**** BEGIN LOGGING AT Wed Jul 25 03:00:02 2018 Jul 25 03:12:11 hello Jul 25 03:12:32 i wrote a simple kernel module for openwrt, Jul 25 03:12:55 there are 2 source file. A.c and B.c Jul 25 03:14:03 and success to build the ipk, but failed to find the function which defined in the second source file Jul 25 03:14:28 after moved all the codes into the first source file, it is works, Jul 25 03:15:06 hence, how to add multiple obj files into the KO file? Jul 25 03:40:34 hey, got an interesting issue with openwrt (LEDE snapshot). Trying to build for the Aerohive AP122 and starting with the netgear R6250 (the closest match). Amazingly, the kernel starts to boot, with it halts with a big nasty error that I don't know how to interpret (https://pastebin.com/Y1HJB3Gh). Anyone willing to give me some pointers on what I should try? Jul 25 04:10:07 mangix: pong Jul 25 04:26:23 russell--: is this for you? https://github.com/openwrt/packages/pull/6496 Jul 25 04:49:59 lol, i have no memory of that. ;-) Jul 25 04:49:59 rmilecki: I think you need to rebuild package/toolchain/ as well to bundle the updated libc.so Jul 25 04:52:20 mangix: there's my name right there ;-) Jul 25 05:11:55 jow: ping Jul 25 05:12:16 just saw the mail from realtek, i'll draft a reply during the day Jul 25 05:15:29 mangix: i'll test the update Jul 25 05:18:58 jow: i think "make V=s" should rebuild that... Jul 25 05:19:04 should I remove some .built file? Jul 25 05:19:38 we got a mail from realtek? Jul 25 05:23:04 first I tried rm -R build_dir/target-arm_cortex-a9_musl-1.1.16_eabi/toolchain/ and then make V=s Jul 25 05:23:07 it didn't help Jul 25 05:23:18 then I did rm -R staging_dir/toolchain-arm_cortex-a9_gcc-5.4.0_musl-1.1.16_eabi/ and make V=s Jul 25 05:23:25 jow: yes, they want their newest SoC officially supported Jul 25 05:23:30 It worked but it took quite a lot of time to compile Jul 25 05:27:09 blogic: oh cool. Jul 25 05:32:53 russell--: i have not tested it, no Jul 25 05:57:49 * russell-- is reconstructing an image he used it from, to make sure it still flashes ap51's ... how i got to be maintainer in the first place, lol Jul 25 06:03:58 shit... I just tried rm staging_dir/toolchain-arm_cortex-a9_gcc-5.4.0_musl-1.1.16_eabi/stamp/.* Jul 25 06:03:59 it didn't help Jul 25 06:04:29 the only reliable way I found to get musl recompiled & libc packaged is rm -R staging_dir/toolchain-arm_cortex-a9_gcc-5.4.0_musl-1.1.16_eabi/ Jul 25 06:05:18 blogic: do you know how to properly rebuild musl & make OpenWrt include updated libc? Jul 25 06:05:44 that method takes too much time to recompile images: rm -R staging_dir/toolchain-arm_cortex-a9_gcc-5.4.0_musl-1.1.16_eabi/ Jul 25 06:26:39 rmilecki: removing stampfile, make toolchain/musl/compile + make package/toolchain/{clean,compile} should work Jul 25 06:42:05 jow: just for sure: which stamfile? Jul 25 06:54:14 rmilecki: the one I mentioned yesterday Jul 25 06:54:22 ok Jul 25 06:54:24 thanks! Jul 25 07:45:02 jow: that didn't work... Jul 25 07:45:38 rm build_dir/toolchain-arm_cortex-a9_gcc-5.4.0_musl-1.1.16_eabi/musl/.built && make toolchain/musl/compile V=s && make package/toolchain/{clean,compile} V=s && make V=s && make V=s Jul 25 07:45:41 didn't work Jul 25 08:02:35 rmilecki: work in what regard exaclty? Updated libc.so not found on the image? Jul 25 08:03:53 jow: my changes in build_dir/toolchain-arm_cortex-a9_gcc-5.4.0_musl-1.1.16_eabi/musl-1.1.16/src/time/strptime.c didn't end in a final images Jul 25 08:08:59 hey, i'm trying to get a dts written for my rspro. it's ancient redboot and i'm trying to load a .elf via tftp. it's basically "cat linux.elf rspro.dtb > appended.elf" and tftpboot appended.elf Jul 25 08:09:52 mangix: update tested