**** BEGIN LOGGING AT Thu May 31 03:00:07 2018
May 31 04:09:37 <schm0> hmm
May 31 04:09:59 <schm0> someone else has the issue that adblock resets itself?
May 31 04:11:14 <schm0> im trying to figue out what the problem is... in the package description -> Selects: PACKAGE_libc [=y] && PACKAGE_libssp [=n] but there is no package libssp
May 31 04:11:24 <schm0> was it renamed to libgcc?
May 31 04:13:49 <schm0> when i apply new config through luci, config gets written properly, when i visit the luci adblock page it displays my settings i applied before, but when i check the config file it is all reset to default
May 31 04:13:53 <schm0> really weird
May 31 04:26:43 <blogic> aloha
May 31 04:28:57 <shm0> also quite interesting adblock fails to download from zeustracker.abuse.ch. ping from router fails, from lan client it works
May 31 04:29:04 <shm0> wtf is going on
May 31 04:29:38 <blogic> one hs v6 the other only v4 ?
May 31 04:30:13 <shm0> ipv4 only
May 31 04:30:50 <shm0> i mean both use dnsmasq how can it fail on the router itself
May 31 04:31:24 <shm0> ping gives bad address
May 31 04:32:12 <shm0> now it works again?
May 31 04:32:26 <shm0> wtf
May 31 04:32:37 <blogic> chinese internet
May 31 04:33:16 <Lantis> afternoon
May 31 04:33:48 <pkgadd> I wouldn't rule out that the list server might have temporarily blocked you for trying too often either
May 31 04:36:34 <shm0> then it just also not work from the lan side?
May 31 04:38:47 <shm0> i think its a problem on the router itself
May 31 04:43:22 <shm0> and for adblock.... it work the first time, after it is done, the config resets to default
May 31 04:56:20 <ds_shadof> luci https://imgur.com/a/28WEoXX
May 31 05:12:10 <jow> dexterji: browser?
May 31 05:12:22 <jow> oops ds_shadof which browser?
May 31 05:12:46 <jow> judging by the font rendering its likely ie
May 31 05:13:36 <ds_shadof> Google Chrome Version 58.0.3029.110 (64-bit)
May 31 05:13:53 <jow> ok
May 31 05:16:42 <ds_shadof> i like my fonts :) they a clear
May 31 05:18:07 <jow> ds_shadof: can reproduce it in chromium; will take a look
May 31 09:02:11 <lede-builder> build #849 of rb532/generic is complete: Success [build successful]  Build details are at http://phase1.builds.lede-project.org/builders/rb532%2Fgeneric/builds/849
May 31 09:11:04 <ds_shadof> Always wanted to know, how luci decides to color the interface? https://imgur.com/a/pt06Lud
May 31 09:15:15 <jow> ds_shadof: the color is derived from the firewall zone
May 31 09:15:51 <jow> https://github.com/openwrt/luci/blob/master/modules/luci-base/luasrc/model/firewall.lua#L459
May 31 09:45:08 * ldir verifies with 'sysctl kernel.panic=0 ; echo c > /poroc/sysrq-trigger' that indeed the hardware watchdog does indeed reboot the box after 30s on an MIR3G
May 31 09:46:51 <ldir> that's /proc/sysrq-trigger  - oh dear can't type or see this morning.
May 31 09:47:35 <lede-builder> build #845 of mpc85xx/p1020 is complete: Success [build successful]  Build details are at http://phase1.builds.lede-project.org/builders/mpc85xx%2Fp1020/builds/845
May 31 10:28:15 <Sphexist> Hello togeher
May 31 10:28:31 <Sphexist> Is someone reachable?
May 31 10:42:07 <lede-builder> build #874 of ar71xx/nand is complete: Success [build successful]  Build details are at http://phase1.builds.lede-project.org/builders/ar71xx%2Fnand/builds/874
May 31 11:37:28 <aparcar> dangole: ping
May 31 11:38:53 <dangole> aparcar: pong
May 31 11:41:16 <aparcar> Hi, currently the LibreMesh distribution offers various packages to simplify mesh networks. All code is Lua or Bash, but to build the packages for all distributions it requires all maintaining SDKs for all possible targets. I found the `all` aka `noarch` flag but it that's not a convenient solution.
