**** BEGIN LOGGING AT Sat Oct 20 03:00:00 2018 Oct 20 06:57:53 Hauke: no Oct 20 07:18:35 build #118 of bcm53xx/generic is complete: Failure [failed kmods] Build details are at http://release-builds.openwrt.org/18.06/images/builders/bcm53xx%2Fgeneric/builds/118 blamelist: Koen Vandeputte Oct 20 08:38:02 * ldir wishes we could rename KGB to GRU or even more freaky.... VLAD :-) Oct 20 08:41:30 or have them spit out stats that are sane? Oct 20 08:41:42 0 and % woudlimply a broken script to me Oct 20 08:43:47 ah now was it you karlp that had some thoughts on the optimisation overrides that were being requested ? Oct 20 08:45:55 what? the lto stuff? Oct 20 08:46:15 I just said, "shit's gonna break, but hey, we might make gcc better by hurting our people in the short term" Oct 20 08:47:13 no, this https://github.com/openwrt/openwrt/pull/858 Oct 20 09:02:43 in Makefiles, what is the practical difference between VARIABLE=value vs VARIABLE:=value ? Oct 20 09:05:57 pkgadd: hi! do you still have issues with wds? Oct 20 09:07:21 ldir: time of evaluation. only relevant if you reference any other variables. Whe you use FOO=$(BAR), you say whenever you use FOO, replace it with the value of BAR (which may be set later) When you use FOO:=$(BAR) then you say FOO should have the value of BAR it currently has (so needs to be defined before) Oct 20 09:17:11 aha, right the former my be recursively expanded whilst the latter is simply expanded. Oct 20 09:18:32 if you do git grep PKG_RELEASE you'll see a couple of discrepancies....and I was curious / trying to avoid cargo culting :-) Oct 20 10:45:28 greearb_: so it's issue #30. I can help test binaries to debug on the 9984 Oct 20 11:15:05 ldir: renaming to it GRU would just make it incompetent and very eager to visit Salisbury Oct 20 11:44:14 ldir: nope, no comment on that sort of thing. Oct 20 12:00:36 is there chance that next realse is near? i'm about to install 18.06.1 and _would_ wait if you say it'll be superseded with next release within a month Oct 20 12:12:46 not a month Oct 20 12:12:59 maybe 3 months at least Oct 20 12:13:53 but if the nand problem gets solved, there may be an 18.06.2 sooner Oct 20 12:17:55 what's with nand? Oct 20 12:20:08 thanks DonkeyHotei , will flash current-stable then Oct 20 13:29:06 i flashed the 18.06.1, on the first contact it tells about non-set root-pw and so on and presents a button to go to "root pw settings" but that one is useless as it doesn't work. login with empty root-pw however lands me on the setting page anyway Oct 20 13:31:29 Piraty (IRC): Reliability problems with kernel 4.14 Oct 20 13:31:39 Piraty (IRC): for details, read http://lists.infradead.org/pipermail/openwrt-devel/2018-October/014355.html Oct 20 13:34:12 ah too bad. thanks luaraneda Oct 20 13:34:27 why this (IRC) suffix? what client you using? Oct 20 14:05:31 he's using an irc <--> matrix gateway Oct 20 14:15:33 * ldir tries to not bend the spoon Oct 20 14:17:22 "there is no spoon" Oct 20 14:46:25 Pubkey auth attempt with unknown algo, so openwrt is unable to ed25519 ? Oct 20 14:47:42 i know that dsa doesn't work Oct 20 14:49:37 dsa is fine, but why no ed2 Oct 20 14:50:40 *rsa Oct 20 15:17:21 Piraty: because it is not supported by dropbear yetr Oct 20 15:17:30 alright Oct 20 16:16:09 Why is "Boot Loaders" section in menuconfig is empty? Oct 20 16:16:21 