successfully Jul 25 08:10:07 this is what i'm seeing https://paste.ee/p/aVyIa#niBM7Qwbgq5WOkKxB3WMsP2hTPsFCACt Jul 25 08:10:15 i wonder if the appended dtb doesn't work with raw elf? Jul 25 08:12:08 jow: ping, query? Jul 25 08:15:16 rmilecki: the above command usually works for me Jul 25 08:15:42 good morning people Jul 25 08:19:30 hrm https://www.linux-mips.org/archives/linux-mips/2015-09/msg00082.html Jul 25 08:19:35 maybe i need to do something like that :( Jul 25 08:21:53 hey it's mainlined tho ;) i have a hunch this will work... Jul 25 08:23:56 14 .appended_dtb 00100000 8079e514 8079e514 0073f514 2**2 Jul 25 08:35:42 m4t: use APPENDED_DTB inside the kernel Jul 25 08:37:27 yeah i enabled MIPS_ELF_APPENDED_DTB Jul 25 08:37:34 i can only boot elf so the .bin doesn't work Jul 25 08:38:21 still isn't finding it though. i used objcopy like the help text shows. Jul 25 08:49:32 just for kicks i objcopy'd the wndr3800 dtb and same thing Jul 25 09:15:47 hmm... whats the best way to test for a ramoverlay? Jul 25 09:18:12 /tmp/root I guess Jul 25 09:22:54 hi, anyone know why ncurses/host does not install any libs to staging_dir/hostpkg/lib/ ? I need this lib for a host compiled helper. Jul 25 09:24:26 andy2244: because the makefile is written in such a way that it only stages the tic compiler Jul 25 09:24:36 mainly to cut down build time Jul 25 09:24:51 yeah i noticed, but it compiles the libs anyway on host? Jul 25 09:25:25 yes because some of them are required for tic Jul 25 09:25:52 if what you need is already built then you can just extend the Host/Install recipe Jul 25 09:26:01 (and send a patch for it) Jul 25 09:27:20 f00b4r0: pong Jul 25 09:27:40 mhh oki, the next lib i need is libpthread, but seems no package is actually building this one? So i'm supposed to grab the root-fs one from /lib(32/64) ? Jul 25 09:27:47 jow: have to run, will be back later. Sent you an email. Jul 25 09:28:23 f00b4r0: allright Jul 25 09:28:38 andy2244: on musl libc there is no libpthread, its part of the core libc Jul 25 09:29:22 ah oki so i have to patch the makefile if a package explicitly looks for this libname? Jul 25 09:30:08 yes Jul 25 09:30:15 cool thanks Jul 25 09:30:30 it might be that musl even provides empty libpthread.a stubs for satisfying any -lpthread Jul 25 09:30:37 but I didn't check Jul 25 09:31:01 at least the cmake script cant find any valid on my test system Jul 25 09:31:32 yeah, if the script explicitely looks for a libpthread then it is going to fail Jul 25 09:31:57 iirc old autoconf had some macros to check if -lpthread, -pthread or nothing is required to obtain pthread interfaces Jul 25 09:32:28 but could be that people do not anticipate !-lpthread anymore in modern glibc times Jul 25 09:32:42 thats the cmake line "find_library(LIB_PTHREAD pthread)" not sure how find_library actually works Jul 25 09:32:51 it looks for a shared library Jul 25 09:33:01 or a static library Jul 25 09:33:52 andy2244: https://stackoverflow.com/a/4774027 Jul 25 09:34:31 that solution should also be acceptable for upstream Jul 25 09:34:41 treating "pthread support" like an ordinary library is not portable Jul 25 09:35:44 ah thanks will to this and try a upstream pr for it than too. Jul 25 09:40:11 oooh 4.14.58 featuring another patch from stable that I had a hand (foot?) in falling over and testing the resultant fix. It will be a pleasure in removing it from openwrt's backports directory :-) Jul 25 09:40:27 yay i got it Jul 25 09:41:06 cat vmlinux.bin rspro.dtb > kernel.bin. then load that as "raw" via tftp at the same load address that the uimages use Jul 25 09:41:27 i.e. load -r -v -b 0x80060000 -m tftp -h 192.168.1.2 vmlinux.bin ; exec -c "" Jul 25 09:43:01 ldir: I