**** BEGIN LOGGING AT Sun Feb 14 02:59:57 2021 Feb 14 06:21:49 hurricos: hey. Did you see the patch for ath9k regarding SMPS? Feb 14 11:12:31 build #782 of archs38/generic is complete: Success [build successful] Build details are at http://buildbot.openwrt.org/master/images/builders/archs38%2Fgeneric/builds/782 Feb 14 13:13:01 reverting "memblock: do not start bottom-up allocations" from the v5.4.97 kernel fixed my mt7621/dir860l kernel crash problem Feb 14 13:35:53 build #642 of ath79/tiny is complete: Failure [failed updatefeeds] Build details are at http://buildbot.openwrt.org/master/images/builders/ath79%2Ftiny/builds/642 blamelist: Petr ?tetiar , Sander Vanheule , Tomasz Maciej Nowak Feb 14 14:51:34 zorun: do you want to commit your changes to opkg? Feb 14 14:51:54 zorun: I would like to wait till the build bot build this once before doing a new release Feb 14 15:06:25 omg what do people have with this /dev/mem interface Feb 14 15:07:56 i guess that is what needs to be determined Feb 14 15:08:49 would seem like a simple question: is it needed for basic operations or not? Feb 14 15:11:03 might have overlooked it but i havent seen a nbd counter argument yet: "there needs to be a good reason to add it" Feb 14 15:14:13 Debian has it enabled! Feb 14 15:17:00 * urjaman chuckles Feb 14 15:18:17 There are other users, but the first that comes to mind for me is flashrom Feb 14 15:19:26 I think you need it for some DOS games Feb 14 15:19:41 and /dev/mem is nice for debugging drivers/HW Feb 14 15:20:05 indeed, but you can have development kernel for that Feb 14 15:21:06 that does make me wonder though (might be a stupid question) but why does openwrt have debufs enabled by default? Feb 14 15:21:14 debugfs Feb 14 15:21:47 yeah i think oldfashioned dosemu used to use it atleast, and xorg when initializing graphics cards by emulating their vbios used to too (these are both like i dunno if that's a thing anymore but...) Feb 14 15:23:07 I am also against activating it by default Feb 14 15:23:57 grift: debugfs is very useful to allow users do dump the rate control stats when they run into wifi issues Feb 14 15:24:16 k just wondering Feb 14 15:25:22 and yeah the more different your dev/debug software is from what you normally have on a device, the harder debugging surprise things becomes Feb 14 15:26:50 nbd: shouldn't then we just limit it to that use case? Feb 14 15:27:09 or just have other means to dump that info Feb 14 15:27:46 enabling debugs by default does feel like a pretty big hammer Feb 14 15:27:50 there's also the crashlog feature Feb 14 15:28:01 and other 802.11 debug things Feb 14 15:28:40 on the upside tracefs isnt enabled Feb 14 15:29:06 i think enabling/disabling debug things by default is always a trade-off Feb 14 15:29:24 if it doesn't cost much, doesn't add many security issues and is used frequently enough, it should be in the default selection Feb 14 15:30:21 when helping people debug wifi issues, being able to rely on debugfs being there by default has saved me personally a lot of time Feb 14 15:30:42 that would have been wasted on walking people through to building a compatible image with debugfs and reflashing their devices Feb 14 15:31:11 grift: debugfs is awesome. It provides GPIO data, clock data. Feb 14 15:31:31 nbd: what about having it in master/release snapshots only? Feb 14 15:31:42 and releases would be compiled without it Feb 14 15:32:25 19.07-SNAPSHOT would have it, 19.07.6 would not Feb 14 15:32:29 Regular distros all have debugfs, just not mounted by default. Feb 14 15:33:23 openwrt is not a regular distro though Feb 14 15:33:30 ynezz: not everybody that runs into wifi issues is using snapshots Feb 14 15:33:42 I've used debugfs many time on OpenWrt, and I'm rather glad it's always there. Feb 14 15:35:10 nbd: ok, but they should be able to mount it manually when they needed it, right? Feb 14 15:35:20 nbd: like we could at least not mount it by default Feb 14 15:35:29 What would you save by doing that? Feb 14 15:35:34 debugfs, bpf, pstore Feb 14 