**** BEGIN LOGGING AT Thu Jan 24 02:59:56 2019 Jan 24 08:53:32 <\x> got my d2 Jan 24 08:53:33 <\x> https://my.mixtape.moe/qfpzko.png Jan 24 08:58:49 jow:ping Jan 24 09:00:02 dedeckeh: pong Jan 24 09:27:15 ldir: ping Jan 24 09:27:29 xback: aye up :-) Jan 24 09:29:27 ldir: just noticed this patch: https://git.openwrt.org/?p=openwrt/staging/ldir.git;a=commit;h=18b17b31b542a2a62afab0abf9927f711273d44d Jan 24 09:29:56 I can confirm this behaviour on rocket-M5 devices. which bugged us a lot Jan 24 09:30:07 i'll give it a good spin in the field here Jan 24 09:31:04 so it can get pushed to master Jan 24 09:31:43 there's a github pr with it in and I wanted to check it didn't break my archer first - which it hasn't :-) Jan 24 09:32:21 do you want me to inform you regarding my findings? or can I take the patch? Jan 24 09:35:00 oh just take it Jan 24 09:35:18 I can't find the PR I stumbled across it in now. Jan 24 09:35:25 i've went through the driver code, and would combine the patch with a few other cleanups Jan 24 09:36:58 this issue could also explain the luci network config issues people reported in the forum Jan 24 09:37:55 https://github.com/ynezz/openwrt/commit/7e3f6878b08e174d29d7d63e624bd311cd14aee2 Jan 24 09:42:36 It's related to http://lists.infradead.org/pipermail/openwrt-devel/2019-January/015387.html - which looks like 'fun' Jan 24 09:43:26 there is also a PR for that in github Jan 24 09:43:59 I noticed the ag71xx network driver is written very similar to the cns3xxx one Jan 24 09:48:29 ldir: https://github.com/openwrt/openwrt/pull/1735 Jan 24 09:49:57 well tracked down Sir - sorry I'm somewhat out of the flow of things at the moment Jan 24 09:50:55 looks like it should be at least tested on your 'active device farm' Jan 24 09:51:06 ldir: still building ikea furniture? ;-) Jan 24 09:52:21 lol - trying to focus a bit more on the thing that earns me money & family Jan 24 09:53:24 and also to be honest, things that I have at least some 'skill' in :-) I keep an eye on dnsmasq of course! Jan 24 09:53:42 I must ask stintel if his static lease issue is still fixed. Jan 24 09:54:28 ldir: I'm testing my ar8216 PR these days. I'll probably start testing PR#1735 on my QCA routers tomorrow. Jan 24 09:55:36 BTW could someone tell me how should I test the port mirroring feature? I don't even know what that's used for. Jan 24 09:57:34 gch981213: In ignorance of what happens to the switch config file in /etc/config I use luci to set up the mirroring. Useful to mirror the WAN port to another spare port to which you connect your mac and run wireshark to see what exactly is going out to/from the Internet :-) Jan 24 09:58:42 Got it Jan 24 09:58:59 ldir: Thanks :) Jan 24 10:01:54 no problem - it's not an exact mirror in that vlan tags get put on the mirrored packets even though the real port doesn't send them - this may be a bug...or maybe it's a symptom of how the switch internally tags packets from ports....I don't know. Jan 24 10:48:37 ldir: I'm travelling so can't test with my laptop Jan 24 10:49:17 ldir: but my printer still appeared as available in my monitoring system, which means it's still on the static lease Jan 24 11:02:23 stintel: oooh good! Jan 24 11:07:32 very! Jan 24 11:24:56 <\x> uhm Jan 24 11:24:57 <\x> hello Jan 24 11:25:09 <\x> how do i regrow /tmp Jan 24 11:25:12 <\x> https://oldwiki.archive.openwrt.org/doc/uci/fstab Jan 24 11:25:22 <\x> this one only regrows /tmp/shm Jan 24 11:26:06 mount /tmp -o remount,size=whatever Jan 24 11:26:22 default is dynamic allocation up to 50% of system memory Jan 24 11:26:33 iirc Jan 24 11:26:35 <\x> root@D2:~# mount /tmp -o remount,size=128M Jan 24 11:26:35 <\x> mount: mounting /dev/zram0 on /tmp failed: Invalid argument Jan 24 11:26:41 ah zram, no idea Jan 24 11:27:05 <\x> https://termbin.com/37oz Jan 24 11:27:14 <\x> the thing is it grows /tmp/shm instead Jan 24 11:27:16 <\x> weird Jan 24 11:27:27 <\x> i dont know if i edit /etc/fstab itself Jan 24 11:28:45 14 local zram_size="$( uci -q get system.