**** BEGIN LOGGING AT Sun Jul 22 03:00:00 2018 Jul 22 03:24:57 aparcar[m]: of course: https://github.com/openwrt/packages/pull/6392 Jul 22 03:51:56 mangix: just curious, my go package PRs are still on hold *sigh* Jul 22 03:53:23 you're not alone Jul 22 03:59:01 wait, there is working go support now? Jul 22 03:59:15 as in, I can package up stuff and have it built for the target? Jul 22 03:59:34 of course Jul 22 03:59:45 oh how delightful Jul 22 04:01:16 i have a feeling someone recently broke travis Jul 22 06:54:52 mangix_: can you send a link? Jul 22 09:15:06 distemper: ping Jul 22 09:16:56 Does openwrt environment allows to create device nodes ? Jul 22 10:32:40 pkgadd: hi! is there any word on dsa for ipq806x? things seem to have stalled a bit. Jul 22 10:40:59 Hi, new versions of ubiquiti nano m xw are becoming more and more restrictive, right? Jul 22 10:41:28 what is the best fit of ubiquiti nano m xw firmware so I can flawessly upload openwrt firmware Jul 22 12:00:20 jwh: late response, but in general yes. you could also do -DLOG_PGSQL=$(if $(CONFIG_PACKAGE_blah_option),TRUE,FALSE) Jul 22 12:26:25 KanjiMonster: oh thats better Jul 22 15:31:01 Hello! Does anybody know how to contact with user *blogic*? Or with other developers for the MTK 7623? Jul 22 15:32:17 hes here, just wait until hes around Jul 22 15:35:33 Ok. I'll try to wait some more... Jul 22 15:36:29 irc is for patient people :) Jul 22 15:38:48 Yes, I noticed it. Unusual routers too. for a year I can not start my router. Jul 22 15:38:53 :) Jul 22 15:42:23 Tohin: I think he'll be back tomorrow Jul 22 15:47:23 he is playing Counter Strike probably.... :-) Jul 22 16:10:57 It's strange. SoC mediatec 7623 has been actively advertising in Russian online media since 2015. In early 2017, the output of bpi-r2 was announced. Already a year as R2 went on sale. And still there is not a single router with full support ... Jul 22 16:13:02 Tohin: why is it strange? Jul 22 16:13:18 apparently there is not much interest in it Jul 22 16:14:06 :( Jul 22 16:15:07 probably the bpi-r2 (I guess BananaPi?) is not a traditional router device Jul 22 16:15:12 *because Jul 22 16:30:55 jow: is the "Force 40MHz mode" tick box under wireless settings new in luci? Jul 22 16:31:22 jow: I can't remember seeing it before Jul 22 16:31:45 question to all: Will be WPA3 implemented in OpenWRT, or do you plan to stay on WPA2? Jul 22 16:33:15 ultito: is there a whitepaper out yet that describes WPA3 in detail? Jul 22 16:33:52 huaracheguarache: no idea :-) Jul 22 16:36:18 just asking, if there is some plan for that Jul 22 16:36:25 ultito: there is no plan Jul 22 16:36:45 it will naturally appear byitself as openwrt jumps to newer version of hostapd and wpa_supplicant Jul 22 16:37:01 drivers without 11w support will not be able to support wpa3 though Jul 22 16:37:47 wasn't having ath10k having issues with 802.11w recently? Jul 22 16:37:56 Borromini: think so Jul 22 16:40:01 hmm weird. luci-nginx was in 18.06 and now it's a broken symlink in the package feed it seems Jul 22 16:40:36 jow: that tick box also appears in the advanced settings of my 5ghz network. Is it there in case I run it in N-mode? Jul 22 16:42:36 Borromini: yeah I dropped it because its not working Jul 22 16:42:57 huaracheguarache: yeah its new Jul 22 16:45:41 jow: ok. btw, I took the consolidated ath10k patch for a spin and I keep getting: daemon.err hostapd: DFS start_dfs_cac() failed, -1 Jul 22 16:46:37 not sure if it is really related to the patch Jul 22 16:47:20 it happens when I try to set up a 160mhz network Jul 22 16:47:49 jow: what means 11w support? Jul 22 16:48:46 ultito: 802.11w Jul 22 16:49:09 protected management frames https://www.networkworld.com/article/2310261/tech-primers/802-11w-fills-wireless-security-holes.html