May 31 11:41:17 <aparcar>  
May 31 11:43:03 <aparcar> Can we expand the SDK to specially treat noarch packages and add them to /packages/noarch/<repo>/? This could drastically reduce resource usage. Also usable for most Luci packages
May 31 11:45:43 <jow> why ?
May 31 11:45:59 <jow> can't you just take one random sdk and build the noarch packages with that?
May 31 11:46:09 <jow> they're noarch, so you can use them everywhere
May 31 11:46:27 <dangole> jow: yes. but how to build -noarch packages with the current tooling?
May 31 11:46:58 <dangole> jow, aparcar: maybe all it takes is properly marking the packages as 'noarch' in their OpenWrt-package Makefiles
May 31 11:47:26 <jow> PKGARCH:=all
May 31 11:47:46 <jow> in the Package/* define
May 31 11:51:54 <aparcar> jow: where will the packages end up? in a special folder for noarch packages?
May 31 11:58:27 * dwmw2_gone updates https://github.com/openwrt/openwrt/pull/940
May 31 12:02:59 <napster> anybody the
May 31 12:04:40 <jow> aparcar: no, in the normal packages folder
May 31 12:42:05 <lede-builder> build #884 of brcm2708/bcm2710 is complete: Success [build successful]  Build details are at http://phase1.builds.lede-project.org/builders/brcm2708%2Fbcm2710/builds/884
May 31 13:05:34 <Tapper> Huawei's Gao Xiang has announced the EROFS open-source Linux file-system intended for Android
May 31 13:05:40 <Tapper> https://www.phoronix.com/scan.php?page=news_item&px=EROFS-File-System
May 31 13:06:37 <Hauke> jow: I used the mbedtls benchmark application and the Lantiq VR9 500MHz MIPS BE SoC supports 1.3 RSA signatures per second and 41 ECDSA-secp256r1 signatures per second
May 31 13:06:38 <Tapper> EROFS  sounds like a good moov for android
May 31 13:06:49 <Hauke> I will try to add ECDSA support to our web server
May 31 13:06:55 <Hauke> in additon to RSA
May 31 13:07:39 <Hauke> jow: here are the benchmark results: https://pastebin.com/nfWvgJd6
May 31 13:08:34 <Tapper> with those compression numbers  could be good for openwrt I think.
May 31 13:12:26 <lede-builder> build #866 of mpc85xx/generic is complete: Success [build successful]  Build details are at http://phase1.builds.lede-project.org/builders/mpc85xx%2Fgeneric/builds/866
May 31 13:27:37 <jow> Tapper: they better focus on open sourcing there drivers. yet another file system is not waht we urgently need
May 31 13:28:01 <jow> Tapper: especially not one that sort of is better than isofs but worse than squashfs because it sacrafices compression for read speed
May 31 13:29:51 <jow> Hauke: interesting ... some impressive ecdsa speedups there from 2.1 to 2.9
May 31 13:44:27 <Tapper> jow in there email they said that the driver will be out soon, but we all know how long soon can be! :-)
May 31 13:46:24 <lede-builder> build #853 of ar7/generic is complete: Success [build successful]  Build details are at http://phase1.builds.lede-project.org/builders/ar7%2Fgeneric/builds/853
May 31 14:03:49 <Hauke> jow: yes
May 31 14:04:07 <Hauke> jow: ECDSA is much faster then RSA now
May 31 14:04:26 <Hauke> primary I wanted to check if the low RSA speed was a regression
May 31 14:05:01 <Hauke> jow: to add support for multiple certs/keys I have to change the ustream-ssl API, is this a problem?
May 31 14:05:11 <Hauke> how backwards compatible should I be?
May 31 14:05:32 <jow> Hauke: maybe consider _v2 or _ex suffixed new functions
May 31 14:05:55 <jow> abi break would be not ideal
May 31 14:05:59 <Hauke> ok
May 31 14:06:14 <jow> depends on the impact of the changes you have to do
May 31 14:06:33 <Hauke> I have to provide a secound cert + key pair
May 31 14:06:54 <Hauke> the problem is that we currently have one function to give a cert and one to give a key
May 31 14:07:01 <Hauke> I would add a new function to give the full pair
May 31 14:07:14 <Hauke> this would then we used by uhttpd
May 31 14:07:20 <jow> that would work
May 31 14:07:59 <Hauke> ok
May 31 14:42:32 <lede-builder> build #875 of ar71xx/mikrotik is complete: Success [build successful]  Build details are at http://phase1.builds.lede-project.org/builders/ar71xx%2Fmikrotik/builds/875
May 31 15:50:26 <paulius> hi. i'm building 4.14 for archer c7 v4 (ar71xx), and apparently the kernel is too big. can the limit be adjusted somehow? or must I abide by what is set?
