**** BEGIN LOGGING AT Fri Nov 30 03:00:00 2018 Nov 30 03:00:05 build #168 of ixp4xx/generic is complete: Success [build successful] Build details are at http://release-builds.openwrt.org/18.06/images/builders/ixp4xx%2Fgeneric/builds/168 Nov 30 03:29:52 build #1097 of ramips/rt3883 is complete: Failure [failed targetupload] Build details are at http://phase1.builds.lede-project.org/builders/ramips%2Frt3883/builds/1097 blamelist: Hans Dedecker Nov 30 03:31:28 build #1111 of ar71xx/mikrotik is complete: Failure [failed targetupload] Build details are at http://phase1.builds.lede-project.org/builders/ar71xx%2Fmikrotik/builds/1111 blamelist: Hans Dedecker Nov 30 03:37:02 build #155 of gemini/generic is complete: Success [build successful] Build details are at http://release-builds.openwrt.org/18.06/images/builders/gemini%2Fgeneric/builds/155 Nov 30 03:42:27 build #154 of ar71xx/nand is complete: Success [build successful] Build details are at http://release-builds.openwrt.org/18.06/images/builders/ar71xx%2Fnand/builds/154 Nov 30 04:01:59 build #154 of ipq806x/generic is complete: Success [build successful] Build details are at http://release-builds.openwrt.org/18.06/images/builders/ipq806x%2Fgeneric/builds/154 Nov 30 04:14:34 christ. too much bot chatter. Nov 30 04:16:28 build #251 of mediatek/mt7622 is complete: Failure [failed kmodupload] Build details are at http://phase1.builds.lede-project.org/builders/mediatek%2Fmt7622/builds/251 blamelist: Hans Dedecker Nov 30 04:16:55 build #154 of ramips/mt76x8 is complete: Success [build successful] Build details are at http://release-builds.openwrt.org/18.06/images/builders/ramips%2Fmt76x8/builds/154 Nov 30 04:28:24 lmao Nov 30 04:39:32 build #153 of lantiq/xway is complete: Success [build successful] Build details are at http://release-builds.openwrt.org/18.06/images/builders/lantiq%2Fxway/builds/153 Nov 30 04:42:33 build #1085 of ixp4xx/harddisk is complete: Failure [failed kmodupload] Build details are at http://phase1.builds.lede-project.org/builders/ixp4xx%2Fharddisk/builds/1085 blamelist: Hans Dedecker Nov 30 04:47:20 build #152 of brcm47xx/legacy is complete: Success [build successful] Build details are at http://release-builds.openwrt.org/18.06/images/builders/brcm47xx%2Flegacy/builds/152 Nov 30 05:04:50 blogic: do you think its possible (with dsa drivers) to use bonding (lacp/802.3ad) on two ports of the same switch? Nov 30 05:05:09 I tried looking at qca8337 datasheet but no mention of it Nov 30 05:05:56 but another router (netgear R9000) advertises 802.3ad support, but that router has two qca8337 switches Nov 30 05:20:48 blogic: I recall even with dsa, only a single cpu port is used, this would be a bottleneck even if 802.3ad worked, no? Nov 30 06:08:53 build #321 of cns3xxx/generic is complete: Success [build successful] Build details are at http://release-builds.lede-project.org/17.01/images/builders/cns3xxx%2Fgeneric/builds/321 Nov 30 06:16:13 nbd: some kind of miracle occurred in mt76, suddenly linkit can connect to all my ap's and has (with one exception where the issue is probably weak signal) zero packet loss. Nov 30 06:16:34 same version as where i saw the kernel panic Nov 30 06:17:00 second test i wrote to /root instead of /tmp so the statistics survived Nov 30 06:19:17 build #318 of kirkwood/generic is complete: Success [build successful] Build details are at http://release-builds.lede-project.org/17.01/images/builders/kirkwood%2Fgeneric/builds/318 Nov 30 06:28:39 build #318 of ixp4xx/generic is complete: Success [build successful] Build details are at http://release-builds.lede-project.org/17.01/images/builders/ixp4xx%2Fgeneric/builds/318 Nov 30 06:40:03 build #335 of x86/geode is complete: Success [build successful] Build details are at http://release-builds.lede-project.org/17.01/images/builders/x86%2Fgeode/builds/335 Nov 30 06:50:12 build #365 of gemini/raidsonic is complete: Success [build successful] Build details are at http://release-builds.lede-project.org/17.01/images/builders/gemini%2Fraidsonic/builds/365 Nov 30 07:01:11 build #362 of ramips/rt288x is complete: Success [build successful] Build details are at http://release-builds.lede-project.org/17.01/images/builders/ramips%2Frt288x/builds/362 Nov 30 07:07:21 build #321 of ipq40xx/generic is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/ipq40xx%2Fgeneric/builds/321 Nov 30 07:11:02 build #358 of ath25/generic is complete: Success [build successful] Build details are at http://release-builds.lede-project.org/17.01/images/builders/ath25%2Fgeneric/builds/358 Nov 30 07:23:36 scenario: I want to gain access to some "shared pins" on a mediatek mt7621. so in my DTSI file, I've gone in and added the register, in my case RGMII2, into the pinctrl0 group and specified