Jul 22 16:49:26 ultito: management frame protection support, basically means that control traffic such as rts, cts, deauth etc. are encrypted as well Jul 22 16:49:28 ldir: ping Jul 22 16:49:39 usually only data frames are encrypted Jul 22 16:49:50 jow: if wpa3 enforces this, won't that mean a lot of clients won't support it as well? Jul 22 16:50:06 i'm thinking smartphones and whatnot, which don't really get major driver updates often Jul 22 16:50:18 Borromini: anything "hardmac" or "fullmac" will not support unless there's silicon or device firmware blob updates Jul 22 16:50:54 i'll have to google that :) Jul 22 16:51:00 look in the past how long the TKIP->CCMP transition took, I guess 11w / SAW will be similar Jul 22 16:51:14 *SAE Jul 22 16:51:48 i barely knew how wireless worked by the time the TKIP CCMP transition came around :P Jul 22 17:06:56 re Jul 22 17:15:40 Borromini: pong Jul 22 17:16:07 ldir: hey. i saw in your exp branch a comment about x86 ext4 images not working with the .56 kernel? Jul 22 17:16:30 i just bumped 18.06 to .57 here and booted a combined ext4 squashfs image from a usb stick, it loads fine Jul 22 17:16:42 I think that's from stintels bump. Jul 22 17:17:44 oh Jul 22 17:17:48 stintel: ping Jul 22 17:17:50 and I'm going to have to take at least a week out. Jul 22 17:18:12 going abroad? Jul 22 17:18:26 if you want 30°C and over you can come to belgium :P Jul 22 17:18:29 no, doing too much, affecting health. Jul 22 17:22:04 sorry to hear. that sucks Jul 22 17:22:39 * Tapper hugs ldir Jul 22 17:23:39 jow: It's strange that when the Asus Wl-500GP appeared, it was very popular because of the flexible settings, GPIO and usb. BPI-R2 is an even more flexible router, but not popular among developers .. Sorry for my English, I use google translate. Jul 22 17:52:10 * blogic hugs ldir Jul 22 17:52:15 why are we hugging him ? Jul 22 17:52:36 went on vacation with my daughter Jul 22 17:52:44 flew for near 4 hours Jul 22 17:52:52 spent all 4 days on a playground :-) Jul 22 17:57:47 nice Jul 22 17:57:59 blogic: because he's working too much Jul 22 17:59:36 Borromini: i dont believe that :-) Jul 22 18:07:39 well if he has difficulties saying no i do :) Jul 22 18:11:49 work, what is that? :-D Jul 22 19:01:14 updated openwrt/upstream, https://sdwalker.github.io/uscan/index.html Jul 22 19:30:04 is stintel on vacay, anybody know? if so, pinging him doesn't make much sense. Jul 22 20:14:42 can someone tell me if this is how i should comments/remarks about a patch i send in? Jul 22 20:14:45 https://lists.infradead.org/pipermail/openwrt-devel/2018-July/013295.html Jul 22 20:17:30 aiui, yes Jul 22 20:18:47 karlp: ok thanks. Jul 22 20:19:36 * Borromini is off Jul 22 20:19:37 goodnight. Jul 22 20:27:12 jow: apparently Broadcom's change for firmware loading broke using NVRAM from MTD partition Jul 22 20:27:20 jow: i'm going to debug/fix that Jul 22 20:28:31 rmilecki: hm, and that breaks booting entirely? Jul 22 20:28:39 I guess because its unable to infer the board anymore Jul 22 20:28:48 no... i'm talking about Netgear R8000 WiFi regression now Jul 22 20:28:51 ah Jul 22 20:28:52 ok Jul 22 20:29:39 neoraider: ping Jul 22 20:35:46 jow: pong Jul 22 20:39:34 neoraider: I wonder if you could help looking into the various reported ath10k 5ghz issues sometime during the next week as you have access to a larger hardware base with gluon. See e.g. FS#1682, FS#1665, FS#1607 Jul 22 20:40:08 blogic said it could be a case of a ath10k firmware / driver support mismatch but idk Jul 22 20:40:19 just guesswork at this stage Jul 22 20:41:57 that, along with the brcm47xx fallout rmilecki is thankfully looking into, is the two blockers I see for final Jul 22 