May 31 15:54:52 <lede-builder> build #786 of ar71xx/generic is complete: Success [build successful]  Build details are at http://phase1.builds.lede-project.org/builders/ar71xx%2Fgeneric/builds/786
May 31 16:17:12 <ausjke> shm0: why ubound? dnsmasq seems solid and widely used and works well, i use dnsmasq-full myself
May 31 16:28:43 <ultito> morning
May 31 16:44:02 <KevinDB> morning
May 31 17:11:56 <ultito> 18.06 will be released tomorrow?
May 31 17:16:51 <rotanid> huh?
May 31 17:17:03 <rotanid> where does it say that?
May 31 17:33:51 <lede-builder> build #890 of brcm2708/bcm2709 is complete: Success [build successful]  Build details are at http://phase1.builds.lede-project.org/builders/brcm2708%2Fbcm2709/builds/890
May 31 17:53:41 <Borromini> hi
May 31 17:58:17 <Hauke> ultito: 18.06 is planned for sometime in june
May 31 18:04:55 <Borromini> ultito: maybe you should install those snapshots huh :)
May 31 18:25:35 <Borromini> ldir: ping
May 31 18:26:47 <ultito> Borromini: never, never, never :-)
May 31 18:26:56 <Borromini> hehehe
May 31 18:27:00 <ultito> I will wait for regural release
May 31 18:27:05 <Borromini> humming along just fine here so far
May 31 18:27:43 <ultito> :-)
May 31 18:28:24 <dangole> blogic: i finally got oxnas to boot (almost) vanilla 4.14
May 31 18:29:24 <paulius> latest git doesn't build on ar71xx
May 31 18:29:30 <paulius> target/linux/generic/backport-4.9/095-v4.12-ipv6-Need-to-export-ipv6_push_frag_opts-for-tunnelin.patch needs to be removed
May 31 18:30:05 <paulius> the changes are apparently already there
May 31 18:30:20 <Borromini> paulius: did you bump your kernel locally?
May 31 18:30:24 <paulius> no
May 31 18:30:27 <Borromini> weird.
May 31 18:30:44 <Borromini> then somebody did a bad kernel bump, if you say the code is already in the kernel.
May 31 18:31:46 <paulius> don't take my word for it, please check, but all I did was git clone, make menuconfig and make
May 31 18:39:43 <Borromini> paulius: master,
May 31 18:39:44 <Borromini> ?
May 31 18:40:11 <paulius> Borromini: openwrt/openwrt.git
May 31 18:40:21 <Borromini> paulius: the master branch?
May 31 18:40:24 <paulius> yes
May 31 18:46:18 <lede-builder> build #8 of mediatek/mt7623 is complete: Failure [failed images]  Build details are at http://phase1.builds.lede-project.org/builders/mediatek%2Fmt7623/builds/8  blamelist: Hans Dedecker <dedeckeh@gmail.com>, Daniel Golle <daniel@makrotopia.org>, Jason A. Donenfeld <Jason@zx2c4.com>
May 31 18:46:30 <ldir> Borromini: pong
May 31 18:46:55 <Borromini> paulius: yop, it breaks
May 31 18:47:02 <Borromini> ldir: wondering if you were bumping any kernals.
