**** BEGIN LOGGING AT Wed Jul 11 03:00:02 2018 Jul 11 03:03:03 bug confirmed Jul 11 03:05:59 Tue Jul 10 20:01:23 PDT 2018 277558.31 1107005.01 total used free shared buffers cached Mem: 253668 47072 206596 948 2760 12904 -/+ buffers/cache: 31408 222260 Swap: 125948 0 125948 Jul 11 03:05:59 Tue Jul 10 20:02:23 PDT 2018 277618.33 1107244.20 total used free shared buffers cached Mem: 253668 48512 205156 948 2760 12904 -/+ buffers/cache: 32848 220820 Swap: 125948 0 125948 Jul 11 03:05:59 Tue Jul 10 20:03:23 PDT 2018 277678.35 1107483.75 total used free shared buffers cached Mem: 253668 48948 204720 948 2760 12904 -/+ buffers/cache: 33284 220384 Swap: 125948 0 125948 Jul 11 03:06:03 etc. Jul 11 03:06:57 leaky driver? Jul 11 03:07:00 buffers? Jul 11 06:52:56 DonkeyHotei: why did you have one of the radios turned off? Jul 11 07:42:13 nbd: ping Jul 11 07:43:28 Hauke: ping Jul 11 08:08:21 russell--: just a random guess, can you try this patch: https://bugs.openwrt.org/index.php?getfile=507 ? Jul 11 08:08:55 Hauke: do you have some BCM4718A1 device maybe? https://bugs.openwrt.org/index.php?do=details&task_id=1647 Jul 11 08:13:35 dedeckeh: thanks for doing the iproute2 bump Jul 11 08:30:38 ldir:np Jul 11 08:38:36 DonkeyHotei: ping Jul 11 08:57:50 jow: trying Jul 11 09:03:48 neoraider: ping Jul 11 09:09:24 neoraider: un-ping Jul 11 09:16:03 jow: ? Is there difference between https://github.com/openwrt/packages/blob/master/utils/lxc/Config.in#L31 and CONFIG_PACKAGE_xz=y ? Jul 11 09:17:04 cuz i can not compile wrt with CONFIG_LXC_BUSYBOX_OPTIONS=y and CONFIG_PACKAGE_xz=y togheter Jul 11 09:17:27 CONFIG_PACKAGE_xz=y is gnu ?\ Jul 11 09:18:07 the former enables xz in busybox, the latter refers to the standalone xzutils package Jul 11 09:19:02 installable packages usually prefer (and are supposed to) other installable packages as dependencies Jul 11 09:19:19 since someone using opkg cannot simply recompile his busybox to contain further applets Jul 11 09:20:08 ah, ok. Thanks Jul 11 09:20:08 and having lxz auto-select xz in busybox would cause the busybox package to enable xz for all users, regardless of whether they use lxc or not Jul 11 09:20:16 *having lxc Jul 11 09:20:34 jow: ping Jul 11 09:20:41 f00b4r0: pong Jul 11 09:20:50 jow: saw my message from monday? Jul 11 09:21:01 netlock stuff? yeah, noted it Jul 11 09:21:15 sounds good to you? Jul 11 09:21:36 yes Jul 11 09:21:45 cool. Still no time to try the new config? :) Jul 11 09:22:40 no Jul 11 09:22:51 *sigh* Jul 11 09:23:09 need to do 17.x today, luci today and tomorrow, backports tomorrow, 18.06 on friday Jul 11 09:23:23 heh, sounds busy alright ;) Jul 11 09:23:24 in the meanwhile also paid work and work on my side job Jul 11 09:23:33 so not before next week Jul 11 09:23:40 backports master git is broken Jul 11 09:23:42 ok. Jul 11 09:23:55 On what email address can I send patches, please? Jul 11 09:24:11 mailing list Jul 11 09:24:16 openwrt-dev Jul 11 09:24:38 iam confused with this https://openwrt.org/contact Jul 11 09:25:00 1. OpenWrt-Devel archives and info - Mailing list for submitting patches and general development related work Jul 11 09:25:15 https://lists.openwrt.org/mailman/listinfo/openwrt-devel Jul 11 09:25:30 openwrt-devel@lists.openwrt.org ? Jul 11 09:25:35 yes Jul 11 09:26:02 does upgrading from 17 to 18 preserve config? Jul 11 09:27:28 it should Jul 11 09:28:02 thx Jul 11 09:32:49 ha. Mikrotik kindly provided their updated kernel patch. One step closer. Jul 11 09:36:38 jow: free is still falling, looks like your patch didn't fix it Jul 11 09:45:29 mkresin: http://paste.debian.net/hidden/8e739f25/ Jul 11 09:55:52 ashkan: are you the guy from the bug report? Jul 11 10:05:22 mkresin: nope, but the same device/problem I guess Jul 11 10:06:11 ashkan: not sure yet, your nand doesn't seem to have bad blocks Jul 11 10:06:43 ashkan: Am I right that my patch fixed at least the ubi auto attach or worked it before? Jul 11 10:08:28 ashkan: can you show me a 17.01.4 bootlog? Jul 11 10:09:57 mkresin: with your patch ~> http://paste.debian.net/hidden/e18a02e3/ Jul 11 10:14:23 ashkan: the same build causes different issues, depending if booted from flash or ram. strange Jul 11 10:18:43 ashkan: there were quite some changes to target/linux/ar71xx/files/drivers/mtd/nand/ar934x_nfc.c for kernel 4.9 support. if LEDE 17.01.4 works, most likely one of these is wonky Jul 11 10:22:18 mkresin: should I try this with your patch ? https://forum.lede-project.org/t/ecc-errors-in-ubi-rb2011uias-2hnd-in-and-solution/16424 Jul 11 10:23:47 curl 7.6.10curl 7.61.0 is out, including a heap-buffer-overflow security fix (CVE-2018-0500). Jul 11 10:31:30 ashkan: are you able to hack up a patch on your own? Jul 11 10:40:47 mkresin: atm I'd do anything to get my router back to life :D Jul 11 10:46:22 if you have serial on the device you may not need to rebuild an image to perform the change reported in the forum post Jul 11 10:46:34 you just need to edit the boot arguments Jul 11 10:52:22 I rather checked if there is an easy way to backport the linked patch. doesn't look like Jul 11 10:53:12 and I still struggle to understand why it worked with kernel 4.4 Jul 11 10:56:43 i don't see nand_samsung.c in 4.4 Jul 11 10:57:08 assuming that implies this chip used some generic default code, based on a quick grep it appears that NAND_NO_SUBPAGE_WRITE is an opt-in Jul 11 10:57:14 so wasn't turned on in 4.4 for that chip. Jul 11 10:57:31 QED? ;) Jul 11 10:59:38 hmm no i'm reading this backwards Jul 11 10:59:44 nevermind what I said Jul 11 11:00:57 f00b4r0: I've tested the change on forum but it didn't work Jul 11 11:03:51 that's why I thought I may build it with mkresin patches on top Jul 11 11:04:14 ashkan: I'm quite sure you have another issue Jul 11 11:10:26 ashkan: your best bett is to define your nand chip at: https://elixir.bootlin.com/linux/v4.9.111/source/drivers/mtd/nand/nand_ids.c#L32 Jul 11 11:10:53 ashkan: with the NAND_NO_SUBPAGE_WRITE option set Jul 11 11:11:08 ashkan: assuming it is really your specific error Jul 11 11:16:17 mkresin: The chip is Samsung K9F1G08U0E Jul 11 11:17:37 ashkan: from the upstream commit: K9F1G08U0E 21nm chips do not support subpage write Jul 11 11:18:06 ashkan: no idea if there are K9F1G08U0E != 21nm Jul 11 11:32:51 [Tue 2018-07-10 11:52:56 PM PDT] DonkeyHotei: why did you have one of the radios turned off? <------ because the wifi bands are saturated here and i already have the archer c7 serving up 2.4 Jul 11 11:33:06 muhaha: pong Jul 11 11:34:53 DonkeyHotei Jul 11 11:34:53 DonkeyHotei: I successfully compiled wrt 18.06 with lxc, I had error with CONFIG_LXC_BUSYBOX_OPTIONS=y vs CONFIG_PACKAGE_xz=y, but now its ok. Jul 11 11:37:02 which option did you disable? Jul 11 11:38:19 muhaha: ^ Jul 11 11:38:22 CONFIG_PACKAGE_xz Jul 11 11:38:42 ? Jul 11 11:42:55 DonkeyHotei: https://github.com/openwrt/luci/pull/1422 Last two comments, that was all that i was able to track to fully enable lxc + luci app in 17.06... Jul 11 11:43:43 https://github.com/openwrt/packages/issues/6409 -> this is what i found in 18.06 Jul 11 11:43:50 *17.01 Jul 11 11:59:45 muhaha: was your error on building lxc, or on building xz? Jul 11 12:00:29 