Why the* Oct 20 16:28:54 huaracheguarache: WDS is not working for me (nbg6817/ QCA9984) with the 4.18 or 4.19 based backports with ath10k, ath10k-ct v4.13 or ath10k-ct 4.16, while ath9k is working - I still have to revert back to the 4.14 based mac80211/ backports package Oct 20 16:38:36 pkgadd: does wds work for you with 4.14 backports + ath10k-ct 4.13? Oct 20 16:40:17 Hauke: that's pretty much the only combination I haven't tested so far, but I can start a build with that configuration now Oct 20 16:56:36 pkgadd: Felix did some changes to mac80211 regrading 4-addr mode here: git.kernel.org/linus/9251a4736ee360d3850644ddaaa1a61dcec96238 Oct 20 16:56:46 and the 2 previous commits Oct 20 16:57:50 I think I may have tried reverting those already, but it's been a while since I tested those (can try it again though) Oct 20 16:58:51 pkgadd: probably nbd can help you better than me Oct 20 17:02:03 I know that 4addr is always a bit fragile in terms of bigger mac80211/ backports upgrades Oct 20 17:04:14 unfortunately I have no x86 system with ath10k around (but plenty ath9k/ ath5k), that doesn't make bisecting any easier Oct 20 17:24:36 Hi just flashed my c7v2 to the snapshot on the download page and it is still on r8281 and k4.9 Oct 20 17:24:52 I was told that r71xx is now on k4.14 Oct 20 17:27:06 current HEAD would be 8369 Oct 20 17:28:05 So what's going on here? The file looke like it flashed rite Oct 20 17:28:24 I flashed through the webinterface Oct 20 17:28:35 should i try over ssh? Oct 20 17:31:50 Tapper: you mean ar71xx Oct 20 17:31:59 yes r71xx Oct 20 17:32:12 your a key is stuck Oct 20 17:33:02 What russell-- Oct 20 17:33:05 rFOO looks like a revision number Oct 20 17:33:32 the arch name is "ar71xx" not "r71xx" Oct 20 17:33:52 O crap sorry Oct 20 17:34:05 ar71xx got it lol. Oct 20 17:34:08 Hauke: current HEAD with old mac80211 (wt-2017-11-01-0-gfe248fc2c180/ v4.14-rc2-1-31-g86cf0e5d) and the corresponding ath10k-ct 4.13 is working fine (http://paste.debian.net/1048278/), as is current HEAD with old mac80211 (wt-2017-11-01-0-gfe248fc2c180/ v4.14-rc2-1-31-g86cf0e5d) and ath10k (http://paste.debian.net/1048278/) Oct 20 17:34:26 My screen reader reads ar and r as the same sound Oct 20 17:34:40 ah Oct 20 17:35:09 technology-- Oct 20 17:35:11 thanks for pointing that out tho! :-) Oct 20 17:36:04 So I need to work out why it's not flashing Oct 20 17:36:31 Is it that the file on the web page has not bin updated or is that my router is not tacking the flash Oct 20 17:36:36 * russell-- looks at the downloads page Oct 20 17:38:53 i haven't reflashed my c7v2 since june 2016 Oct 20 17:39:05 Tapper: a few weeks ago it was the case for me that occasionally sysupgrades didn't stick on my archer. It would stick on the second attempt. Oct 20 17:39:30 file i used is https://downloads.openwrt.org/snapshots/targets/ar71xx/generic/openwrt-ar71xx-generic-archer-c7-v2-squashfs-sysupgrade.bin Oct 20 17:39:43 and the date is Tue Oct 9 13:37:27 2018 Oct 20 17:39:49 so old file Oct 20 17:41:35 the switch to 4.14 was committed on Oct 10 Oct 20 17:41:39 All the files on the page are old Oct 20 17:42:28 lol as if the ar71xx builds have bin broken for 11 days. Oct 20 17:42:39 since kernel 4.14 the images are getting too big for some devices and the build breaks Oct 20 17:42:51 and nobody pointed it out! Oct 20 17:46:56 