May 31 18:47:56 <ldir> running 4.9.105 for the past 15 minutes
May 31 18:48:34 <Borromini> ok
May 31 18:48:48 * Borromini rummages through ldir's trees
May 31 18:49:41 <ldir> https://github.com/ldir-EDB0/openwrt/commit/0e714d6fb46416adddefcca441cdf69cfcaa9b98
May 31 18:49:51 * Borromini tips his hat
May 31 18:52:19 <Borromini> will use that for 18.06 :)
May 31 18:54:09 <ldir> use it in any way you see fit :-)
May 31 18:54:19 <Borromini> good lord
May 31 18:54:40 <Borromini> i wonder if anyone was ever murdered with a kernel patch :^)
May 31 18:59:10 <Borromini> paulius: it looks like ldir's patch addresses it, if you're interested
May 31 19:03:16 <paulius> Borromini: thanks :)
May 31 19:06:32 <Borromini> thank ldir :)
May 31 19:08:37 <ldir> I was amused/surprised at how quickly 105 followed 104
May 31 19:09:43 <turova> blogic, I talked to you a little while ago about implementing hnat on a c7. Currently working on porting it over. One thing I wanted to do was try your work on a device you tested on, but couldn't find any ipq806x + qca833 hardware
May 31 19:09:55 <turova> what device did you use to test your hnat code?
May 31 19:10:21 <mamarley> ldir: If you think that is fast, have a look at 4.14.45, 4.14.46, and 4.14.47. :P
May 31 19:10:45 <lede-builder> build #249 of ar71xx/tiny is complete: Failure [failed toolchain]  Build details are at http://phase1.builds.lede-project.org/builders/ar71xx%2Ftiny/builds/249  blamelist: Hans Dedecker <dedeckeh@gmail.com>, Jason A. Donenfeld <Jason@zx2c4.com>
May 31 19:12:44 <ldir> ^^^^^ caused by daa73b63d5
May 31 19:14:24 <Tapper> Hi is the thing with too many active connections when using ofloading fixt yet?
May 31 19:20:06 <Tapper> The was a email to the list about, "18.06 bug: Flow Offload Active Connections"
May 31 19:32:10 <Borromini> darn it
May 31 19:32:15 <Borromini> 4.14 breaking as well of course
May 31 19:32:49 <mamarley> Breaking what?
May 31 19:33:01 <Borromini> a kernel bump
May 31 19:33:10 <Borromini> so nothing in-tree so far :)
May 31 19:45:07 <pkgadd> paulius: to me it looks as if you could resize the partitioning easily, by changing target/linux/ar71xx/image/generic-tp-link.mk and tools/firmware-utils/src/tplink-safeloader.c - either by changing the split between kernel and rootfs or by combining both into a single "firmware" partition, making use of MTD_SPLIT_TPLINK_FW, there is precedent for both in the git history, but you should have a
May 31 19:45:13 <pkgadd> robust/ tested recovery mechanism before testing this
May 31 19:45:41 <pkgadd> paulius: do you happen to have the patches for bumping ar71xx to 4.14 around?
May 31 19:46:53 <paulius> pkgadd: currently it's in the form of modified files
May 31 19:47:26 <pkgadd> paulius: this https://git.openwrt.org/?p=openwrt/openwrt.git;a=commitdiff;h=b72b36653a3fc8347456ab9c45d26a3144688a4c would be an example for the conservative approach of just changing the split between kernel and rootfs, but there've also been examples to migrate it to a common "firmware" partition
May 31 19:48:19 <pkgadd> o.k., instead of changing the dts, you'd have to modifiy MTDPARTS in target/linux/ar71xx/image/generic-tp-link.mk
May 31 19:48:28 <paulius> pkgadd: my 4.14 hacks work on a tl-wr1043nd v1, but i got crashes on an archer c7. need to investigate
May 31 19:49:05 <paulius> pkgadd: do you want it?
