**** BEGIN LOGGING AT Thu Oct 31 02:59:58 2019 Oct 31 07:21:20 build #32 of ramips/mt7621 is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/ramips%2Fmt7621/builds/32 blamelist: Yousong Zhou Oct 31 07:53:28 /builder/shared-workdir/build/scripts/ubinize-image.sh: 12: /builder/shared-workdir/build/scripts/ubinize-image.sh: hexdump: not found Oct 31 07:57:20 oh, we fixed it already in c6d41c320c795b0bb9a9350c7d4b1ce55f25769a Oct 31 08:06:14 weee Oct 31 09:01:39 any idea why I can't edit https://openwrt.org/about ? Oct 31 09:02:11 just wanted to add blocktrron to the People list Oct 31 09:07:48 anyone happen to have a copy of the ralink SDK? I need it for an MT7628 Oct 31 09:35:59 ynezz: maybe only wiki admins like tmomas? are you a wiki admin? Oct 31 09:36:26 ynezz: you have to be wiki admin Oct 31 09:39:49 ynezz: added blocktrron Oct 31 09:49:01 thanks! Oct 31 09:49:39 btw anyone with usable emails for kaloz and cyrus? Oct 31 09:51:32 all publicly available emails have @openwrt.org which probably doesn't work anymore Oct 31 10:42:10 Steven Barth Oct 31 10:42:18 Imre Kaloz Oct 31 10:42:50 csrf: what do you need from the SDK ? Oct 31 10:57:27 guys, for what it's worth: i've been running the hostapd 2.9 backport from ynezz on my 19.07 builds for like a week and no issues so far. Oct 31 10:57:47 qualcomm based smartphone and intel 8260AC wifi as clients. Oct 31 11:07:47 blogic: thanks! Oct 31 11:18:27 ynezz, i don't have a vote, but maybe you should propose a rule to put a reasonable time limit on the voting period. Oct 31 11:19:29 russell--: this is not an issue I would say, the issue is to reach majority of votes Oct 31 11:20:21 has that not happened already? Oct 31 11:22:23 I don't follow, what do you mean? Oct 31 11:22:36 Adrian is solved, he has 18 votes out of 29 I would say Oct 31 11:23:29 this "easy" votes, where almost everybody agrees are not an issue Oct 31 11:23:33 i guess the ambiguity is what constitutes "unreachable" Oct 31 11:24:49 well, there was some similar discussion off-list, and I really don't like this off-list stuff unless it's really necessary, like the SPI->SFC switch for example Oct 31 11:25:00 so I've just brought it into open Oct 31 11:26:17 a non-profit i'm involved with has some similar issues with making sure people have due notice Oct 31 11:31:18 "warrant canary"-like periodic messages to openwrt-admin, maybe, so if there are delivery problems, silence is the warning Oct 31 11:33:04 indeed, this was the cause with xback for example :p Oct 31 11:35:43 true :-) I simply forgot to subscribe to openwrt-admin .. Oct 31 11:37:39 :D Oct 31 11:42:53 another thought is, if there are substantial numbers of no votes, that implies there isn't an overwhelmning consensus, and maybe that's a reason to slow down and think about how to resolve the controversy better. if you are close to 50-50, that's a problem right there. Oct 31 11:43:27 anyway, those are my deep thoughts for the evening, i'm going to sleep Oct 31 12:00:04 blogic: btw, that auth_cache thing ldir linked does not fix 802.11r (assuming i put it in the right place in uci) Oct 31 12:02:12 Any ideas why I'm getting these weird "selected processor does not support `something ' in ARM mode" errors? https://paste.ubuntu.com/p/T7KjRDbXdh/ Oct 31 12:04:05 I'm trying to create a mt7629 subtarget, and it's using cortex-a7 without neon and fpu. It compiles fine if I add CPU_SUBTYPE=neon-vfpv4 but of course I'll then get illegal instruction errors. Oct 31 12:06:59 gch981213: maybe the kernel cflags still specify neon and/or fpu Oct 31 12:16:27 nbd: But those aren't even float point instructions. Oct 31 12:17:33 whoops...I found an -march=armv5t in kernel gcc argument. Oct 31 12:32:56 I have seen a patch was accepted on Patchwork, but it is not available in git log from openwrt master. How long does it take for patch to appear as actual commit in master? Oct 31 12:34:11 varies. which member accepted it? you might ask them directly Oct 31 12:36:21 Found the problem. fpu appears in the FEATURES field in target/linux/Makefile. Oct 31 12:37:32 *target/linux/mediatek/Makefile Oct 31 12:37:37 blogic actually Oct 31 12:40:20 yeah, blogic tends to accept patches into his staging tree first, while publishing that tree only sporadically Oct 31 12:40:24 be patient Oct 31 12:43:28 Necrosporus: which one, maybe i lost one in flight ? Oct 31 14:18:48 * Monkeh yawns Oct 31 19:14:36 swalker: oh interesting. I thought 2.1a was alpha. but it's stable. Oct 31 19:22:11 build #33 of ramips/mt7621 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org/openwrt-18.06/images/builders/ramips%2Fmt7621/builds/33 Oct 31 21:51:58 blogic: I wanna see if I can build a small, custom (2.6?) kernel for a cheap wavlink wifi extender. it has a MT7628KN Oct 31 21:58:59 why 2.6 Oct 31 21:59:20 modern kernels are far less painful Oct 31 22:17:08 DonkeyHotei: cuz I *believe that's all that was available for that device via the SDK, and I think it can't handle anything much newer Oct 31 22:17:26 only has 8MB RAM Oct 31 22:17:35 2MB flash Oct 31 22:17:39 8mb, ouch Oct 31 22:17:50 i just want super basic, CLI routing via a simple kernel Oct 31 22:17:58 i don't need openWRT or anything fancy Oct 31 22:18:12 basic *linux/CLI routing Oct 31 22:18:19 might be better with an rtos Oct 31 22:18:54 the stock firmware on this device is crap, so I'd rather have something I can configure manually, even if there's no web config Oct 31 22:22:50 anyway, the sdk seems to be buried in chinese file-sharing sites, which are really difficult to work with Oct 31 22:23:12 if someone already has a copy of the SDK laying around, I'd appreciate it Oct 31 22:34:25 DonkeyHotei: rtos? is that a better way to go for low-ram setups? Oct 31 22:35:04 that was my assumption, yes Oct 31 22:36:31 I have zero experience with rtos Nov 01 00:05:12 greearb_: used remote loging, 1.8M file with 3 crashes... u want all that or should I cut it down? Nov 01 00:05:48 also dump came from logread, I can used sed to remove the header or just leave it be Nov 01 00:06:00 header on each line that is Nov 01 02:41:24 csrf: Nope. 8M memory is still not enough for 2.6 kernel if you want any wifi support. MTK has an eCos SDK for mt7628kn (which I believe is ugly&buggy&discontinued) Nov 01 02:42:21 what's the difference between luci-mod-rpc vs uhttpd-mod-ubus, both are for remote rpc-over-http calls it seems **** ENDING LOGGING AT Fri Nov 01 03:00:13 2019