am sure they sneaked in another bug for it ;) Jul 25 09:45:20 lol - yeah probably :-) People leave this sort of stuff for me to fall over.... https://pbs.twimg.com/media/DJyAs4rXoAIqyMb.jpg is what happens every single time :-) Jul 25 09:46:35 although the first frame of that cartoon is wrong... it should be me minding my own business, tripping over the bug, then saying "oh my a little bug!" Jul 25 09:48:01 rmilecki: regarding that pcie errata thing that breaks booting, are you going to push a revert-patch for 18.06 ? Jul 25 09:48:24 rmilecki: asking because I planned to tag on friday/saturday so it would be good to have the stuff done by then Jul 25 09:49:09 i'll have that done by then Jul 25 09:49:16 I just want to finish my musl struggle Jul 25 09:49:29 i'm almost done Jul 25 09:51:52 rspro Linux (none) 4.14.54 #0 Wed Jul 25 06:15:28 2018 mips GNU/Linux Jul 25 09:52:06 well. that's the easy part i guess. i need to figure out how to describe this switch :o Jul 25 09:53:07 rmilecki: thanks, didn't mean to pressure Jul 25 10:07:37 jow: i'll push the 2 ubus fixes to master and the 2 stable trees right ? Jul 25 10:08:06 i've sent ~50 of our patches upstream the last few days btw Jul 25 10:19:48 blogic: yes please Jul 25 10:41:11 Does anyone know why stintel has this big ext4 warning for 4.14.56 ? https://git.openwrt.org/?p=openwrt/staging/stintel.git;a=commit;h=676b4f3e1e3c31c6288897d99104408b22ceea6e And is it still relevant for 4.14.57 (which I have as a refresh in my tree) and 4.14.58 which I may have done by the end of the day depending on more important stuff. Jul 25 10:59:42 * karlp cheers for upstreaming. wooo Jul 25 11:00:22 karlp: more to come Jul 25 11:00:37 ldir: it was related to some build option he had Jul 25 11:02:43 blogic: lol - I turned the volume up on my macbook a little earlier so your message just triggered the loudest, unexpected 'ting' ever and I nearly jumped off the sofa. Mrs ldir laughed. Jul 25 11:03:01 ldir: lol Jul 25 11:03:23 anyway, the message is bogus, right, ok. Jul 25 11:04:15 by which I mean his warning is bogus. Jul 25 11:04:45 ldir: not 100% sure yet Jul 25 11:04:56 but its not that superblock fuzzying issue he assumed Jul 25 11:05:08 jow: messed up the shortlog on the 18.06 ubus update Jul 25 11:05:20 rebased the cherry manually but frgot to update shortlog .... Jul 25 11:05:53 * blogic opens the 4th bottle of water of the day Jul 25 11:06:07 its crazy warm here Jul 25 11:06:07 * jow wishes he'd open the fourth bottle of beer at work Jul 25 11:06:33 jow: i was actually considering to go buy 1-2 beers and put them in the fridge for tonight Jul 25 11:06:43 oh that's interesting 'cos I was expecting it to be fallout from the ext4 'more fussy' CVE fixes. Jul 25 11:07:03 ldir: that is what we assumed but apparently not Jul 25 11:07:39 https://github.com/openwrt/openwrt/pull/597 is this pull going to be merged ? Jul 25 11:08:49 a__pi_: reluctant to do so Jul 25 11:09:01 why? Jul 25 11:13:32 a__pi_: yet another target and an out of tree driver that will never go upstream, that i will have to keep alive till ever for 1 single board Jul 25 11:13:57 specially with recent changes in nand upstream, all out of tree drivers will heavily break with the next bump Jul 25 11:14:08 and then i'll need to spend a week fixing them Jul 25 11:14:12 and i dont want to do that Jul 25 11:14:39 nobody wants to do it Jul 25 11:22:50 jow: pushed trunk and 18.06 Jul 25 11:23:01 building a x86 image for 17.01 just to be sure Jul 25 11:39:47 * ldir hits head against qnap support. Jul 25 12:43:33 Hi. Just wanted to say thanks for all the effort that's gone into OpenWRT. I just upgraded from Lede 17.01 to OpenWRT 18.06-rc2 and everything