May 31 19:49:31 <paulius> pkgadd: thanks for the info, i'll try
May 31 19:50:38 <pkgadd> paulius: if you can provide it easily, I'd like to look into it (ar71xx is the only target I care about which is still on kernel 4.9, while ath79 isn't quite there for my devices so far - I've started rebasing it to 4.14 myself, but ran out of time in the middle of the mdio (switch) changes)
May 31 19:51:06 <paulius> ah, the phy_addr -> mdio_addr thing? :)
May 31 19:51:50 <pkgadd> yep
May 31 19:51:51 <paulius> the reason it works on the tl-wr1043nd v1 is because it lacks mdio
May 31 19:51:54 <paulius> heh
May 31 19:52:01 <pkgadd> ;)
May 31 19:52:33 <paulius> I just tried changing phy_addr to mdio_addr in the mach file. that's probably why it didnt work
May 31 19:54:06 <paulius> the new mdio_board_info function seems to works in a similar way, but mycode reading skills are limited
May 31 19:56:23 <paulius> how do I diff my locally modified files against git master? :p
May 31 19:56:32 <Borromini> git diff
May 31 19:56:48 <Borromini> well, you add them as a commit, then git format-patch -1
May 31 19:57:06 <Borromini> but if you just want to see the differences git diff works, but i guess that's not what you're after
May 31 20:00:01 <paulius> Borromini: git diff doesn't seem to show new files I created
May 31 20:00:43 <paulius> nor the ones i removed
May 31 20:01:46 <paulius> read: i need to learn git
May 31 20:02:21 <Borromini> git status will show you what's not tracked and what is
May 31 20:02:32 <Borromini> you need to git add the new stuff and anything you changed/deleted
May 31 20:02:40 <Borromini> then git commit -m "your message here"
May 31 20:02:45 <Borromini> then you can generate a patch
May 31 20:06:08 <paulius> I see. I'll do that a bit later then and provide a diff. But it's basically just enough hacks toget it to compile and run on a tl-wr1043nd v1.
May 31 20:06:25 <paulius> 1 month uptime now, though
May 31 20:07:12 <paulius> pkgadd: if you figure out the mdio thing do ping me :)
May 31 20:09:08 <pkgadd> will do ;)
May 31 20:13:01 <pkgadd> paulius: this https://git.openwrt.org/?p=openwrt/openwrt.git;a=commitdiff;h=55c77b3d3c7a85a2bbf891437e0f6aee4021fc96 is an example to move from dedicated kernel+rootfs to mtdsplit
May 31 20:13:44 <pkgadd> or this https://git.openwrt.org/?p=openwrt/openwrt.git;a=commitdiff;h=ca6002c0efab2e847be91b13d558e5c38184b50a
May 31 20:17:08 <Borromini> maybe a stupid question, but: does mtdsplit allow to resize kernel partitions 'dynamically'? as in: you can flash another image, it won't break things if the kernel partition is e.g. bigger on the new image?
May 31 20:17:57 <pkgadd> Borromini: yes (padded to erase block sizes, so 64 KB on most devices), but this only works on devices using spi-nor, not NAND flash
May 31 20:18:11 <Borromini> ok and old devices have nor in general no
May 31 20:19:19 <pkgadd> devices up to 32 MB generally do, NAND usually starts in the 128 MB range
May 31 20:19:53 <pkgadd> but you can't avoid testing on the actual device if it actually works
May 31 20:23:06 <lede-builder> build #874 of ixp4xx/generic is complete: Failure [failed toolchain]  Build details are at http://phase1.builds.lede-project.org/builders/ixp4xx%2Fgeneric/builds/874  blamelist: Hans Dedecker <dedeckeh@gmail.com>, Jason A. Donenfeld <Jason@zx2c4.com>
May 31 20:26:11 <huaracheguarache> pkgadd: hi, I saw that you mentioned the potential issues associated with removing the netgear partition to make some space. Do you know if removing it would somehow affect the firmware upgrade process in OEM firmware when flashing to OpenWrt (factory image)?
May 31 20:29:31 <Borromini> would only affect it the other way around i reckon
May 31 20:29:58 <huaracheguarache> I guess I should try it out
May 31 20:30:23 <Borromini> :)
May 31 20:32:00 <pkgadd> huaracheguarache: changing the partitioning requires tftp, that's all - the really interesting question is, what is this netgear partition used for in the OEM firmware, what data does it contain - and can it be recreated, if lost (by repurposing the partition)
May 31 20:33:05 <huaracheguarache> pkgadd: my bet is on streamboost
May 31 20:36:03 <pkgadd> huaracheguarache: quite possibly (my nbg6817 reserves 3.2 GB for streamboost (no real idea what that is/ why it needs storage) and minidlna)
May 31 20:38:56 <huaracheguarache> pkgadd: it's basically cake with some datamining: https://www.smallnetbuilder.com/images/stories/lansrouters/streamboost/streamboost_optin.jpg
May 31 20:40:14 <pkgadd> 'nice'
May 31 20:41:28 <pkgadd> that would make it not too difficult to reformat/ reinitialize that partition from the OEM firmware, but no idea if the OEM firmware has that capability
May 31 20:42:01 <huaracheguarache> I'll check tomorrow
May 31 20:42:53 <lede-builder> build #478 of layerscape/armv8_32b is complete: Failure [failed toolchain]  Build details are at http://phase1.builds.lede-project.org/builders/layerscape%2Farmv8_32b/builds/478  blamelist: Hans Dedecker <dedeckeh@gmail.com>, Jason A. Donenfeld <Jason@zx2c4.com>
May 31 20:43:15 <pkgadd> make sure to make a backup before, check what's inside (perhaps it can be mounted) and compare it to whatever (migh) gets recreated
May 31 20:44:43 <huaracheguarache> what tools should I use to make a backup?
