**** BEGIN LOGGING AT Mon Dec 03 03:00:02 2018 Dec 03 06:13:31 *yawn* Dec 03 06:52:44 the packages repo mast is on github right ? Dec 03 06:54:16 rmilecki: kind reminder -> https://patchwork.ozlabs.org/patch/980504/ Dec 03 06:54:39 blogic: thanks, today Dec 03 06:55:46 no rush Dec 03 07:00:10 rmilecki: those block changes all look good Dec 03 07:00:28 blogic: thanks for looking at them Dec 03 07:00:30 they are quite intrusive and i am always a bit wary of these as block has some implicit logic at places Dec 03 07:00:31 i'll push them Dec 03 07:00:35 yeah Dec 03 07:00:49 i was going to suggest that and then we just respond to possible fall out quickly Dec 03 07:00:56 that first patchset (4 patches) should not change a thing Dec 03 07:01:04 let me have a second look again in a sec Dec 03 07:01:08 sure Dec 03 08:58:08 Hi, I have problems with some (not all) applications when using r10k-ct on zyxel nbg6817 (ipq806x) Dec 03 08:58:54 the last couple of times I just switched to non -ct r10k mod but now it seems I can no longer install non -ct firmware anymore!? Dec 03 09:14:20 build #999 of malta/be is complete: Failure [failed targetupload] Build details are at http://phase1.builds.lede-project.org/builders/malta%2Fbe/builds/999 blamelist: Russell Senior , Tomasz Maciej Nowak , Eneas U de Queiroz Dec 03 09:39:09 blogic: did you take a second look at block patches? Dec 03 10:08:35 build #663 of layerscape/armv8_64b is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/layerscape%2Farmv8_64b/builds/663 Dec 03 10:26:51 i've slightly changed the way OpenWrt's partitioning works to create /subpartitions/ Dec 03 10:27:12 if I broke anything for anyone, ping me! Dec 03 11:12:07 oh, I think I broke something actually... Dec 03 11:28:33 so after waiting for a bit I saw this: Dec 03 11:28:34 [ 225.648074] jffs2: Erase at 0x01ac0000 failed immediately: errno -22 Dec 03 11:28:47 not sure if that's my device failing or is that a result of my recent changes Dec 03 11:31:00 KanjiMonster: will check the imagebuilder later today. but if it breaks, it breaks for ar71xx and ath79 as well Dec 03 11:31:57 mkresin: imagebuilder has no toolchain, so it can't compile anything; assuming building relocate needs it Dec 03 11:55:44 build #377 of mvebu/cortexa9 is complete: Failure [failed targetupload] Build details are at http://phase1.builds.lede-project.org/builders/mvebu%2Fcortexa9/builds/377 blamelist: Rafa? Mi?ecki , Russell Senior , Tomasz Maciej Nowak , Florian Eckert , Eneas U de Queiroz Dec 03 11:55:45

Dec 03 12:28:34 blogic: Did you get a chance to ping MT about the rest of the padconf docs? Dec 03 12:46:08 KanjiMonster: Hi, could you please take a look at https://github.com/openwrt/openwrt/pull/750 ? Dec 03 13:44:29 I have to share this - it made me smile - https://www.youtube.com/watch?v=hjjydz40rNI Dec 03 13:51:48 Monkeh: indeed i did not Dec 03 13:53:05 Monkeh: but now i did Dec 03 13:55:21 I'm just going through the oldest PR requests on GitHub, and so far it seems to me, that nobody wants to merge PRs for new devices to ar71xx anymore, so I'm wondering if this "ath79 first" could be made explicit? Dec 03 13:55:46 ynezz: send the question to the -adm list Dec 03 13:55:51 ok Dec 03 13:57:36 What is the target for in-tree kernel modules? For example if I want to build nf-conntrack-netlink what is the target? My understanding is that make KernelPackage/nf-conntrack-netlink is for out-of-tree kernel modules. And anything inside kernel/linux/modules is for in-tree kernel modules. So how would I build for the latter? Or is it just a case of building the kernel with make target/install? Dec 03 13:57:56 blogic: Thanks Dec 03 13:58:20 Writing a driver for all that is likely to be an adventure, but.. :) Dec 03 13:58:24 make target/compile will create the ko files Dec 03 13:58:42 