xz Jul 11 12:01:43 my error was on building lxc Jul 11 12:01:58 and still is Jul 11 12:02:37 Do you have a docker installed? Jul 11 12:02:42 no Jul 11 12:03:38 Thats how I do it.. Complete isolation between builds.. Jul 11 12:04:19 and how do you generate your .config? Jul 11 12:05:29 mv .config.seed; make defconfig Jul 11 12:07:18 i have also a weird step (https://github.com/lukasmrtvy/docker-lede-apu2c4/blob/snapshots-18.06/kconfig.sh) to edit kernel .config .. but it works eventually Jul 11 12:07:19 can you show your .config.seed? Jul 11 12:07:44 https://github.com/lukasmrtvy/docker-lede-apu2c4/blob/snapshots-18.06/.config Jul 11 12:10:40 unfortunately is not possible to do a defconfig to kernel .config, so I have to manually add "new" (enabled by my changes) options manually to kernel .config Jul 11 12:28:02 Sebastiii: ping Jul 11 12:33:50 Strykar: ping Jul 11 12:59:00 f00b4r0: ping, query? Jul 11 13:31:46 muhaha: i am building with -O3 and you are building with -Os Jul 11 13:31:55 this may be the crucial difference Jul 11 13:35:27 DonkeyHotei: mabye Jul 11 13:58:43 muhaha, rmilecki, luka: confirmed, in the 18.06 branch, lxc fails to build with gcc 7.3 for x86/64 with -O3 but builds with -O2 or -Os Jul 11 14:02:32 good. Do you fill a bug for it? :) Jul 11 14:02:46 just now discovered it Jul 11 14:04:31 now you are responsible for filling a bug report for it :D Jul 11 14:04:57 isn' Jul 11 14:05:07 t it pretty much expected that -O3 breaks things? Jul 11 14:06:07 with gcc 5.5 it built fine Jul 11 14:14:17 is this #gentoo Jul 11 14:14:32 no Jul 11 14:18:46 jow: seems identical to https://github.com/openwrt/packages/issues/6004 but on a different package Jul 11 14:23:48 jow: "Allegedly this has been fixed in later gcc versions..." citation? Jul 11 15:05:11 What is with that rat. Jul 11 15:08:29 * Monkeh still hasn't got this memory leak to occur Jul 11 15:40:55 Monkeh: if you have an mt7621 device, run it with both radios enabled and no clients associated Jul 11 15:50:32 SpaceRat Is still comming ina and out. What for? Jul 11 15:53:03 SpaceRat Hi why do you keep comming in and out? Jul 11 15:54:43 obviously a misconfigured irc client Jul 11 15:59:24 DonkeyHotei it's bin going on now for weeks. Jul 11 16:10:15 if i knew how & had privs I would ban from this channel. If actually notices can go to openwrt-adm to discuss/complain etc/ Jul 11 16:10:54 "simples" as those infected by UK advertising might say. Jul 11 16:15:10 Monkeh: you need to install 'worn-washer' Jul 11 16:18:30 DonkeyHotei: nice debugging work! Jul 11 16:18:33 thanks Jul 11 16:19:01 rmilecki: i'm about to submit a pr to fix it Jul 11 16:19:14 DonkeyHotei: great, cc me Jul 11 16:19:24 i'll ping you Jul 11 16:23:10 DonkeyHotei: Well I've been doing single radio tests to see if it's a specific driver Jul 11 16:23:18 Now running with both. Jul 11 16:23:33 Monkeh: the driver in use was mt76 Jul 11 16:23:42 DonkeyHotei: 'mt76' consists of two drivers. Jul 11 16:27:08 Now running both radios, memory usage is varying but not increasing so far Jul 11 16:30:44 nice, enabling LTO on busybox reduces the .ipk size from 214k to 207k Jul 11 16:31:08 LTO all the things Jul 11 16:31:16 ha! thought .55 came out soon after .54 - ext4 CVEs Jul 11 16:32:51 juicy Jul 11 16:36:54 rmilecki: https://github.com/openwrt/packages/pull/6460 Jul 11 16:41:12 Hello! I have downloaded the install .bin file from here (https://openwrt.org/toh/hwdata/tp-link/tp-link_archer_c7_v4), and I can't find a checksum on the page to compare to. Can anyone point me in the right direction? Jul 11 16:43:08 ChrisTYB: https://downloads.openwrt.org/releases/18.06.0-rc1/targets/ar71xx/generic/ Jul 11 16:50:36 ChrisTYB: also, this is the development channel; you want #openwrt for that question Jul 11 16:53:02 Oh, sorry, misread the web page. Thanks for the info! Jul 11 17:00:18 ping jow Jul 11 17:00:29 Can we have a short chat about nls.mk? Jul 11 17:06:00 Monkeh: are there clients connected to the wifi? Jul 11 17:06:56 No. Jul 11 17:07:48 russell--: when you disabled one of the radios, which radio was it? Jul 11 17:08:05 DonkeyHotei: 2.4 Jul 11 17:08:19 I shall experiment more later Jul 11 17:08:26 lsmod here says mt76x2e Jul 11 17:09:03 i have mt7603e and mt76x2e Jul 11 17:19:40 hey guys Jul 11 17:23:00 ldir: ping Jul 11 17:30:02 hello! Jul 11 17:30:38 can someone help me with some tips on flashing a TP-LINK WR740N v7 (EU) router? Jul 11 17:30:49 i cant seem to find it in the downloads section Jul 11 17:31:40 at the end of the FCC-ID it says v6, so i tried using that but i get a wrong filename error, that i cant fix even by changing the filename Jul 11 17:37:00 nice, by LTOing a few more things, i'm now at 17k gzipped size reduction Jul 11 17:37:17 i'd estimate around 10-12k after LZMA Jul 11 17:37:20 in the default image Jul 11 17:37:48 nbd: hi. is koen hanging out on irc by any chance? i don't know his nick if he is. Jul 11 17:38:55 nbd: Awesome Jul 11 17:39:17 Borromini: no idea Jul 11 17:39:37 nbd: ok thanks. Jul 11 17:41:58 sleepenthusiast: look at the file names tp-link is using for its own firmwares Jul 11 17:42:34 Borromini: pong Jul 11 17:43:28 sleepenthusiast: scratch that, you have a v7 you say and there's no openwrt support for it afaict, so unless you start hacking no way to get it running Jul 11 17:43:45 ldir: hey. thanks for chiming in on my 4.14 bump Jul 11 17:43:59 any one in the uk will not talk for a wile now because football! Jul 11 17:44:51 Np Am erratic at moment due to silly work hours. Sorry Jul 11 17:44:53 It's coming home! Jul 11 17:47:17 pfff Jul 11 17:47:37 still recovering from yesterday's game Jul 11 17:48:08 ldir: i have to re-read the mail convo again but if i'm understanding this right koen suggested to do the bump right and only look at the backport patch after? Jul 11 17:50:02 stintel I cant remember who were you backing? Jul 11 17:50:15 I'm from Belgium Jul 11 17:50:29 O sorry dude Jul 11 17:50:59 tsk tsk Jul 11 17:51:02 You boys gave it a good go! Jul 11 17:51:30 Borromini, im pretty sure the v7 is just the UI Jul 11 17:51:38 the hardware is the same in both V6 and V7 Jul 11 17:51:53 sleepenthusiast: tp-link will have a firmware header check that looks for v7 instead of v6 Jul 11 17:52:00 even if it's just cosmetics. wouldn't be the first time Jul 11 17:52:02 i guess ill just have to open it up and see for myself, if thats the case, im assuming an openwrt v6 should work Jul 11 17:52:06 ah damn Jul 11 17:52:12 how would i change that myself Jul 11 17:52:30 it's been way too long since i once attempted that Jul 11 17:52:57 well, other answers suggested only renaming the file Jul 11 17:52:57 i don't know if you can do it without compiling your own image Jul 11 17:53:06 as thats the only thing the tplink router looks at, but ive failed with it Jul 11 17:53:13 yeah sure im not afraid of compiling it Jul 11 17:53:16 it's the only thing the *web UI* looks at Jul 11 17:53:24 fair enough Jul 11 17:53:39 ill check the hardware soon enough and see what i can do Jul 11 17:53:55 although even their website has two v7 files, so im not sure about that either Jul 11 17:57:55 https://forum.openwrt.org/viewtopic.php?id=64002 < check that thread Jul 11 17:58:08 that's actually my old patch for the 740 v6, funny :) Jul 11 17:58:59 what you probably need is TPLINK_HWID := 0x07400006 - which should logically become 0x07400007 Jul 11 17:59:23 however, you do need to double check the tp-link binary for its header before changing openwrt's and building an image Jul 11 18:00:13 sleepenthusiast: do check hardware is identical first though, because doing this might fubar your hardware. Jul 11 18:10:34 unfortunately, the proper way to do it would be to compare the v6 and v7 gpl tarballs Jul 11 18:12:46 yeah this is the lazy way :) Jul 11 18:16:24 there's no v7 listed on tp-link.com Jul 11 18:16:54 probably an oversight Jul 11 18:17:51 pretty weird Jul 11 18:18:01 european site version: no v6 Jul 11 18:18:22 indian site version: does list a v7 for the EU and UN (no idea what UN means) Jul 11 18:21:43 740Nv7 appears to be a 741Nv6 Jul 11 18:22:19 you went through the tarball or someone else already did? Jul 11 18:22:47 FCCID of the 740Nv7 is that of the 741Nv6. Jul 11 18:22:53 oh Jul 11 18:22:59 that's a giveway i suppose :) Jul 11 18:23:27 I'm assuming this is the same guy who was asking about it in #openwrt recently Jul 11 18:23:48 no idea Jul 11 18:24:11 Anyway, I could get one but the odds of me spending money on a 4/32 device to port are low.. :) Jul 11 18:24:32 Especially as Stihl are probably about to bend me over for a new bearing. Jul 11 18:28:15 hehe Jul 11 18:52:40 boa tarde Jul 11 18:54:45 brazil left the world cup sorry. Jul 11 18:57:43 wasnt around but yeah it was me that asked in the other channel aswell Jul 11 18:57:55 ill look into the 741Nv6 aswell, thanks monkeh! Jul 11 18:57:59 how do i ping people btw Jul 11 19:00:08 ping $nick Jul 11 19:14:08 rmilecki: I think I have a similar device, I will look into FS#1647 at the weekend Jul 11 19:14:17 jow: ping Jul 11 19:21:11 Tobias-tSYS: ping perhaps Jul 11 19:26:05 Ahaha Jul 11 19:26:07 1:1 Jul 11 20:41:10 Well It'a all over now for us in the football :-( Jul 11 20:46:40 yeah, there goes my night sleep. Jul 11 20:47:23 (we got a huge serbian population here. they are already driving around honking.) Jul 11 20:47:54 * mamarley writes drmr's neighbors tickets for "improper use of horn." Jul 11 20:48:59 Can I somehow add symbol for busybox tar to support --numeric-owner option ? LXC_BUSYBOX_OPTIONS enabling BUSYBOX XZ, but PACKAGE_tar (gnu, have --numeric-owner option available ) also enabling PACKAGE_xz at the same time, so there is a clash when I try to compile it. Jul 11 21:11:58 ldir: disgraceful Jul 11 21:27:00 jwh ldir I know! Jul 11 21:42:26 is posix real-time librt in any way tied to kernel's RT patch? or they're totally irrelevant? Jul 11 21:43:09 i know this is a general kernel question, my understanding is that they're not related...RT patch applies to all processes, be it normal process or posix-rt-process Jul 11 21:44:24 RT makes the kernel near-full-pre-emptable for syscalls and interrupt threads, while posix-rt-process just makes sure those processes have a high priority under the scheduler Jul 11 22:44:48 https://arstechnica.com/information-technology/2018/07/year-old-router-bug-exploited-to-steal-sensitive-dod-drone-tank-documents/ Jul 11 23:02:44 lol Jul 11 23:06:19 Tapper: https://cdn.arstechnica.net/wp-content/uploads/2018/07/Screen-Shot-2018-07-11-at-5.42.23-PM-1440x978.png Jul 11 23:07:24 "The certificate for completion of a cyber-security awareness "challenge" found amongst the files stolen from an Air Force captain's computer." **** ENDING LOGGING AT Thu Jul 12 03:00:01 2018