is working great :) Jul 25 13:12:25 hi, is there any way to have two users on Luci? Like one admin with full access and one guest with limited options? Jul 25 13:32:42 alex___: no Jul 25 14:20:26 * ldir tries to melt his laptop again - yes it's another compilathon. Jul 25 14:20:36 just go outside Jul 25 14:20:38 that will help Jul 25 14:20:55 no it won't! :-) Jul 25 14:21:16 is it not melting where you are? Jul 25 14:21:39 yes it is, but I suspect it's warmer inside than out. Jul 25 14:21:55 * ldir turns into a puddle under the desk Jul 25 14:22:03 direct sunlight tends to heat things up though :D Jul 25 14:23:03 hm, wonder if BIRD v2 supports MP-BGP properly Jul 25 14:23:06 just put it in a car and place some aluminum foil under it. that should help. Jul 25 14:23:29 would quite like only v6 sessions (and maybe addressing) Jul 25 15:02:18 hi guys. What packages do I need to connect this -> https://duckduckgo.com/?q=huawei+e5785&atb=v108-2__&iax=images&ia=images to this -> https://duckduckgo.com/?q=netgear+wndr3800&atb=v108-2__&iar=images&iax=images&ia=images via USB? It will connect to LTE Jul 25 15:19:26 aversario: aither uqmi or umbim Jul 25 15:19:36 and the wwan package Jul 25 15:26:01 blogic: I'm compiling master@a5c3bba(ubus: update to latest git HEAD) for ipq40xx. Do you need any test performed? you mentioned something about usb3 Jul 25 16:05:15 rmilecki: does make toolchain/musl/{clean,compile} not work? Jul 25 16:10:50 on a 16-core machine, my typical build took 6 seconds longer than a change-nothing build (1m25s vs 1m19s) Jul 25 16:19:58 luaraneda: let me know if usb3 works Jul 25 16:20:07 luaraneda: what unit do you have ? Jul 25 16:20:32 luaraneda: i am 2/3 done reworking the ethernet part giving us DSA support, i'll nudge you to test it the next days Jul 25 16:20:37 ;-) Jul 25 16:20:43 tetsing is always appreciated Jul 25 16:22:24 Ok. I have an ASUS RT-AC58U Jul 25 16:23:05 * blogic checks dts file Jul 25 16:23:38 yep, there is usb3 Jul 25 16:23:55 do you have a superspeed usb3 device that you can attach ? Jul 25 16:24:06 i verified the high speed phy driver Jul 25 16:24:45 but superspeed phy needs to be runtime tested Jul 25 16:25:08 i unfortunatley only have a unit with usb2 HS port Jul 25 16:25:24 Yes, do you need a benchmark? like drive transfer speed Jul 25 16:26:35 nah Jul 25 16:26:44 just need to know if the device is registered Jul 25 16:27:09 its only the phy driver, i did not touch the actual usb host driver Jul 25 16:27:51 you'll see a "failed to probe dwc3" or similar message in the bootog if i foobar"ed Jul 25 16:28:01 *bootlog Jul 25 16:34:34 Kernel log when connecting a device (control date before testing with your patches) Jul 25 16:34:35 [1052110.875847] usb 2-1: new SuperSpeed USB device number 2 using xhci-hcd Jul 25 16:34:58 blogic: You need the same but with your patches applied? Jul 25 16:35:47 s/date/data Jul 25 17:17:53 rmilecki: ping Jul 25 17:18:00 nbd: pong Jul 25 17:18:37 wow - that was a quick game :-) Jul 25 17:19:31 rmilecki: i just re-read your email regarding making musl changes Jul 25 17:19:43 and noticed that the path you edited refers to gcc 5.4.0 Jul 25 17:19:53 do you have a custom config that still uses 5.4.0? Jul 25 17:20:06 just making sure you didn't accidentally edit a file in a build dir that's no longer used Jul 25 17:20:26 I think that was on 18.06 Jul 25 17:20:42 ah, okay Jul 25 17:22:52 oh, wait I was working on 17.01... Jul 25 17:22:55 maybe that's why? Jul 25 17:27:13 ldir: cool, thx Jul 25 17:27:16 oops Jul 25 17:27:20 luaraneda: cool, thx Jul 25 17:27:31 you're welcome anyway :-) Jul 25 17:27:34 luaraneda: i was cheeky and already sent the patch upstream :-) Jul 25 17:28:10 * ldir