make package/linux/install will package them Dec 03 13:58:46 bleh, pulleda bunch, ran make menuconfig, exited, still ran into "make oldconfig" type of errors :) Dec 03 13:59:04 Monkeh: had to make it sound profitable Dec 03 13:59:11 heh Dec 03 13:59:22 Appeal to their wallets, not their brains Dec 03 13:59:41 Thanks blogic Dec 03 13:59:50 Monkeh: have you worked in this industry before ? :-) Dec 03 14:00:06 blogic: More a question of 'have you lived on this planet long', surely Dec 03 14:00:25 true its rather generic Dec 03 14:00:39 while talking about wallet Dec 03 14:00:46 i wanted to get some abstract art prints Dec 03 14:00:59 and normally an art print is like 30-40 euros Dec 03 14:01:12 somehow if you go abstract its 100-150 Dec 03 14:01:15 whyever Dec 03 14:01:31 Go to China, anything you want is 5. :P Dec 03 14:02:44 i wanted decent quality Dec 03 14:10:56 does this config symbol problem ring a bell for anyone: https://zerobin.net/?8e93ed54d055356f#Kb0LDUqdc5DF41Yv43Rbzgoy1QnmKUo/nLkqPDEPHDM= Dec 03 14:11:12 I've not built master in a while, but pulled, ran make dirclean, make menuconfig, first, Dec 03 14:27:02 do we really need MTD_ERASE_PARTIAL? Dec 03 14:27:07 does any target really use it? Dec 03 14:29:24 I think we always align rootfs_data Dec 03 14:29:33 so unaligned access support should not be needed Dec 03 14:33:13 karlp: not for me - I pull master *all* the time. Dec 03 14:34:37 * ldir must stop being distracted by computing stuff and get on with doing bookwork & invoicing Dec 03 14:35:35 rmilecki: I thought it is used on ath25 Dec 03 14:37:44 jow: any chance you understand it? Dec 03 14:37:50 jow: i thought I'm familiar with mtd Dec 03 14:37:59 but it appears i don't, or that is just buggy Dec 03 14:38:20 411-mtd-partial_eraseblock_write.patch adds something like if (mtd->flags & MTD_ERASE_PARTIAL) Dec 03 14:38:31 that triggers special part_erase() code path for every call Dec 03 14:38:44 no matter if address is aligned or not... Dec 03 14:38:48 is that correct at all? Dec 03 14:40:03 rmilecki: I vaguely recall this partial mtdblock erasing as something that was introduced for the combined sysupgrade support of the legacy ath25 target Dec 03 14:40:45 since there was no common firmware partition on this devices, we needed a way to write cleanly across the partition boundary Dec 03 14:41:03 uh Dec 03 14:41:07 Seriously? Does someone really using those those ancient AR231x chips nowadays? Dec 03 14:41:16 that will be hard to get it sorted out without a hw Dec 03 14:41:50 gch981213[w]: how does that matter to the discussion at hand? Dec 03 14:41:59 ath25 is uspported and uses this code Dec 03 14:42:07 thats all I am saying Dec 03 14:42:34 Just out of curiosity :) Dec 03 14:42:47 jow: do you have some ath25 device? Dec 03 14:43:02 I'm surprised that this target is still maintained Dec 03 14:43:46 Hey jow, the external toolchain is working great! I need to build kernel modules because certain packages depend on them. I'm currently using external kernel tree. I copied the ARM realview target and created a new one. And then copied .config from my targets kernel to target/linux/myrealviewcopy/config-default. I'm not sure if that was the right thing to do or not. But my current issue is that it doesn't appear Dec 03 14:43:46 to build the OpenWRT in tree modules (e.g. netfilter). I.e. when I look in build_dir/target-<>/linux-<>/ I only see linux kernel directory and none of the in tree modules like I do on other OpenWRTs. Dec 03 14:43:51 rmilecki: yeah, I've a fold fon device iirc Dec 03 14:44:09 jow: could you get me a boot log from it? Dec 03 14:44:49 rmilecki: yes, but not today, tomorrow earliest Dec 03 14:44:59 sure, no rush Dec 03 14:45:19 Do you have any suggestions? Is using a external kernel tree the way to go with this? Otherwise