@system[0].zram_size_mb )" Jan 24 11:29:25 so that would be option zram_size "..." in /etc/config/system Jan 24 11:29:33 <\x> # uci -q get system.@system[0].zram_size_mb Jan 24 11:29:33 <\x> 256 Jan 24 11:30:20 <\x> i do have 512MB of ram, so i guess half of that is the optimal zram. ye ye, i know 0 zram is the best Jan 24 13:11:41 build #752 of layerscape/armv8_64b is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/layerscape%2Farmv8_64b/builds/752 Jan 24 13:56:40 both the nightly image and my own compilation of trunk fails to boot on WDR 4900 v1. I get "Bad data CRC" when it's trying to verify the checksum on the kernel at boot. https://pastebin.com/HxttQPKN 18.06.1 release image works properly Jan 24 13:57:27 openwrt-mpc85xx-generic-tl-wdr4900-v1-squashfs-factory.bin downloaded just now didn't boot Jan 24 13:57:53 penwrt-18.06.1-mpc85xx-generic-tl-wdr4900-v1-squashfs-factory.bin works as expected. Jan 24 13:58:48 SwedeMike: break into u-boot and check the boot command copying the kernel into memory, it likely will be less than "Data Size: 2900991 Bytes = 2.8 MiB" (so the kernel grew larger than what u-boot copies) Jan 24 14:05:48 spiboot=sf probe 0;sf read 2000000 60200 29f000;sf read 3000000 50000 10000;bootm 2000000 - 3000000 Jan 24 14:09:03 right, changing 29f000 to 39f000 worked Jan 24 14:09:31 unfortunately there is no "saveenv" command in this uboot. Jan 24 14:10:49 a proper solution would then probably be a second stage bootloader/u-boot Jan 24 14:11:55 https://pastebin.com/EWvWrpCF shows the default env settings Jan 24 14:11:58 unfortunately a common issue found on multiple devices Jan 24 14:13:09 can the uboot env variables be set from within openwrt after boot? Jan 24 14:14:42 yes, there is a pkg for that Jan 24 14:16:34 uboot-envtools Jan 24 14:27:01 at what point is /lib/uboot-envtools.sh run? Jan 24 14:53:14 rmilecki: ping Jan 24 14:53:24 ldir: pong Jan 24 14:54:49 ldir: ever seen this? [709739.855248] ltq_etop 1e180000.etop eth0: tx ring full Jan 24 14:55:26 i remember you also had a netgear dgn3500 at one point Jan 24 14:57:25 DonkeyHotei: I did/do, but it's in a box unused and no realistic way of plugging it up to use it to see if it gets the error. Jan 24 14:58:55 yeah, mine was unused for a bit, but i deployed it at the other place i spend a lot of time, with a recent 18.06 snapshot, and the above happens whenever spotify is in use, breaking all connectivity Jan 24 15:00:34 i'm expecting to go back there monday, and i could do any requested testing then Jan 24 15:01:15 no idea Jan 24 15:02:21 not sure who's dealing with that driver lately Jan 24 16:47:54 just throwing in a glib2 call error i get, maybe someone can help: Jan 24 16:47:54 a glib2 call "g_convert()" cant convert from “UTF8” to “UTF16LE” Jan 24 16:47:54 So where should the UTF16LE support come from? Can i enable this somehow? Jan 24 16:49:30 The error message i get is "Could not open converter from “UTF8” to “UTF16LE”" Jan 24 16:50:11 andy2244: you need to build with full nls support Jan 24 16:50:35 otherwise glib2 gets linked against a stub iconv which only handlers a limited number of one-byte charsets + utf-8 Jan 24 16:51:02 is this a package somewhere? Jan 24 16:51:23 or a option ? Jan 24 16:54:01 ah ok i see: CONFIG_BUILD_NLS Jan 24 16:54:02 andy2244: second... Jan 24 16:54:07 right, that Jan 24 16:54:14 thanks Jan 24 16:54:26 afterwards make pacakge/glib2/{clean,compile} V=s Jan 24 17:21:09 mhh now i get "cannot find -lintl" at the linking stage. Jan 24 17:21:10 The lib is build under /usr/lib/libintl-full, but i see no link to /usr/lib or an extra -L search option in the build process? How should this work normally? Jan 24 17:24:40 ok