15:35:53 reducing attack surface for the start Feb 14 15:35:59 Hauke: I have already committed it some days ago Feb 14 15:36:17 https://lwn.net/Articles/429326/ Feb 14 15:36:24 Hauke: I am currently debugging the ramips switch issue and I should be done today Feb 14 15:37:08 grift: yeah, but if most of the stuff runs under root Feb 14 15:37:34 right, dont have to convince me, i use selinux to block all access to it Feb 14 15:37:48 i don't really see how we're reducing attack surface by not mounting debugfs Feb 14 15:37:59 zorun: ah sorry, I didn't notcie Feb 14 15:38:22 "Eugene Teo recently posted a request for CVE numbers for 20 separate vulnerabilities involving world-writable files in debugfs and sysfs. Some of the debugfs vulnerabilities would seemingly allow any local user to write arbitrary values into device registers - a situation from which little good can be expected to emerge." Feb 14 15:39:17 that was in `11 though Feb 14 15:39:39 if we limit debugfs to root-only, it should be fine Feb 14 15:44:15 but yes that is exactly where that /dev/mem thread derailed Feb 14 15:44:30 the argument probably shouldnt be about security directly Feb 14 15:44:38 it should just be simple: do we need it? Feb 14 15:44:59 and then the answer to that may lead to security Feb 14 15:48:12 i guess the harsh reality is that we need debugging facilities on stable releases Feb 14 15:49:27 actually what do i know about stable releases, i only use snapshots, but i assume its enabled on stable as well Feb 14 15:49:37 on debian /sys/kernel/debug is root-only Feb 14 15:49:46 it's case in OpenWrt as well Feb 14 15:49:55 0700 and root owned Feb 14 15:50:00 same for bpf, just checked it Feb 14 15:50:31 nothing uses bpf though? Feb 14 15:50:48 bpftools isnt even packaged? Feb 14 15:50:57 http://sprunge.us/2EvG6P (just run https://github.com/a13xp0p0v/kconfig-hardened-check on X86_64 kernel .config) Feb 14 15:50:58 theres no compiler on-device Feb 14 15:53:18 any reason for not having clang in the github workflow? Feb 14 15:54:16 seems like the wrong question: any (good) reason for having it? Feb 14 15:55:31 grift: check 217877d046414878126e5c8013162e3b7edc774c for the background Feb 14 15:56:50 grift: xdp cross compiling Feb 14 16:05:01 Hauke: is it ok to wait this evening (or possibly tomorrow morning) before tagging? Feb 14 16:14:28 zorun: yes no problem Feb 14 16:18:47 Hm, something's wrong with the wiki today, works slowly if at all. Feb 14 16:48:10 build #683 of imx6/generic is complete: Failure [failed pkgbuild] Build details are at http://buildbot.openwrt.org/master/images/builders/imx6%2Fgeneric/builds/683 blamelist: Petr ?tetiar , Sander Vanheule , Rosen Penev , Hauke Mehrtens , Tomasz Maciej Nowak Feb 14 17:09:11 swig attack Feb 14 17:13:43 ynezz: see "OpenWrt add dependency to swig for U-Boot builds" on the mailing list Feb 14 17:33:07 PaulFertser: I applied it and it works wonderfully. My 8265 immediately sticks to MCS15 when in the kitchen Feb 14 17:33:14 Radio was not moved Feb 14 17:33:33 hurricos: nbd is just awesome like that :) Feb 14 17:33:44 I've tried resetting the radio a few times to kick it back to MCS7 but no luck Feb 14 17:33:51 so it's a fantastic patch. Feb 14 17:34:28 I have a dense network of OpenWrt MR16s which are presently powered off. Post-COVID this will basically double throughput for us :^) Feb 14 17:34:42 great time to have also discovered https://github.com/rsa9000/atheepmgr to quiet down my radios Feb 14 17:34:55 not just put a cap using txpower in /etc/config/wireless, but actually bring the highest rates real low Feb 14 17:35:33 show just how well ath9k hw can perform in in-person-conference level densities Feb 14 17:36:10 hurricos: it would be rather cool if you wrote a wiki article about it Feb 14 17:36:15 which wiki? Feb 14 17:36:26 Ah yeah Feb 14 17:36:36 hurricos: I think OpenWrt wiki is a good place and the article is going to be saved and indexed there nicely. Feb 14 17:37:02 For sure. I want to avoid telling people too much about atheepmgr