Tapper: building from source should work Oct 20 17:47:08 I am quite proud I worked that out being the resident dumb ass! Oct 20 17:47:22 russell-- that was not snark at you mate Oct 20 17:47:38 lol Oct 20 17:47:39 I was working out how to spell as you typed it Oct 20 17:48:14 i built for wzr600dhp recently, it worked fine Oct 20 17:48:39 OK thanks Oct 20 17:49:00 It's for my AP so there is no grate need to build just yet Oct 20 18:00:31 I think I saw discussion about making 'running out of space' a soft error on the buildbot rather than the present hard error and no images at all. Oct 20 18:11:37 Tapper: for quite a few devices it isn't 'only' about the total firmware size, but the size of the kernel partition in particular (kernel 4.14 is significantly larger than 4.9 - on top of that there have been unrelated changes which increase the size of the kernel recently) Oct 20 18:16:08 Hauke: in comparison the non-working current HEAD, including current mac80211 (v4.19-rc5-0-g6bf4ca7fbc85/ v4.19-rc5-1-0-g05571dcd) and current ath10k, http://paste.debian.net/1048288/ normal WLAN clients work just fine, but WDS/ 4addr doesn't transfer any data (no ping, no ssh, etc.) Oct 20 18:18:51 nbg6817-dualboot --toggle-rootfs Oct 20 18:18:53 grrr Oct 20 19:59:12 huaracheguarache, your crash was rate-ctrl mem corruption I've been chasing. To concentrate my effort, I'm forking off a branch that is reproducing this, and larding it up with debugging Oct 20 19:59:33 it is not 'b' fw, but I'll port the bugfix forward when I find it Oct 20 19:59:58 and older 'b' fw has rate-ctrl issues anyway, but changing too many things at once is going to make it harder to debug and fix Oct 20 20:04:41 huaracheguarache, please watch github bug 38 and try FW when I upload it... Oct 20 20:04:55 https://github.com/greearb/ath10k-ct/issues/38 Oct 20 20:14:55 image posted...off to do non-firmware related weekend things! Oct 20 21:02:58 Hauke, nbd: just tested reverting "mac80211: support station 4-addr mode fast-rx", "mac80211: support fast-rx with incompatible PS capabilities when PS is disabled" and "mac80211: support AP 4-addr mode fast-rx" from current HEAD and with current mac80211 (v4.19-rc5-0-g6bf4ca7fbc85/ v4.19-rc5-1-0-g05571dcd) using ath10k, no improvement with WDS/ 4addr Oct 20 21:25:24 EFI builds are failing on master: https://paste.fedoraproject.org/paste/bphu5mRQIrgJM~yQLOCw9Q Oct 20 21:30:08 @jow: seems to be keying on this line: "grub-mkimage-efi: error: relocation 0x4 is not implemented yet." Oct 20 22:02:28 Hauke, nbd: running hostapd_cli on the nbg6817 (WDS-AP), I see "<3>WDS-STA-INTERFACE-ADDED ifname=wlan1.sta1 sta_addr=14:cc:20:XX:XX:XX" and "<3>AP-STA-CONNECTED 14:cc:20:XX:XX:XX" with the old 4.14 based mac80211/ backports package, while the current/ 4.19 based one only gives me "<3>AP-STA-CONNECTED 14:cc:20:XX:XX:XX" Oct 20 22:08:43 mangix: https://github.com/openwrt/packages/blob/master/lang/python/python-crypto/patches/100-CVE-2013-74459.patch needs renamed to 100-CVE-2013-7459.patch Oct 20 22:15:49 the WDS-client has not been touch between those tests with different mac80211 version on the WDS-AP (nbg6817), while it currently runs the old mac80211 package as well, there is no difference with both running the new mac80211 version (ath9k seems to be fine) **** ENDING LOGGING AT Sun Oct 21 03:00:00 2018