wonders if blogic has prepared the clown outfit? Jul 25 17:51:28 blogic: ok. Now that I know what kind of data you are looking for, I'll sysupgrade and post the updated data Jul 25 17:57:38 hello itsa meee, any examples of go packages? :D Jul 25 17:58:08 ldir: :-D Jul 25 17:58:19 ldir: davem is the worste in that regard Jul 25 17:59:56 gninrom Jul 25 18:05:32 he he - he had a bit of a grumble at the conntrack lookup in cake but let it through, I'd like to think 'cos he could see the value and then fact there wasn't any other way of doing what we needed to do. Jul 25 18:12:15 was looking at this https://bugs.openwrt.org/index.php?do=details&task_id=896 since I'm affected, found it a bit odd that psyborg is suggesting that if one pays him he'll fix the bug. Jul 25 18:12:18 isn't that a big no no? Jul 25 18:15:24 f00b4r0: pfff Jul 25 18:15:34 pfff? Jul 25 18:15:58 pay for foss code ..... Jul 25 18:16:22 if i got 1 euro / line of code i produced i'd be retired by now Jul 25 18:16:29 that's not the point. Jul 25 18:16:46 i code so that i can undestand and explain the problem Jul 25 18:16:58 ... which my code solves Jul 25 18:17:15 I take it it's fine to go fishing on the flyspray Jul 25 18:17:26 sure, go for it Jul 25 18:17:35 make sure to wear protective gear Jul 25 18:20:27 blogic: don't think I need any. I'm very aggressive and brain dead, remember? ;P Jul 25 18:20:39 goo Jul 25 18:20:42 d Jul 25 18:20:48 go go go Jul 25 18:21:00 jow: am back. Jul 25 18:21:28 f00b4r0: will you be at the summit ? Jul 25 18:22:16 blogic: no Jul 25 18:22:24 why ? Jul 25 18:22:27 would be fun Jul 25 18:22:30 sure. Jul 25 18:22:38 drinks, nergasm, ... Jul 25 18:22:53 *nerdgasm Jul 25 18:22:56 last summit I attended was OLS 2007. Kinda retired after that. Not sure I'd know how to behave socially anymore. Jul 25 18:23:10 lulz Jul 25 18:24:01 when and where is this summit? :D Jul 25 18:36:40 blogic: USB3 working: Jul 25 18:36:41 [ 138.713395] usb 2-1: new SuperSpeed USB device number 2 using xhci-hcd Jul 25 18:37:12 But one of the commits broke my internet connection. I'm bisecting now Jul 25 18:38:35 f00b4r0: you know you want to go really - you can point and laugh at this Englishman for a start :-) Jul 25 18:39:05 :) Jul 25 18:39:14 f00b4r0: i went to a nerd event in 2014 after not going to any sort of nerd event since 2002. Jul 25 18:39:31 f00b4r0: wouldn't say paid offers for help are a big no no, as long as its sort of a private transaction and no big commercial operation Jul 25 18:39:32 this will be my first nerd event. Jul 25 18:39:38 f00b4r0: a lot had changed then. for one, there weren't enough drunk people. Jul 25 18:40:19 someone tried to fight me in a glass elevator Jul 25 18:40:40 jow: ok. I'm afraid whoever falls for that bait will be quite disappointed, given his trackrecord and the fact that he doesn't appear in upstream conversations about this bug ;P Jul 25 18:40:55 f00b4r0: I agree Jul 25 18:41:15 f00b4r0: I can can laugh in your general direction...and you can taunt me a second time ;-) Jul 25 18:41:25 even if something comes out of it it'll likely a hackjob breaking all other models Jul 25 18:41:29 *be a Jul 25 18:41:36 jow: rmilecki what is the problem ith this set_c0_config7(MIPS_CONF7_ES); change? Jul 25 18:41:53 I also was not able to really try this as CFE sets it Jul 25 18:41:57 Hauke: all we know is that it causes a kernel hang during boot on a wrt610n v2 / e3000 Jul 25 18:41:59 but I haven't runtime tested it Jul 25 18:42:07 where hanetzer :( Jul 25 18:42:11 jow: indeed! Jul 25 18:42:23 fu guys :( Jul 25 18:42:39 and where's all the nand peoplez Jul 25 18:43:19 i have ar300m and ar750s is coming in a few weeks. have serial. have jtag. have