however. Feb 14 17:37:34 Less because they'll make too-noisy radios -- I really haven't seen cheap boards with high-power PAs -- more because stupid stuff easily fries it. Feb 14 17:38:05 or, rather, does nothing useful. But maybe the interest in high-power 802.11 has faded? ath9k hardware is quite old by now. Feb 14 17:38:11 I do not think anyone is going to complain about fried radios if you put a visible disclaimer at the beginning, so that shouldn't be an issue. Feb 14 17:38:49 fair enough. Fwiw tweaking ARTs is also not usually that great of an idea all-around, I really need to wait until I can properly test Feb 14 17:39:18 nbd: ping. trying to compile mt76 HEAD but seeing this: https://paste.debian.net/1185476/ Feb 14 18:19:27 Borromini: the update needs some makefile changes. i plan on pushing it today Feb 14 18:20:44 nbd: ok thanks :) Feb 14 18:27:10 nbd: in the mac80211 updated I added some patches to mt76, can you please add them to your repo if you haven't done so Feb 14 18:32:37 Hauke: i already had the change in a branch that i was using for testing with the new version Feb 14 18:32:57 it'll be in mt76 master with my next update Feb 14 18:39:52 nbd: thanks+ Feb 14 18:40:00 mangix: Testing pr update, compiled out fine.. fiber seems to build Feb 14 18:56:14 Borromini, Hauke: mt76 update pushed Feb 14 18:56:28 hurricos: ath9k smps fix is also pushed (and sent upstream) Feb 14 18:57:14 nbd: danke! Feb 14 19:28:16 build #713 of zynq/generic is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/zynq%2Fgeneric/builds/713 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 19:28:16 , Hauke Mehrtens Feb 14 19:34:01 build #720 of bcm53xx/generic is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/bcm53xx%2Fgeneric/builds/720 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 19:34:01 , Hauke Mehrtens Feb 14 19:39:13 adrianschmutzler: this looks related to your change: cp: cannot stat '/builder/shared-workdir/build/LICENSE': No such file or directory Feb 14 19:40:48 nbd: it looks like your patch refresh script works different Feb 14 19:41:06 when I run "make package/mac80211/{clean,refresh,compile} -j9 V=99 Feb 14 19:41:10 I get changes Feb 14 19:41:30 build #783 of archs38/generic is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/archs38%2Fgeneric/builds/783 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 19:41:30 , Hauke Mehrtens Feb 14 19:42:10 build #753 of omap/generic is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/omap%2Fgeneric/builds/753 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 19:42:10 , Felix Fietkau , Hauke Mehrtens Feb 14 19:48:00 build #783 of layerscape/armv8_64b is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/layerscape%2Farmv8_64b/builds/783 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 19:48:03 , Felix Fietkau , Hauke Mehrtens Feb 14 20:00:15 Hauke: where is it looking for the license? Feb 14 20:03:26 build #760 of apm821xx/nand is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/apm821xx%2Fnand/builds/760 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 20:03:26 , Felix Fietkau , Hauke Mehrtens Feb 14 20:05:43 aparcar[m]: I do not know what is happening there, I just saw this error mesage Feb 14 20:05:58 aparcar[m]: look into one of the error messaegs from build bot for details Feb 14 20:11:41 build #785 of lantiq/xway_legacy is complete: Failure [failed pkgbuild] Build details are at http://buildbot.openwrt.org/master/images/builders/lantiq%2Fxway_legacy/builds/785 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 20:11:42 , Felix Fietkau , Hauke Mehrtens Feb 14 20:14:01 build #793 of ramips/rt3883 is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/ramips%2Frt3883/builds/793 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 20:14:01 , Felix Fietkau , Hauke Mehrtens Feb 14 20:29:13 Hauke: Okay found the issue Feb 14 20:29:27 Curious that only the SDK contains the license file but not the imagebuilder Feb 14 20:33:15 wääääh Feb 14 20:33:50 adrianschmutzler: it's in