20:42:02 the current ath10k firmware (now with the 160 MHz issue fixed) should be o.k., but for updated firmware a newer mac80211/ ath10k snapshot would be needed (as there are plenty of changes for the newer firmware ABI) Jul 22 20:43:37 brcmfmac regression is caused by 2811c97803 ("mac80211: backport brcmfmac firmware & clm_blob loading rework") Jul 22 20:43:39 jow: hmm, interesting, we did not get any complaints about ath10k from our bleeding edge testers after switching the Gluon master to 18.06 Jul 22 20:44:05 out of curiosity, what's the plan for https://patchwork.ozlabs.org/project/openwrt/list/?series=37466 "base-files: add new backup options" for improving sysupgrade? I'm using it successfully since april. Jul 22 20:44:31 neoraider: maybe it is something completely stupid like a changed sysfs path or something... but the fact that there's various reports across various devices makes me suspicious Jul 22 20:44:48 jow: the first one you linked is actually mt76, not ath Jul 22 20:45:04 https://git.openwrt.org/?p=openwrt/openwrt.git;a=commit;h=02f815d1907cdd7e042415a2b4a749c819087168 Jul 22 20:45:08 is this maybe related to the first one? Jul 22 20:45:35 achterin found the re350 issue and blocktrron found said commit that was reverted by blogic Jul 22 20:45:56 neoraider: ah indeed Jul 22 20:46:09 hexa-: unlikely, the report it about 18.06-rc2, that fix is much older Jul 22 20:48:36 jow: regarding FS#1665: I remember we had a hack in earlier versions that made OTP calibration failures non-fatal Jul 22 20:48:59 jow: it was removed a few months ago because it was supposedly not necessary anymore because of other improvements in ath10k Jul 22 20:49:10 my main issue is that I lack any ath10k hw, should probably organize myself some archers or so Jul 22 20:49:23 my local hardware zoo is about ten years old on average, except the apu Jul 22 20:49:28 so not much use for testing :D Jul 22 20:53:17 Hi jow I have a hi trafic network using a c7 if you let me know witch snapshot I should test I will Jul 22 20:57:56 the cheapest option for ath10k hardware (QCA9880-BR4A) would be the BT Home Hub 5 type A Jul 22 20:58:34 which requires somewhat heavy modifications afer purchase right? Jul 22 20:59:37 karlp: it requires opening the case (easy) and then flashing via serial console, that is a bit tricky (because very small test points/ vias) - but it's doable Jul 22 20:59:59 neoraider: after reviewing the tickets again I noticed that one is about mt76 as you pointed out, one is broadcom which rafal is looking into and the remaining one is a ar71xx which might have been affected by https://git.openwrt.org/?p=openwrt/openwrt.git;a=commitdiff;h=f3bdc4a76aa2ee74c16495991e4ce5b95680f0db Jul 22 21:00:14 neoraider: so maybe it is not widespread after all and just some conincidence Jul 22 21:02:05 QCA9984 (ipq8065/ nbg6817) and QCA9880 (BT HH5a) are working pretty well for me (not yet running the latest ath10k changes, will probably build that later tonight) - but it apparently depends a lot on the client devices in use Jul 22 21:03:03 pkgadd: thanks for the suggestion, will see if I can obtain some cheaply Jul 22 21:03:46 jow: I can spare you an HH5A with a predilection for hard locking ;P Jul 22 21:04:36 jow: you'll find them with the vendor firmware for ~8 GBP, the bigger issue are ~11 EUR shipping to .de (preflashed with LEDE would be ~30 EUR), a variant of this https://openwrt.ebilan.co.uk/viewtopic.php?f=14&t=965 worked quite well for me (without soldering, although I assume you might prefer a permanent serial connection) Jul 22 21:05:03 pkgadd: That is horrible. :) Jul 22 21:06:04 Monkeh: I wouldn't suggest to do it exactly that way (that much tape comes with a high risk of ripping SMD components