May 31 20:45:38 <pkgadd> hmm, I think nanddump might be needed
May 31 20:45:42 <pkgadd> and cat
May 31 20:45:57 <pkgadd> e.g. that's my streamboost partition http://paste.debian.net/1027502/
May 31 20:49:07 <huaracheguarache> ok, thanks. I'll have a look
May 31 21:01:47 <ausjke> https://zerobin.net/?e9a8303bd3d436cd#QPe5xHc/5IJ60Zzj046oe5IBPR6yBsm+PvPRSGWpKHQ=
May 31 21:02:15 <ausjke> zbt1326 still shows one radio for wifi after the wifi-fix-patch was git-pulled
May 31 21:15:14 <ausjke> https://zerobin.net/?a95f21ee858577ba#3LWElOLdvaz06VgtPce/ml4azE7/hUIspAo8ya+JvFo=  dmesg showed mt7603e failed to run, i.e. the missing 2.4Ghz radio, iw list also showed only phy1
May 31 21:34:33 <ausjke> that mt76 fix is not real the fix, ignore above msgs
May 31 22:00:34 <lede-builder> build #877 of brcm63xx/generic is complete: Failure [failed toolchain]  Build details are at http://phase1.builds.lede-project.org/builders/brcm63xx%2Fgeneric/builds/877  blamelist: Hans Dedecker <dedeckeh@gmail.com>, Jason A. Donenfeld <Jason@zx2c4.com>
May 31 22:20:07 <ausjke> hold on, are you comparing to bsd's pf filter?
May 31 22:31:10 <schm0> hmm
May 31 22:31:55 <schm0> somehow my custom feed images are not ending up in the final image anymore. the packages get build but are not included in the final image
May 31 22:35:34 <schm0> -images +packages
May 31 22:44:03 <abenz> jow: I have a local copy of the feeds repos, I need to switch from 18.06 and trunk from time to time.. when I append ;18.06 at end of each entry (in feeds.conf), it doesn't work for a local repo (only seems to work if its remote repo)
May 31 22:44:12 <lede-builder> build #892 of ar7/ac49x is complete: Failure [failed toolchain]  Build details are at http://phase1.builds.lede-project.org/builders/ar7%2Fac49x/builds/892  blamelist: Hans Dedecker <dedeckeh@gmail.com>, Jason A. Donenfeld <Jason@zx2c4.com>
May 31 22:44:37 <abenz> is there a way I can specify a different branch or do I have to manually cd into the feed dir and checkout either trunk or 18.06 branch ?
May 31 23:33:27 * mangix hates pcie with a passion
May 31 23:45:39 <mangix> ugh I give up. my head hurts.
May 31 23:49:39 <shm0> does someone here use stubby with dnssec? i can't get it working anymore.
May 31 23:51:09 <shm0> it fails now to create the appdata_dir folder, stubby -i gives no path. tried to create manually, stubby -i shows path now but doesnt download anything
Jun 01 00:00:33 <shm0> nvm got it working
Jun 01 00:00:55 <shm0> somehow its a permission issue now. but worked yesterday
Jun 01 00:01:00 <shm0> interesting again
Jun 01 00:02:02 <Lantis> shm0: have you moved away from Gargoyle these days, or are you still using some of the builds?
Jun 01 00:02:24 <shm0> moved from it long time ago^^
Jun 01 00:03:54 <shm0> but was also nice. i liked the autorate for ingress qos. but my isp seems to limit icmp
Jun 01 02:32:04 <mangix> karlp: ping
**** ENDING LOGGING AT Fri Jun 01 03:00:09 2018