that the function is "gpio". is that how to the pins controlled by that register should be treated as GPIO pins? Nov 30 07:24:13 https://pastebin.com/geEp9HUz Nov 30 07:24:55 problem is that i don't understand why this works. i can now control the pins (they are LED's) but did I do it correctly? Nov 30 08:23:07 karlp: we perhaps don't know all the use cases of the libubox, but one of mine use cases is development/prototyping of ubus (and related openwrt solutions) on the host (not cross-compiling) Nov 30 08:26:28 but you're correct it needs some love, and we perhaps should consider converting it to some CMake module or maybe extend upstream Lua51 CMake module so it could provide Lua's cpaths as some kind of list or something like that Nov 30 08:29:24 for me it's not a big deal to provide the proper Lua cpath in LUAPATH CMake's variable, so it works for me somehow for all my use cases and there are always more important things to work on right now :) Nov 30 09:25:48 Hauke: how do you think, why did I add a line like this? ;) Nov 30 09:25:50 BUILD_BUG_ON(LINUX_VERSION_CODE >= KERNEL_VERSION(4, 15, 0)); Nov 30 09:26:09 hint: it's right below pr_warn("Support for the generic \"linux,part-probe\" has been deprecated and will be removed soon"); Nov 30 09:26:30 that was to make sure noone will ever use that code on newer kernels Nov 30 09:27:04 but instead of dropping a patch you dopped the line... Nov 30 09:27:24 Hauke: please completely drop that patch, I absolutely don't want to see that in new kernels Nov 30 09:27:34 161-mtd-part-add-generic-parsing-of-linux-part-probe.patch Nov 30 09:57:14 blogic: hi, I slowly start to understand fstools / block :) Nov 30 09:58:05 there is some headache code, like e.g. switch (type) in the mount_device() where you need to carefully analyze all the cases Nov 30 09:58:08 but it starts making sense Nov 30 10:27:14 biangbiangmian: https://www.kresin.me/files/pinmux_review_ramips.ods might give you a clue Nov 30 10:29:14 build #450 of ar71xx/tiny is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/ar71xx%2Ftiny/builds/450 Nov 30 10:30:19 Hi jow, the external toolchain is working great. Really cool. I've run into a minor issue when building dnsmasq though and would like your advice. Nov 30 10:32:18 Dnsmasq needs netlink conntrack kernel module apparently (bit surprising). As you suggested I just set the target to arm realview, so the wrong kernel is used Nov 30 10:33:56 What's the best way to handle this? Should I try make a target with this old kernel or should I just patch dnsmasq and to use library that's built manually? Nov 30 10:34:49 I've never added a target. Is it a large amount of work? I'd only need it for building kmors Nov 30 10:34:58 Kmods* Nov 30 10:41:19 AndroUser2: AFAIK dnsmasq builds with conntrack dep if built with HAVE_CONNTRACK - ie. you don't have to build it with conntrack support. Nov 30 10:43:05 Interesting, what does it use conntrack for anyway? Regardless, I'd like to solve this kmod issue as I'm sure some future package I require will depend on one. Nov 30 10:53:19 mkresin: Hi, do you know what are the possible values for `/sys/kernel/debug/ieee80211/phy0/mt76/led_pin` ? It seems like my 5GHz LED is connected to either LED_A or LED_B pin, I've tried some numbers already but none of them works Nov 30 10:53:53 could find datasheet with description of that LED S0/S1 registers Nov 30 10:54:04 s/could/couldn't/ Nov 30 10:54:38 AndroUser2: http://www.thekelleys.org.uk/dnsmasq/docs/dnsmasq-man.html search for conntrack Nov 30 10:55:45 A Nov 30 10:56:11 ldir: thanks! Pretty cool feature Nov 30 11:04:36 ynezz: same problem with the datasheet, hence no idea Nov 30 11:04:46 ynezz: have you tried to change /sys/kernel/debug/ieee80211/phy0/mt76/led_active_low as well? Nov 30 11:05:03 yes indeed Nov 30 11:16:51 thanks mkresin! saved Nov 30 11:38:12 jow: just wanted to notify that I haven't forgot you regarding testing the patch :) it's just been really busy here Nov 30 11:39:58 build #5 of ath79/generic is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/ath79%2Fgeneric/builds/5 Nov 30 12:05:57 has anyone used hotplug scripts to trigger LED's based on the return value of a command? that's a pretty reasonable thing to do isn't it? Nov 30 12:10:57 build #1232 of x86/64 is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/x86%2F64/builds/1232 Nov 30 12:45:44 build #289 of mediatek/mt7623 is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/mediatek%2Fmt7623/builds/289 Nov 30 13:10:44 build #1129 of apm821xx/sata is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/apm821xx%2Fsata/builds/1129 Nov 30 13:22:06 build #1056 of ar71xx/generic is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/ar71xx%2Fgeneric/builds/1056 Nov 30 13:27:26 