun 06 13:55:59 And yes ntfs-3g is installed Jun 06 13:56:40 df -h Jun 06 13:56:51 do you see your drive and mountpoint listed there? Jun 06 13:56:57 yes Jun 06 13:57:12 It's even picked up the name of the drive in the mount-points page of LUCI Jun 06 13:57:21 so it's clearly reading the drive on some level Jun 06 13:57:42 ls -lah /your/mount/point shows absolutely nothing? Jun 06 13:58:15 empty folder Jun 06 13:58:17 correct Jun 06 13:58:30 how about partitioning? Jun 06 13:58:42 How do you mean? It's one single partition Jun 06 13:59:50 do you see the correct size when you issue df -h ? Jun 06 14:00:15 yes Jun 06 14:00:27 absolutely everything checks out, just no files Jun 06 14:00:36 hence my stumpedness :) Jun 06 14:01:12 can you try mount it manually instead of using luci ? Jun 06 14:01:25 see if it throws any errors Jun 06 14:01:51 umount /dev/ Jun 06 14:01:53 I also did this and no errors, it showed up in the mounted drives list of LUCI correctly Jun 06 14:02:13 ntfs-3g /dev/yourdrive /mnt/mountpoint -o rw,sync Jun 06 14:02:22 yes I did exactly that Jun 06 14:02:34 fun stuff Jun 06 14:02:58 if you cd into /mnt/mountpoint/ and run "echo *" (without quotes) is it also empty? Jun 06 14:03:27 checking Jun 06 14:04:35 I just get the asterisk Jun 06 14:04:39 ok Jun 06 14:05:46 do you see any ntfs related errors in dmesg ? Jun 06 14:06:33 [88300.777285] blk_update_request: I/O error, dev mtdblock14, sector 161536 [88300.784474] blk_update_request: I/O error, dev mtdblock14, sector 161536 [88300.791194] Buffer I/O error on dev mtdblock14, logical block 20192, async page read Jun 06 14:06:42 I'm seeing several of these lines Jun 06 14:07:00 that does not look healthy but is unrelated to your ntfs issue I suppose Jun 06 14:07:10 mtdblock* is internal flash memory Jun 06 14:07:47 Wouldn't surprise me with this overpriced hulk! Jun 06 14:08:36 this is werid indeed, one thing worth trying is mounting the filesystem with the kernel ntfs driver Jun 06 14:08:43 insteado f ntfs-3g Jun 06 14:09:10 is that just ntfs instead? Jun 06 14:09:57 it would be kmod-fs-ntfs iirc and then mount -t ntfs /dev/drive /mnt/foo/ Jun 06 14:10:04 instead of ntfs-3g Jun 06 14:10:04 ah great, thanks - trying that Jun 06 14:10:32 no it's not that Jun 06 14:12:07 do you have a usb stick? if yes try mount that and see if you detect the files Jun 06 14:12:23 or if you have a powered USB hub, try using it instead of attaching the drive to the router directly Jun 06 14:12:39 Hm, would a Mint Live USB be suitable? Jun 06 14:12:54 anything with files on it Jun 06 14:12:56 Not NTFS obviously Jun 06 14:12:59 something with ntfs preferably Jun 06 14:13:08 I haven't got any alas Jun 06 14:13:08 right, to rule out issues with ntfs-3g itself Jun 06 14:13:20 ANd no spare sticks right now Jun 06 14:18:17 Hmm... I mounted a vfat stick successfully Jun 06 14:18:49 or I thought I did Jun 06 14:19:36 OK yes, it's mounted properly, showing correct size with df -h etc. Jun 06 14:19:38 But still no files Jun 06 14:20:22 Is there anything wrong with this command: Jun 06 14:20:23 mount -t vfat /dev/sda1 /mnt/usb -o rw,sync Jun 06 14:20:56 schlimig: no obvious error in it. Jun 06 14:21:08 Very puzzling Jun 06 14:23:02 schlimig: look into dmesg for messages regarding your mount Jun 06 14:23:25 schlimig: maybe there is an error message or warning available Jun 06 14:23:38 Nothing alas Jun 06 14:23:42 I see the usual: Jun 06 14:23:43 [89071.444777] usb-storage 2-1:1.0: USB Mass Storage device detected [89071.451216] scsi host2: usb-storage 2-1:1.0 [89072.471114] scsi 2:0:0:0: Direct-Access Kingston DataTraveler 2.0 PMAP PQ: 0 ANSI: 4 [89074.463446] sd 2:0:0:0: [sda] 122915328 512-byte logical blocks: (62.9 GB/58.6 GiB) [89074.471807] sd 2:0:0:0: [sda] Write Protect is off [89074.476620] sd 2:0:0:0: [sda] Mode Sense: 23 00 00 00 [89074.477299] sd 2:0: Jun 06 14:23:53 Which suggests everything is fine Jun 06 14:25:51 maybe a filesystem size limitation. Does it mount on a normal linux machine? Jun 06 14:26:15 Yes, both the USB stick I'm trying and the HDD do Jun 06 14:27:18 You could try to mount it without telling about filesystem type Jun 06 14:27:25 mount /dev/sda1 /mnt/usb Jun 06 14:27:41 Then again looking into kernel log Jun 06 14:28:27 The lines from dmesg you posted only contains the detection of the USB stick. Jun 06 14:28:38 I get a mount failure if I do that Jun 06 14:29:00 Check with lsmod for some kernel module with vfat in it Jun 06 14:30:02 fat 41632 2 msdos,vfat nls_base 4804 9 vfat,fat,ntfs,cifs,nls_utf8,nls_iso8859_1,nls_cp437,exfat,usbcore vfat 7296 0 Jun 06 14:31:09 Hmm, seems it's not recognising any vfat at /dev/sda1 Jun 06 14:31:33 When you mount the stick on a normal linux machine, what device is mounted? Jun 06 14:31:55 Maybe you could try mounting /dev/sda Jun 06 14:33:00 I don't recall on my linux machine (and it's out of action at the moment) Jun 06 14:33:04 I'll try sda Jun 06 14:33:35 nope, invalid argument Jun 06 14:34:49 strange... Jun 06 14:42:25 I have to shoot so I'll have to grapple with this some more later. @smay @jow @SwedeMike @abenz thank you all for your help, much appreciated :) Jun 06 14:42:33 Bye# Jun 06 16:15:23 hi Jun 06 16:24:52 * ldir collapses in a heap after a tussle with his MIR3G - but has beaten it into submission. Jun 06 16:30:18 * Borromini applauds Jun 06 16:30:43 was it resisting your gentle touch while being forcibly flashed? Jun 06 16:36:02 mm flashing Jun 06 16:36:04 i've got an ubnt-erx that is losing its port 0 link for 3 seconds, 9 times in the last 12 hours Jun 06 16:36:42 OpenWrt SNAPSHOT, r7104-987900f2de Jun 06 16:37:01 speaking of which, does it have dsa yet? :D Jun 06 16:37:12 need nice .1q Jun 06 16:37:16 :P Jun 06 17:01:03 Borromini: no it was more a dts understanding thing. It did get flashed...several times. Jun 06 17:07:15 nbd: the longer I look at https://git.openwrt.org/?p=project/netifd.git;a=commitdiff;h=509ffb22475ebdd5291d510a098f996473951344 the more it seems to me as if the default values got swapped Jun 06 17:07:34 jow. long time Jun 06 17:07:44 ldir: ok Jun 06 17:08:51 nbd: at least judging from the diffs before and after that commit it seems to me like multicast_to_unicast was silently changed to default-enabled with this commit Jun 06 17:09:14 while the commit message claims to fix something else Jun 06 17:13:23 hello people Jun 06 17:13:44 any body can help my Jun 06 17:14:53 nbd: also I cannot see where it is getting default-enabled elsewhere to actually invoke system_bridge_set_multicast_to_unicast() in netifd Jun 06 17:16:26 I wanna to install openwrt in my litlebeam M5 Jun 06 17:16:43 but I canĀ“t Jun 06 17:47:36 djowpong Jun 06 17:47:39 jow: pong Jun 06 17:49:12 Hauke: it was about that recent mtd cherry-pick in lede-17.01 that caused older IBs to stop functioning Jun 06 17:49:18 Hauke: I added a hack now to address this issue Jun 06 17:49:39 jow: thanks Jun 06 17:49:52 I hoped I would improve the situation ;-) Jun 06 17:49:55 marking a previously shared