all the toys. want to help get this nand stuff completed. Jul 25 18:43:26 salcedo: struck to death by bad blocks Jul 25 18:43:38 jow: like IRL death by bad blocks?! Jul 25 18:43:56 no, bad attempt at humour Jul 25 18:45:08 we left off at trying to get the latest kernel working for ar300m nand and i suspected the issue is an off-by-one somewhere with ECC Jul 25 18:45:33 Hauke: here's a boot log up to the hang: https://bugs.openwrt.org/index.php?do=details&task_id=1688 Jul 25 18:45:47 Hauke: I think next thing that should happen at this stage is bringing up the wifi driver Jul 25 18:46:10 blogic: I started some yers ago to write a dma channel driver but never finished Jul 25 18:46:22 if you can still find your it would be nice if you can send it to me Jul 25 18:49:12 jow: hmm Jul 25 18:49:27 jow: did you revert the change and then it does not happen? Jul 25 18:51:19 jow: I will be back home on saturday evening, I can search for such a device and test it on sunday Jul 25 18:51:29 or do you also have such a device? Jul 25 18:52:42 Hauke: mamarley successfully tested reverting it, and the boot completes Jul 25 18:53:10 Hauke: maybe the war should be restricted to newer SoCs Jul 25 18:53:39 I think this SoC was also effected by this problem Jul 25 18:54:09 jow: we should write a mail to this guy who send this patch he has access to the documentation Jul 25 18:55:48 jow: the brcmsmac driver says: bcm4716 (which includes 4717 & 4718), plus 4706 on PCIe can reorder transactions. As a fix, a read after write is performed on certain places in the code. Older chips and the newer 5357 family don't require this fix. Jul 25 18:57:24 I will go to sleep now Jul 25 19:08:02 hola Jul 25 19:08:34 no puedo acceder al router con openwrt instalado alguien me ayuda? Jul 25 19:08:48 chulis: this is an english speaking channel. Jul 25 19:09:38 ok where is the spanish chat? Jul 25 19:29:44 Hi all, when using the files directory are permissions maintained? My dropbear directory permissions appear to be changing? Jul 25 19:33:48 chulis: As far as I know, the Spanish channel does not exists. Can you state your problem in English? Jul 25 19:34:20 Ups, he/she left :s Jul 25 20:16:16 blogic: thanks, but are you sure that isn't https://openwrt.org/docs/guide-user/network/wan/wwan/ethernetoverusb_ncm or https://openwrt.org/docs/guide-user/network/wan/wwan/ethernetoverusb_rndis ? Jul 25 20:38:33 blogic: good and bad news. All of your ipq40xx commits are fine Jul 25 20:39:48 blogic: nbd: The last commit to ubus (libubus: check for non-NULL data before running callbacks) broke something Jul 25 20:39:55 I can't get internet access Jul 25 20:42:29 I'm connecting through PPPoE, the pppoe-wan interface is working fine, because I have internet connectivity if I ssh into the router, but local clients can't access internet (not even ping to 8.8.8.8) Jul 25 20:47:37 It's seems to be firewall/nat related, but I'm not sure Jul 25 23:00:23 hello all! Jul 25 23:01:19 what is the preferred way to use when trying to share a hotel wifi that has a login and password prompt. I am trying to setup my router as the device that logs in but, how to log in? Jul 25 23:01:34 are there any text based web clients that I can use? Jul 25 23:02:16 in the past there were some packages that were available, but currently I dont know what to use. Jul 25 23:02:29 buy a subscription and stop breaking the terms of service :D Jul 25 23:02:49 that would be one route to go... Jul 25 23:02:57 also, #openwrt Jul 25 23:03:04 this is development (hence the -devel) Jul 25 23:03:26 ok, my bad. see you there! =) Jul 25 23:04:07 * karlp thinks ssh -D would probably have been enough? **** ENDING LOGGING AT Thu Jul 26 03:00:01 2018