from the PCB), but doing it with a bit more care works surprising well Jul 22 21:06:39 fortunately there is enough empty space around it without risking anything Jul 22 21:06:54 Eh, solder no scare. Jul 22 21:07:29 Monkeh: I gave it a chance, knowing that I could solder it down later, if needed Jul 22 21:08:09 ...and it succeeded nicely (but involved me keeping my thumb firmly pressed on it for ~40 minutes) ;) Jul 22 21:16:32 Yeah, nuts to that. Jul 22 22:13:27 jow: got my email? Jul 22 22:36:32 hrmpf. just notice my ath9k usb wifi stick works with 17.01.5 - but not with snapshot images (neither one from 2017-06-22 nor one from today) Jul 22 22:38:13 rotanid: and 18.06? Jul 22 22:38:35 the kernel module kmod-ath9k-htc says the following using a snapshot build - compared to 17.01.5: https://pastebin.com/raw/j3YEBK5u Jul 22 22:38:54 I've used ath9k_htc on ar71xx (tl-wdr3600) until april, regularly updated master builds. it always worked (but not as reliable as I wanted it to be, many wireless stalls) Jul 22 22:40:09 Borromini: silly me, trying 18.06 rc2 now Jul 22 22:41:52 rotanid: either way, if master breaks that's worth fixing as well Jul 22 22:42:07 o/ pkgadd Jul 22 22:43:34 pkgadd: testing on a wdr3500 now. maybe it's also target-related, as i tested master on a lantiq device. we will know in a few minutes. Jul 22 22:44:26 I used the ar9271 USB stick in order to avoid the repeater effect (internal 2.4 GHz wlan as dedicated uplink, ar9271 and internal 5 GHz as AP), I wouldn't recommend doing that anymore - the repeater effect is much less of a problem Jul 22 22:50:31 ok, it Jul 22 22:50:53 ok, it's working on ar71xx wdr3500. any idea why it's not working on lantiq xrx200 (fb 3370) ? Jul 22 22:51:43 "[ 12.857799] ath: phy1: Mac Chip Rev 0x0f.3 is not supported by this driver" Jul 22 22:52:24 both are big endian, so that shouldn't be the reason Jul 22 22:53:13 and same cpu architecture as far as i know Jul 22 22:55:39 ah, different kernel version! Jul 22 22:55:48 4.9 with the wdr3500, 4.14 with the 3370 Jul 22 23:09:58 can't find a ath79 build for the wdr3500, might have to build myself to see if it's the target or the kernel version Jul 22 23:10:19 seems ath79 is not part of the "official" snapshot builds yet Jul 22 23:11:03 the wdr3600 works well with ath79, the only difference between them is the hardware switch - getting it ported will take some effort, but not a whole lot Jul 22 23:12:17 it looks like it's ported already, there just aren't builds for this target yet (or am i blind?) Jul 22 23:12:56 nope, it's still marked as source-only and not built on the buildds Jul 22 23:13:03 i know how to build, it just takes much more time to try out things if you have to build yourself Jul 22 23:13:21 ~20 minutes Jul 22 23:15:30 but no, there is no support for the tl-wdr3500 in ath79 yet (only wdr3600 and wdr4300) Jul 22 23:20:03 ah, mixed it up with the folder ath79 in ar71xx sources Jul 22 23:20:44 ok, so i'll search for a ath79 supported device and check again if the kmod-ath10k-htc works there for the ar9271 stick Jul 22 23:21:37 there is a closed pull request for forward porting ar71xx to kernel 4.14, that one needs rebasing as ar71xx has changed in relevant ways in the mean time, but it's doable, if you want to test kernel 4.14 on ar71xx Jul 23 00:01:17 ok, tried with wdr3600 on ath79 now, kernel 4.14, got the binary build from https://forum.lede-project.org/t/ath79-builds-with-all-kmod-packages-through-opkg-flow-offloading/15897 Jul 23 00:01:45 with this, even the error message doesn't come up, like with my own build from 2018-06-22 Jul 23 00:02:07 no message whatsover regarding a new wifi device, only one line about a new usb device Jul 23 00:05:51 did you really installed