I was just going to patch the packages that depended on kmods to not depend on them then manually build them in the other sdks. But I'm much rather OpenWRT take care of it all for many reasons including package dependency management. Dec 03 14:50:29 actually jow, sorry ignore that. I got confused with other packages. It actually looks like it worked! Dec 03 14:50:38 really amazing how versatile it is Dec 03 15:19:45 rmilecki: http://paste.ubuntu.com/p/9mHTGghGSJ/ Dec 03 15:20:12 ynezz: is that ath25? Dec 03 15:20:16 yep Dec 03 15:20:33 thanks, i'll look at that later today Dec 03 15:20:35 ! **** BEGIN LOGGING AT Mon Dec 03 15:54:29 2018 Dec 03 16:08:22 build #1068 of lantiq/falcon is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/lantiq%2Ffalcon/builds/1068 Dec 03 16:08:58 hrm, looks like one of the profiling/tracing options being turned on doesn't do the right thing magiclly when you move from 4.9 to 4.14. re-genned the config file from the snippet stub and it's fine. Dec 03 16:09:13 rmilecki: I have an aht25 around too if you need them later. Dec 03 16:09:26 (I don't use them much anymore, but still got them) Dec 03 16:09:40 they're still deployed and happily running away doing their thing Dec 03 16:36:24 rmilecki: do you have any staging branch with all your collected fstools/block/mountd patches? Dec 03 16:37:18 also, if you share your little notefile, I'll test out things and help write/update a wiki page for you. Dec 03 16:39:36 so uh, what was the reason behind reverting microcode stuff for x86? Dec 03 16:40:40 which one? Dec 03 16:42:52 what packages are essential for openwrt ? Dec 03 16:43:46 jwh: grub error Dec 03 16:44:22 gilly_: default selection minus opkg, odhcpd, odhc6c, dnsmasq, pppd (+related kmods) Dec 03 16:45:03 what packages are essential for openwrt ? Dec 03 16:45:16 17:44 < jow> gilly_: default selection minus opkg, odhcpd, odhc6c, dnsmasq, pppd (+related kmods Dec 03 16:45:33 i am trying to build for 4mb Dec 03 16:45:41 i dont need router functions Dec 03 16:45:46 just a linux box Dec 03 16:46:21 i tried removing above packages Dec 03 16:46:37 are there anymore that i can remove safely ? Dec 03 16:48:24 well if you write custom init scripts and have serial access then you can remove almost everything execpt busybox, libc and the kernel Dec 03 16:48:47 the resulting system will not work ootb though Dec 03 16:49:05 e.g. if you remove procd (openwrt's init) then you need to reconfigure busybox to enable init Dec 03 16:49:44 i dont have serial access Dec 03 16:50:08 well then stick to the default selection minus the packages above Dec 03 16:50:37 if you do not want ssh access you can enable telnetd in busybox and disable dropbear as well Dec 03 16:50:57 jow: ah Dec 03 16:54:57 i need about 200kb more. can you help me figure out what i need to do more ? Dec 03 16:55:04 here are the logs Dec 03 16:55:05 https://pastebin.com/cfSMAw2P Dec 03 16:57:12 -usign -openwrt-keyring Dec 03 16:58:07 if you've already stripped out everything and it doesn't fit, what do you think you're going to be able to add to do something useful again with? Dec 03 16:58:56 karlp: aircrack Dec 03 16:59:16 gilly1500: maybe try passing CONFIG_TARGET_SQUASHFS_BLOCK_SIZE=1024 Dec 03 16:59:28 gilly1500: e.g. make images PACKAGES="..." CONFIG_TARGET_SQUASHFS_BLOCK_SIZE=1024 Dec 03 17:01:50 probablyu dfon't need uboot-envtools and uclient-fetch either? Dec 03 17:02:28 CONFIG_TARGET_SQUASHFS_BLOCK_SIZE=1024 didnt help Dec 03 17:15:53 hm, kinda wanna unify the ipip stuff with 6in4 but I'm not really sure how the proto hooks work - is there a primer on the wiki? Dec 03 17:25:54 also can there be aliases or does it literally just exec the hook with thta name? Dec 03 17:25:57 that* Dec 03 17:26:00 guess symlinks work Dec 03 17:58:48 build #777 of ramips/mt76x8 