package nonshared in the release branch will cause it to vanish from the package repo Jun 06 17:50:05 which is okay and intended Jun 06 17:50:09 ok I didn't know Jun 06 17:50:22 but older IBs will not function anymore (if this package happens to be one in the default package set) Jun 06 17:50:47 since they neither have it in their local verison tagged nonshared repo, nor find it in the global shared anymore Jun 06 17:51:18 anyhow... just a heads-up, I don't anticipate further bumps to mtd in 17.01 anymore Jun 06 17:51:31 but we should be careful with moving shared packages to nonshared between minor releases Jun 06 17:52:06 the other way around is actually no problem (given that PKG_RELEASE and/or version is increased) Jun 06 17:52:20 ok Jun 06 17:52:25 thanks for the information Jun 06 17:59:33 jow: thank you for always cleanung up behind me ;-) Jun 06 18:09:53 clear Jun 06 18:09:58 clean Jun 06 19:04:14 hello folks, i installed OpenWRT to a asus WL-500gP V1 router. Problem is that when i connect the LAN interface to PC and supposed WAN interface to ethernet wall mount the PC gets the ISP's IP directly. Is there a config that enables the router mode or equilevant? Jun 06 19:07:54 https://www.onetransistor.eu/2017/04/wan-port-openwrt-lede-vlan.html Jun 06 19:07:59 i have answered my own question Jun 06 19:08:04 but now the router is a brick Jun 06 19:08:07 >___< Jun 06 19:27:14 build #11 of omap/generic is complete: Failure [failed sourceupload] Build details are at http://release-builds.openwrt.org/18.06/images/builders/omap%2Fgeneric/builds/11 blamelist: Tony Ambardar Jun 06 19:48:51 build #10 of mpc85xx/generic is complete: Failure [failed sourceupload] Build details are at http://release-builds.openwrt.org/18.06/images/builders/mpc85xx%2Fgeneric/builds/10 blamelist: Tony Ambardar Jun 06 21:04:03 nbd: i just figured that ramips/patches-4.14/0053-mtd-spi-nor-add-w25q256-3b-mode-switch.patch breaks reboots on WrtNode2* boards Jun 06 21:07:04 nbd: i reckon we should read the flash chip's Non-Volatile Status Register Bit ADP(S17) setting and then make sure it's in that mode when rebooted Jun 06 21:40:51 dangole: there's a variant of said patch in linux-next. Maybe use that? Jun 06 21:44:09 https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?h=next-20180606&id=f134fbbb4ff813dd227c9ce40b5c0b2078a77b07 Jun 06 21:58:32 magix: so EX4B corrupts the EAR. we never did EX4B before that patch Felix made. Jun 06 21:59:33 mangix: and apparently the WrtNode2 romloader doesn't mind the chip being in 4B mode, but it does mind those left-overs in EAR Jun 06 22:02:03 mangix: i'll try the patch you mentioned on top and see if it fixes the issue Jun 06 22:18:15 is libluci-lua mandatory? im trying to figure it why sometimes luci doesnt properly save settings Jun 06 22:43:41 how do i turn some C source code into a .ipk? Jun 06 23:14:49 biangbiangmian: https://wiki.openwrt.org/doc/devel/packages Jun 06 23:32:09 abenz: btw. all my sysupgrades of recent firmwares succeeded for me (ar71xx, ipq8065 and lantiq), not quite from the exact same time frame, but several updated twice over the last ~week Jun 06 23:36:48 bah, screw you t mobile de Jun 06 23:37:01 their ridiculous dns setup is breaking my tunnel Jun 06 23:53:53 pkgadd: mine only failed on mt7621 (checking with serial shows wifi/hostapd wont go down so upgrade aborts) Jun 07 00:44:29 what should i be looking at if i'm not able to detect a SIM slot with modemmanager? Jun 07 00:57:39 pkgadd: ping Jun 07 00:58:10 