build #997 of malta/be is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/malta%2Fbe/builds/997 Nov 30 13:27:42 build #1128 of at91/legacy is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/at91%2Flegacy/builds/1128 Nov 30 13:31:06 build #1130 of brcm47xx/legacy is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/brcm47xx%2Flegacy/builds/1130 Nov 30 14:17:52 build #1120 of mpc85xx/generic is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/mpc85xx%2Fgeneric/builds/1120 Nov 30 14:18:39 build #1142 of mxs/generic is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/mxs%2Fgeneric/builds/1142 Nov 30 14:22:20 build #1128 of omap/generic is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/omap%2Fgeneric/builds/1128 Nov 30 15:09:04 blogic: i'll need your help with fstools / block Nov 30 15:09:17 blogic: my problem is I can't generate mount "remove" event for hotplug.d Nov 30 15:09:37 blogic: with mountd it was simple because it kept internal state of all mount points Nov 30 15:09:59 it contained info of every block device that it mounted Nov 30 15:10:23 so when device was disappearing, it knew if it was mounted or not and could generate a "remove" event Nov 30 15:11:20 with fstools / block we get "ACTION=remove DEVNAME=sda1 /sbin/block hotplug" called Nov 30 15:11:55 but block does not know if /dev/sda1 was previously mounted or not... so it does not know if it should generate "remove" mount event for hotplug.d scripts Nov 30 15:12:03 blogic: does my explanation make sense to you> Nov 30 15:12:39 rmilecki: I suppose the event is triggered after umount? So no chance to check /proc/mounts ? Nov 30 15:12:56 right Nov 30 15:13:21 block does: mount_point = find_mount_point(path); Nov 30 15:13:24 but it returns null Nov 30 15:13:40 then I fear you'll need to implement some mount accounting Nov 30 15:14:10 i'm ok with that, i'll just wait to see if blogic is ok with that too then Nov 30 15:14:30 a simple list with device names should do, so that won't be complex Nov 30 15:15:01 yeah, stored in /var/run or /var/state Nov 30 15:15:27 oh, i thought of internal list in C in block Nov 30 15:15:53 but block isn't a running daemon? Nov 30 15:15:58 so it doesn't keep state Nov 30 15:16:03 actually... just plugging USB device out shouldn't make entry in /proc/mounts disappear, right? Nov 30 15:16:15 not sure Nov 30 15:16:16 ah, block is not a deamon... i forgot Nov 30 15:16:23 build #1098 of ramips/rt3883 is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/ramips%2Frt3883/builds/1098 Nov 30 15:16:38 so you have to write that info out somewhere Nov 30 15:17:03 maybe I can query blockd... it keeps a list of mounts Nov 30 15:19:37 build #1112 of ar71xx/mikrotik is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/ar71xx%2Fmikrotik/builds/1112 Nov 30 15:47:45 oh, i realized correctly that mount point can't just disappear on its own Nov 30 15:47:53 it's blockd calling umount Nov 30 15:48:25 so in mount_action() block notifies blockd about disappeared block device, then blockd unmounts it, then block looks for a mount point Nov 30 15:49:52 build #252 of mediatek/mt7622 is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/mediatek%2Fmt7622/builds/252 Nov 30 15:52:59 ok, that's the problem... blockd unmounts device in its device_free() Nov 30 15:53:08 so we have both: block and blockd trying to unmount device Nov 30 15:53:13 that's fishy I suppose Nov 30 15:53:33 and doesn't allow writing reliable code for sending "remove" event Nov 30 16:03:22 hmm this comparison seems to give me issues with mwlwifi.. https://github.com/torvalds/linux/blob/master/net/mac80211/ethtool.c#L166 Nov 30 16:03:22 is there a better way to compare two ieee80211_channel structs? Nov 30 16:11:25 build #1086 of ixp4xx/harddisk is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/ixp4xx%2Fharddisk/builds/1086 Nov 30 18:53:55 rmilecki: I will add an extra patch to drop it Nov 30 20:22:43 build #940 of armvirt/64 is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/armvirt%2F64/builds/940 Nov 30 22:27:29 Hauke: thanks Nov 30 23:43:28 nbd: wishlist for mt76 is rssi information Nov 30 23:54:40 rmilecki: what triggers the "remove" event anyway? is it thing slike CD pin on sd slots or what? isn't it too late to do much then anyway? Nov 30 23:55:02 youcan "cleanup after unclean unmounts", but that's it right? Dec 01 00:05:45 so, I had OpenWRT failing to compile the host tools because gmp was incorrectly detecting x32 and then causing mpfr to fail to build Dec 01 00:07:05 I "fixed" it by deleting the ifeq ($(GNU_HOST_NAME),x86_64-linux-gnux32) Dec 01 00:07:12 from gmp's Makefile Dec 01 00:07:22 but I don't understand why it was detecting that in the first place **** ENDING LOGGING AT Sat Dec 01 03:00:00 2018