the ath9k_htc modules? Jul 23 00:06:28 "opkg list-installed | grep ath9k" says: yes (and i rebooted) Jul 23 00:24:47 lol been away a bit guess everyones back in here Jul 23 00:25:35 Package mtr is missing dependencies for the following libraries: libcap.so.2 Jul 23 00:43:23 nyt: good news. that's partially my fault :) Jul 23 00:44:08 hrm buildbot had no issues with it Jul 23 00:46:20 i have libcap enabled Jul 23 00:46:25 after i saw it the first time Jul 23 00:46:27 then built again Jul 23 00:46:30 same error Jul 23 00:46:33 so dunno whats up Jul 23 00:48:34 not sure where the dependency is even coming from Jul 23 00:48:45 mtr only lists libncurses Jul 23 00:51:08 it's from the configure script Jul 23 00:51:27 where? Jul 23 00:52:10 configure.ac. try this patch. https://gist.github.com/neheb/c014747ff9f6a1247f6de8c188901195 Jul 23 00:53:23 gtg for a while Jul 23 00:55:07 will give it a go Jul 23 00:56:23 last build was in january, so just merged everythign into my tree and getting it building again Jul 23 01:07:52 i've built ath79 for wdr3600 myself now. i even cannot use "lsusb" now, "failed to initialize libusb: -99" Jul 23 01:08:18 libusb-1.0 is installed Jul 23 01:41:01 i managed to get the ar9721 wifi usb stick working with kernel 4.14 (ath79 target, device wdr3600) Jul 23 01:42:00 but i still fail on the lantiq xrx200 device with the same message... "ath: phy1: Mac Chip Rev 0x0f.3 is not supported by this driver" Jul 23 01:43:33 nyt: did that work? Jul 23 01:44:35 I changed IRC browsers (from Chatzilla to Adium) and now I can't get back into #lede-dev… Jul 23 01:44:58 philipp64: the channel lede-dev is deprecated. therefore new joins are blocked. Jul 23 01:45:09 philipp64: both are crummy IRC clients IMO Jul 23 01:45:13 try textual? Jul 23 01:45:17 or colloquy? Jul 23 01:45:24 They run on OS X? Jul 23 01:45:46 yes Jul 23 01:45:47 I'm trying to have a unified (cough) messaging agent. Jul 23 01:45:53 fair enough Jul 23 01:45:57 philipp64: irssi? Jul 23 01:46:00 but you're in the dev channel Jul 23 01:46:24 so? Jul 23 01:47:05 #lede-dev effectively no longer exists. you're here. Jul 23 01:47:09 rotanid: how long before the channel gets turned off? Jul 23 01:47:54 philipp64: you can't turn off a channel. you can block joining it. if you delete it, the next one joining creates it again Jul 23 01:48:10 #lede-dev forwards users here. Jul 23 01:48:10 at least as far as i know Jul 23 01:48:13 Fishman: when I type "colloquy" in the OS X App Store, it shows me "textual 7". odd. Jul 23 01:48:32 philipp64: it may not be on the MAS Jul 23 01:48:35 rotanid: thanks, didn't know that. Jul 23 01:50:20 mangix: irssi run on Mac? Jul 23 01:50:29 of course Jul 23 01:50:29 it can Jul 23 01:50:46 not seeing it on the App Store. Jul 23 01:50:52 it's cli Jul 23 01:50:57 ah. Jul 23 01:51:08 i'm running it on my router right now Jul 23 01:51:15 since it's up 24/7 Jul 23 01:52:05 rotanid: well, if you're sure it's not a cabal. Jul 23 01:52:33 cabal? Jul 23 01:53:13 "a secret political clique or faction", according to Websters. Jul 23 01:53:34 Wikipedia says, "A cabal is a small group of people united in some close design, usually to promote their private views of or interests in an ideology, state, or other community, often by intrigue and usually unbeknownst to those outside their group." Jul 23 01:54:09 just to feed the imaginations of the conspiracy theorists. Jul 23 01:55:30 * philipp64 notes the long silence… Jul 23 01:55:48 * mangix is busy with build failures Jul 23 01:56:22 philipp64: pretty sure Jul 23 01:57:05 well, okay then. Jul 23 02:12:17 blogic: where the hell did you go 4 hours away that has a playground wherever you live doesn't? :D **** ENDING LOGGING AT Mon Jul 23 03:00:01 2018