is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/ramips%2Fmt76x8/builds/777 Dec 03 18:15:16 also umm, jow you might know this... will pppd come up by default without IPCP? Dec 03 18:16:25 or... can I tell it to disable IPCP in interface config sections? Dec 03 18:24:25 I guess what I'm really asking is where do the params for pppd get generated? Dec 03 18:25:10 nm Dec 03 19:09:36 build #1142 of brcm47xx/mips74k is complete: Failure [failed targetupload] Build details are at http://phase1.builds.lede-project.org/builders/brcm47xx%2Fmips74k/builds/1142 blamelist: Florian Eckert , Chen Minqiang , Mathias Kresin , Rafa? Mi?ecki , Russell Senior Dec 03 19:09:36 , Eneas U de Queiroz , Pawel Dembicki Dec 03 19:35:54 https://github.com/joeholden/openwrt/commit/a654093caba5f2bacca744d93f1940fdad15d85b Dec 03 19:36:06 pretty sure thats garbage, but you can see what I mean Dec 03 19:58:00 build #1000 of malta/be is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/malta%2Fbe/builds/1000 Dec 03 22:39:15 build #378 of mvebu/cortexa9 is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/mvebu%2Fcortexa9/builds/378 Dec 04 00:12:14 build #325 of ipq40xx/generic is complete: Failure [failed pkgbuild] Build details are at http://phase1.builds.lede-project.org/builders/ipq40xx%2Fgeneric/builds/325 blamelist: Rafa? Mi?ecki , Chen Minqiang , Mathias Kresin , Pawel Dembicki Dec 04 00:13:49 build #284 of oxnas/ox820 is complete: Failure [failed targetupload] Build details are at http://phase1.builds.lede-project.org/builders/oxnas%2Fox820/builds/284 blamelist: Rafa? Mi?ecki , Chen Minqiang , Mathias Kresin , Pawel Dembicki Dec 04 00:14:10 Is there any firmware that can be loaded on the FVS318G Broadcom BCM53118. Mine will not let me into the administrative pages without aborting session even if I had logged in and went direct to the tab. Dec 04 00:17:31 chris-owen: it's not in the ToH, which means it's not supported - that doesn't mean it would be unsupportable, but it will need device specific work Dec 04 00:20:05 pkgadd: sir, what do you think to my proposal to disable IPCP? Dec 04 00:20:16 you're the only other I know who still uses dsl/ppp Dec 04 00:22:50 ok Dec 04 00:23:06 what would be a good starting place? Dec 04 00:24:31 jwh: I haven't really looked into the details, if you'd want me to test something specific, I can do that Dec 04 00:24:51 chris-owen: get access to serial console, obtain stock firmware image, and bootlog, find a GPL sdk from the vendor Dec 04 00:25:12 pkgadd: just feature parity really, 'ipv4' option akin to the ipv6 one (just appends noip to the pppd params) Dec 04 00:25:17 that means, I can test PPPoE, not PPPoA Dec 04 00:25:24 for those crazies who only want v6 Dec 04 00:26:00 I looked at just adding it to pppd_options but I thought it might be better being the same as the ipv6 option Dec 04 00:27:00 it's currently invoked as: /usr/sbin/pppd nodetach ipparam wan ifname pppoe-wan lcp-echo-interval 1 lcp-echo-failure 5 lcp-echo-adaptive +ipv6 set AUTOIPV6=1 nodefaultroute usepeerdns maxfail 1 user xxx@xxx password xxx ip-up-script /lib/netifd/ppp-up ipv6-up-script /lib/netifd/ppp6-up ip-down-script /lib/netifd/ppp-down ipv6-down-script Dec 04 00:27:10 yeah Dec 04 00:27:34 so, basic idea is that doing uci set network.wan.ipv4=0 will just stop pppd bothering to negotiate IPCP Dec 04 00:28:19 theres some shell magic I don't quite understand in there though heh Dec 04 00:28:54 it's been very long ago since I last looked into ppp specifics and configurations, early ISDN days... Dec 04 00:30:18 the +ipv6 got me too, pppd should negotiate that by default unless noipv6 is specified Dec 04 00:31:46 the option on its own is umm, extraneous? Dec 04 00:31:55 I have to admit that my own domain/ mail setup is still IPv4-only - simple reason, I just needs