lol... i should probably add "include $(INCLUDE_DIR)/nls.mk" in my package :p Jan 24 17:48:03 right, that should take care of it Jan 24 17:48:26 you'll also need DEPENDS:=$(ICONV_DEPENDS) Jan 24 18:04:42 mhh still getting the same error: "Could not open converter from “UTF8” to “UTF16LE”" Jan 24 18:04:42 But the build is using the full headers/libs from libintl-full, any other idea? Jan 24 18:10:27 is ldd showing linkage to libiconv.so ? Jan 24 18:12:04 its a static binary Jan 24 18:12:22 ah you mean glib Jan 24 18:13:27 no wait, the binary is static so glib2 is linked into it? Jan 24 18:14:49 maybe you need to tweak include/nls.mk and add -lcharset as well Jan 24 18:15:35 hm no, thats unrelated Jan 24 18:17:03 is "strings /your/binary | grep ISO-10646-UCS-2" returning a match? Jan 24 18:17:20 ok was on wrong terminal: Jan 24 18:17:20 ldd /usr/sbin/cifsadmin Jan 24 18:17:20 /lib/ld-musl-armhf.so.1 (0xb6f36000) Jan 24 18:17:20 libcifsdtools.so.0 => /usr/lib/libcifsdtools.so.0 (0xb6f1f000) Jan 24 18:17:20 libglib-2.0.so.0 => /usr/lib/libglib-2.0.so.0 (0xb6e22000) Jan 24 18:17:20 libintl.so.8 => /usr/lib/libintl.so.8 (0xb6e09000) Jan 24 18:17:21 libnl-genl-3.so.200 => /usr/lib/libnl-genl-3.so.200 (0xb6df4000) Jan 24 18:17:21 libnl-3.so.200 => /usr/lib/libnl-3.so.200 (0xb6dd0000) Jan 24 18:17:22 libgcc_s.so.1 => /lib/libgcc_s.so.1 (0xb6db5000) Jan 24 18:17:22 libc.so => /lib/ld-musl-armhf.so.1 (0xb6f36000) Jan 24 18:17:23 libiconv.so.2 => /usr/lib/libiconv.so.2 (0xb6d9e000) Jan 24 18:17:32 looks correct Jan 24 18:17:49 thats definitely the full iconv (the stub one is always static) Jan 24 18:18:27 next step would be strace, see if iconv maybe requires some data files not being present Jan 24 18:21:48 ok will try this or experiment with some code changes Jan 24 19:17:23 hi Jan 24 19:17:30 *yawn* Jan 24 19:17:44 * blogic is fighting with a 3.18 kernel provided by a customer Jan 24 19:17:49 bollocks shit crap Jan 24 19:17:55 :-D Jan 24 19:25:22 anyone familiar with Enable NAT Loopback? Jan 24 19:26:43 build #280 of gemini/generic is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/gemini%2Fgeneric/builds/280 Jan 24 19:42:47 build #1173 of ipq806x/generic is complete: Failure [failed images] Build details are at http://phase1.builds.lede-project.org/builders/ipq806x%2Fgeneric/builds/1173 blamelist: Koen Vandeputte , Jo-Philipp Wich , Christian Lamparter , Petr ?tetiar Jan 24 20:51:26 xback: are you there? Jan 24 20:54:24 nvm Jan 24 20:57:32 <\x> weird Jan 24 20:57:34 <\x> hmmm Jan 24 20:57:50 <\x> my sysupgrade is failing everytime Jan 24 20:57:54 <\x> https://pomf.pyonpyon.moe/tuhwok.png Jan 24 20:58:14 <\x> why is /dev/zram0 mounted as /tmp and why is it only 16MB? Jan 24 20:59:52 that's an option in menuconfig Jan 24 21:00:41 lach1012: if it's about his last terrible kernel build breakage, he is aware already https://github.com/openwrt/openwrt/commit/528508ae8b33c76deaf18f313973144a7e920678#commitcomment-32049736 :) Jan 24 21:00:50 <\x> where DonkeyHotei Jan 24 21:00:57 ynezz, yes I saw Jan 24 21:01:30 could someone please help, i think this is a bug (or a massive understanding on my part) Jan 24 21:01:53 is anyone familiar with the "Enable NAT Loopback" setting? Jan 24 21:12:43 malwar3hun73r: I am Jan 24 21:14:35 need to run, bbl Jan 24 21:24:16 please ping me when you're around Jan 24 21:24:18 i think i found something interesting Jan 24 21:24:21 jow, ^ Jan 24 22:01:35 Hey guys. I'm trying to install python3-light into a destination I've setup on an mmc card and it's not going too well. It works for other packages, but this one keeps ignoring my --dest and trying to put it on root instead. opkg --noaction --dest mydest install python3-light Jan 24 22:02:41 Anybody know why it would be doing this? Jan 24 22:02:54 oh yeah, I'm using 18.06.1 **** ENDING LOGGING AT Fri Jan 25 02:59:58 2019