hmm Jun 07 00:59:08 when i use the subnet option for zone config, why does it break dhcp? Jun 07 00:59:25 there is also nothing in the logs Jun 07 01:01:32 but clients that are already have their lease, can access the network just fine Jun 07 01:03:35 https://ibb.co/kOHUq8 Jun 07 01:04:40 the top one is default and works fine. the bottom one is modified one. the purpose is to prevent leakage between the two networks on input/output Jun 07 01:05:53 i think because the client as no ip at this stage yet. how to drive around that? Jun 07 01:10:43 there must be way to prevent services that are bound to br-lan to leak to the isolated network and vice versa without blocking input/output on both networks Jun 07 01:28:13 mangix: ping (although I'm almost on the way out) Jun 07 01:30:32 pkgadd: my ath10k patch broke on my r7800. Jun 07 01:30:40 so that's that Jun 07 01:31:23 interesting, it's fine on my nbg6817 Jun 07 01:31:46 oh you have that, right Jun 07 01:32:37 SOC and wlan side of it are identical, so that's weird Jun 07 01:32:57 archer c7 fails as well Jun 07 01:36:36 mangix: fixed that winbond flash issue... it was related to the patch you mentioned, but the solution was to indeed read the configuration and decide upon that whether to use 4B mode or the 3B mode with 4B READ instruction... Jun 07 01:38:24 commit fec205f654 ("ramips: mmc: Sync with staging driver") breaks microSD on MT7688AN (eg. WrtNode2 boards) Jun 07 01:39:13 dangole: blogic reverted the fix for that Jun 07 01:39:45 https://github.com/openwrt/openwrt/commit/8a570921b5ba49a2d3824f1220e4c53809063468 Jun 07 01:39:56 i think that was accidental Jun 07 01:44:28 something seems bugged to me... when i have zone_*_input and zone_*_output that are configured to specific interfaces. And i bind a service to one of the interfaces how is this even possible that the traffic can reach the other zone Jun 07 01:46:21 i dont want to create a bunch of firewall rules. imagine someone has like 10 zones that is just buggos Jun 07 01:48:22 *bogus Jun 07 01:54:00 for example miniupnpd i configure it to br-lan interface but the ssdp broadcast is send to all zones. another example luci interface bound to br-lan host of the isolated zone can still access it. so i have to nlock input on all zones and create bunch of firewall rules? Jun 07 01:54:50 or did i miss something in switch config? Jun 07 02:08:15 hello all Jun 07 02:09:12 can anyone help me to get libmraa working with lede? Jun 07 02:22:09 is anyone using libmraa? Jun 07 02:26:37 if no one uses libmraa, what do you use for GPIO control? Jun 07 02:28:20 mangix: jackpot. reverting john's previous revert-commit fixes the issue and microSD works well with the upstream driver Jun 07 02:29:35 @blogic: commit 8a570921b Revert "ramips: Add back some non-mt7621 code that staging removed" breaks stuff on MT7688AN Jun 07 02:35:11 build #522 of ramips/mt76x8 is complete: Failure [failed sourceupload] Build details are at http://phase1.builds.lede-project.org/builders/ramips%2Fmt76x8/builds/522 blamelist: Tony Ambardar , Matthias Schiffer , Daniel Golle , Hans Dedecker Jun 07 02:52:12 any linkit7688 users here? Jun 07 02:54:57 andrewmuck: i got a duo Jun 07 02:55:10 but i will be no help, sorry. never messed with it beyond realizing the default fw sucked. Jun 07 02:55:29 i tried lede some time ago and the wireless would receive but not send (or vice versa, i forget) Jun 07 02:57:16 * m4t is curious if it works now **** ENDING LOGGING AT Thu Jun 07 03:00:04 2018