a simple dumping ground (shared webhosting) with 'unlimited' email possibilities - and price was a major selling point (6.60 EUR/ year) Dec 04 00:32:06 heh Dec 04 00:33:00 I'm not happy with that situation and would very much prefer IPv6 support for it, but so far I haven't found anything that could compete in that price region Dec 04 00:33:20 yeah, can't really argue with that unsustainable pricing :D Dec 04 00:34:01 https://wint.global//linux-hosting-pakete-plesk/ classic 1.0 Dec 04 00:34:08 I'm just going through my todo list really, bunch of little things like pppd Dec 04 00:34:25 probably not that useful to be able to disable ipv4 on ppp for most people, but feature parity+++ Dec 04 00:35:02 lol thats crazy Dec 04 00:35:16 the domain probably costs that Dec 04 00:36:01 I've registered my .de domain elsewhere, but yes, that costs (using a different registrar) 4.85 EUR/ year Dec 04 00:36:35 s/4.85/4.65/ Dec 04 00:37:43 I've been using it for just over 3 years now, no IPv6 (and limited to 100 mails per hour), but other than that it does the job Dec 04 00:37:53 And all you have to do is sell your soul to shared hosting Dec 04 00:37:56 pretty reasonable Dec 04 00:38:01 I wouldn't but heh Dec 04 00:38:41 there are some crazy deals on lowendbox occasionally though, I got a crappy openvz (yeah yeah, sucks but ridiculous spec) for $10/year Dec 04 00:38:47 or 2 years Dec 04 00:39:01 Monkeh: I really don't want to maintain my own mailserver, the webspace contains of a single -static- page I haven't touched in over 15 years. plus some hidden dumping ground for stuff I might need Dec 04 00:39:10 so my needs are very, very low Dec 04 00:39:18 ah heh Dec 04 00:40:02 ok what am I missing Dec 04 00:40:10 I can't find docs for ipip support in uci Dec 04 00:40:24 but it serves my mailing needs for four domains (just had to set up something local as a rss2email gateway, as that blew the 100 mails/ hours limit) Dec 04 00:42:24 ok so the support is mostly missing Dec 04 00:42:29 I shall fix that tomorrow maybe Dec 04 00:42:43 once I figure out how the proto handlers work Dec 04 00:46:47 theres a missing link somewhere Dec 04 00:47:18 does netifd have a generic tunnel handler? is that what I'm missing? Dec 04 00:48:13 I don't really know how all the different (but oh so similar) tunnel types are configured... if its the same API that would make sense Dec 04 00:49:41 bah, lldpd 1.0.2 checks for libcap without pkgconfig now too, and finds it because I have it as dependency for something else, but it's in packages feed so shouldn't depend on it, and I'm unable to convince autocrap to not link against libcap :/ Dec 04 00:49:57 :( Dec 04 01:23:31 Hi! I got a mt76 device that doesn't have the correct MAC address set in EEPROM data and the correct MAC address needs to be extracted with a script. It seems that current mt76 driver doesn't have a firmware hotplug for eeprom data available. How should I set the MAC address for them? Should I use the fix_wifi_mac script found in other target? Dec 04 01:24:06 macaddr no good? Dec 04 01:27:04 The manufacturer uses 00:0c:43:e1:76:28 in their eeprom while writing the real mac in other partition using a format similar to u-boot-env. Dec 04 01:29:03 So the MAC address for wifi is wrong. Dec 04 02:14:39 build #106 of samsung/s5pv210 is complete: Failure [failed kmodupload] Build details are at http://phase1.builds.lede-project.org/builders/samsung%2Fs5pv210/builds/106 blamelist: Rafa? Mi?ecki , Chen Minqiang , Mathias Kresin , Pawel Dembicki Dec 04 02:14:47 build #229 of at91/sama5d2 is complete: Failure [failed targetupload] Build details are at http://phase1.builds.lede-project.org/builders/at91%2Fsama5d2/builds/229 blamelist: Rafa? Mi?ecki , Chen Minqiang , Mathias Kresin , Pawel Dembicki **** ENDING LOGGING AT Tue Dec 04 03:00:15 2018