traget/sdk/Makefile Feb 14 20:34:24 - $(TOPDIR)/LICENSE \ Feb 14 20:34:24 + $(TOPDIR)/LICENSES \ Feb 14 20:35:44 there is also something in target/toolchain Feb 14 20:36:19 build #772 of mediatek/mt7623 is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/mediatek%2Fmt7623/builds/772 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 20:36:19 , Felix Fietkau , Hauke Mehrtens Feb 14 20:37:29 adrianschmutzler: I see now Feb 14 20:37:46 probably I only grepped target/linux, because I checked ... Feb 14 20:38:17 readme.toolchain is a wild rename for LICENSE Feb 14 20:38:34 oh nevermind it's a multiline statement Feb 14 20:38:42 so, what do we do there now? Feb 14 20:39:02 build #784 of apm821xx/sata is complete: Failure [failed pkginstall] Build details are at http://buildbot.openwrt.org/master/images/builders/apm821xx%2Fsata/builds/784 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 20:39:02 , Felix Fietkau , Hauke Mehrtens Feb 14 20:39:10 https://termbin.com/qv84w Feb 14 20:39:20 that should do the trick Feb 14 20:39:28 is $(CP) recursive? Feb 14 20:40:27 cp -fpR, so yes Feb 14 20:40:28 yeah, it's ~cp -a Feb 14 20:40:43 "does what you want" ... :) Feb 14 20:41:07 so, what's the purpose of this copying? Feb 14 20:41:47 It's multiline, the entire folder is copied Feb 14 20:41:54 It's not renamed to readme Feb 14 20:42:02 cp -a is usually not a good idea Feb 14 20:42:14 better to be specific and use --preserve i believe Feb 14 20:42:18 cp -fpR according to rules.mk Feb 14 20:42:19 -a is too coarse Feb 14 20:42:50 but why do we need this in TOOLCHAIN_BUILD_DIR and SDK_BUILD_DIR? Feb 14 20:43:02 and depending on the answer, do we need COPYING, too? Feb 14 20:43:26 Hauke: nbd - might be worth checking brew diffutils is installed - I've had problems in the past with subtly different patch refreshes on macos due to a different diff Feb 14 20:44:20 adrianschmutzler: we theoretically offer the toolchain as it's own archive Feb 14 20:44:32 so everything we offer as archive should contain the licenses Feb 14 20:44:38 aka toolchain, sdk aaand imagebuilder Feb 14 20:45:04 build #782 of mvebu/cortexa72 is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/mvebu%2Fcortexa72/builds/782 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 20:45:04 , Felix Fietkau , Hauke Mehrtens Feb 14 20:45:20 so, copying, too Feb 14 20:45:32 I'm working on it Feb 14 20:46:24 https://zerobin.fff.community/?2c21426ce68418e7#K4BouoEVzJsnL07uYCdrTXzlzXPmhmTzTszvkxMfk1o= Feb 14 20:46:26 that's my diff Feb 14 20:46:41 build #768 of armvirt/32 is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/armvirt%2F32/builds/768 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser , Feb 14 20:46:41 Felix Fietkau , Hauke Mehrtens Feb 14 20:46:52 but if you are working on it, that would be nice, because I should actually go back to the kitchen ... Feb 14 20:46:55 updated openwrt/upstream, https://sdwalker.github.io/uscan/index.html Feb 14 20:47:09 nbd: Thakn you! Feb 14 20:47:48 adrianschmutzler: imagebuiler missing Feb 14 20:48:00 there is no "LICENSE" in that makefile Feb 14 20:48:10 build #765 of mediatek/mt7622 is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/mediatek%2Fmt7622/builds/765 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 20:48:10 , Felix Fietkau , Hauke Mehrtens Feb 14 20:49:09 one could add it in a separate commit, if you think it makes sense Feb 14 20:49:38 adrianschmutzler: okay please finish if it you can Feb 14 20:50:08 okay, I'll push a quick one and see Feb 14 20:52:30 Does any method exist for using a serial connection between two OpenWrt hosts as a network? I do see SLIP but it looks like slattach package may have been dropped Feb 14 20:53:22 build #644 of layerscape/armv7 is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/layerscape%2Farmv7/builds/644 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 20:53:23 , Felix Fietkau , Hauke Mehrtens Feb 14 20:54:09 adrianschmutzler: thank you Feb 14 20:54:24 thanks for pointing me to the source Feb 14 20:56:42 okay, so let's see. Feb 14 20:56:59 feel free to take care of imagebuilder as well, but that's not time-critical Feb 14 20:58:07 hurricos: well, ppp? Feb 14 20:59:24 adrianschmutzler: thanks for the quick fix Feb 14 21:00:11 karlp: ahhh. That sounds like it was the obivous answer. This looks promising: https://tldp.org/HOWTO/PPP-HOWTO/direct.html Feb 14 21:00:13 Thank you Feb 14 21:00:22 For those who are at a bored point, can you look at your toolchain directory and give me the filename you show for your -gcc? mips64-openwrt-linux-musl-gcc if the example for mine Feb 14 21:01:50 ldir: i'm using diffutils 3.7 (via brew) Feb 14 21:01:53 build #644 of mediatek/mt7629 is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/mediatek%2Fmt7629/builds/644 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 21:01:53 , Felix Fietkau , Hauke Mehrtens Feb 14 21:02:01 Grommish: https://paste.c-net.org/UnholyHangout Feb 14 21:02:38 hurricos: I need what it shows for your *-gcc :) THe mips64 was mine Feb 14 21:02:54 build #761 of mvebu/cortexa53 is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/mvebu%2Fcortexa53/builds/761 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 21:02:54 , Felix Fietkau , Hauke Mehrtens Feb 14 21:03:01 it's in staging_dir/toolchain-*/bin/*-gcc Feb 14 21:03:19 Oh, I msiread that Feb 14 21:03:22 I need others :D Feb 14 21:03:29 Foigures you'd be one I have hahaha Feb 14 21:04:06 I need to finish setting up triple targets for rust and I really don't want to build devices I don't have if I can avoid it Feb 14 21:04:16 lol Feb 14 21:04:25 I see, the SDK used by OpenWrt Feb 14 21:04:28 and realized I don't actually have to.. I just have go get poeple with other devices to ook Feb 14 21:04:55 nbd: in that case I have no idea Feb 14 21:05:39 should be a variation of *-openwrt-linux-musl*, but it's the *'s I need to fill in Feb 14 21:08:42 ah, i found it out Feb 14 21:09:09 i use quilt refresh outside of the build tree and my .quiltrc includes -d in the diff opts Feb 14 21:14:00 build #531 of bcm63xx/generic is complete: Failure [failed pkgbuild] Build details are at http://buildbot.openwrt.org/master/images/builders/bcm63xx%2Fgeneric/builds/531 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 21:14:00 , Felix Fietkau , Hauke Mehrtens Feb 14 21:36:54 build #539 of bcm47xx/legacy is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/bcm47xx%2Flegacy/builds/539 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 21:36:54 , Felix Fietkau , Hauke Mehrtens Feb 14 21:41:02 build #773 of ramips/rt288x is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/ramips%2Frt288x/builds/773 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 21:41:02 , Felix Fietkau , Hauke Mehrtens Feb 14 21:41:59 build #833 of ath25/generic is complete: Failure [failed pkgbuild] Build details are at http://buildbot.openwrt.org/master/images/builders/ath25%2Fgeneric/builds/833 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 21:41:59 , Felix Fietkau , Hauke Mehrtens Feb 14 21:49:54 build #293 of mpc85xx/p1010 is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/mpc85xx%2Fp1010/builds/293 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 21:49:54 , Felix Fietkau , Hauke Mehrtens Feb 14 21:51:46 build #760 of kirkwood/generic is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/kirkwood%2Fgeneric/builds/760 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 21:51:46 , Felix Fietkau , Hauke Mehrtens Feb 14 21:54:36 build #753 of x86/geode is complete: Failure [failed images] Build details are at http://buildbot.openwrt.org/master/images/builders/x86%2Fgeode/builds/753 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser , Feb 14 21:54:37 Felix Fietkau , Hauke Mehrtens Feb 14 21:58:01 build #826 of oxnas/ox820 is complete: Failure [failed pkginstall] Build details are at http://buildbot.openwrt.org/master/images/builders/oxnas%2Fox820/builds/826 blamelist: Rosen Penev , Paul Spooren , John Audia , Adrian Schmutzler , Paul Fertser Feb 14 21:58:01