**** BEGIN LOGGING AT Fri Mar 12 03:00:02 2021 Mar 12 05:13:07 Ooo.. It's like a giant game of chicken! Suicide Linux - https://qntm.org/suicide Mar 12 05:14:41 Esp since WSL auto-mounts the Windows NTFS drives as /mnt/c Mar 12 08:33:00 whut? my bash doesn't "correct" my typing Mar 12 08:35:59 And the day it starts doing it, I'll change my shell. Mar 12 08:38:09 the intarwebs suggest this is a zsh optional feature Mar 12 08:48:40 ENABLE_CORRECTION="true" Mar 12 08:50:36 What a great idea! Mar 12 08:51:36 yeah, typin ls and you obviously mean sl ;) Mar 12 08:56:11 And if you don't have sl, it will install the package for you. :P Mar 12 08:58:40 nbd: ping, can you check /query from me 3 days ago? Mar 12 09:28:47 I have a question relating to the git submissions. Do I really have to use my real name and email address? Mar 12 09:30:07 "signed by imaginary people" is not exactly reassuring, wouldn't it? :) Mar 12 09:30:15 olmari: true, I suppose Mar 12 09:30:16 its a weird question to ask here Mar 12 09:30:50 if youre going to use fake name then you probably wouldnt even be asking this Mar 12 09:31:06 grift: I guess I will use my real name then, I suppose Mar 12 09:31:07 its not like the community does a background check Mar 12 09:31:48 either that or just set up a fake name and email but asking here whether its allowed seems weird Mar 12 09:32:02 grift: well, I was just looking through https://openwrt.org/submitting-patches Mar 12 09:32:12 grift: Actually, it's not a weird question at all. In the Linux kernel, for example… https://marc.info/?l=linux-crypto-vger&m=141795521524440&w=4 Mar 12 09:32:30 Or in the tree itself… https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/log/?qt=author&q=George+Spelvin Mar 12 09:33:16 hmm, maybe, i dont see it like that though, either you do it or you dont. its not like anyone will ever find out if you do right anyway Mar 12 09:33:21 mm.. openwrt specifically is still.. well... as far as I know an hobby to everyone involved, as in no one is getting paid to work with it, correct me if I'm wrong... and for broad general using actual name is just logical Mar 12 09:34:05 For context: https://en.wikipedia.org/wiki/George_Spelvin Mar 12 09:34:39 And like implied, rarely there are actual checks nor desires for one either Mar 12 09:35:14 rsalvaterra: lol. 'got out of use since a porn actress used it as her on-stage name' Mar 12 09:35:31 openwrt requires your real name afaik Mar 12 09:35:45 Festivenari: what's the issue with using your real name here? Mar 12 09:35:59 yes ofcourse, it wouldnt make sense to say otherwise Mar 12 09:36:46 but practically its not required because practically if you do it right then no one will even question it Mar 12 09:37:12 Borromini: oh, well, hmm... I guess I will Mar 12 09:38:10 Festivenari: Think it this way… if you're contributing, your contributions are unquestionably attributed to you. Mar 12 09:38:26 probably should register this nick on freenode, actually Mar 12 09:38:58 Festivenari: if you're in IT it also helps potential employers to assess your previous work, when looking for a job Mar 12 09:39:03 I suppose so Mar 12 09:39:10 anonimity is just a thing Mar 12 09:39:15 if you say you're a kernel dev e.g. but can't link to any commits... Mar 12 09:40:37 https://daniel.haxx.se/blog/2021/02/19/i-will-slaughter-you/ Mar 12 09:40:41 And in broad general an identifiable nickname can be as effective as realname, but no one believes "Bugs Bunny" etc Mar 12 09:41:08 agreed Mar 12 09:41:09 Obviously still not same thing, but somewhat related Mar 12 09:41:59 but you should not ask whether its allowed and expect people to say yes Mar 12 09:42:06 just do it or dont do it Mar 12 09:50:36 stintel: i wonder who 'Dick ridin' Dave' is Mar 12 09:50:41 sounds like a nice nickname. Mar 12 09:50:59 disclaimer: i am not gaybashing. Mar 12 09:58:19 Alright, guys, let's do this… https://lore.kernel.org/lkml/20210312095545.53765-1-rsalvaterra@gmail.com/ Mar 12 10:00:17 neat, i didn't know you could tell jffs2 what kind of compression to use Mar 12 10:01:02 Borromini: You can. When this is accepted upstream, I'll take care of the OpenWrt (and fstools) part… ;) Mar 12 10:02:32 ;) Mar 12 10:02:57 what's the compression gain? Mar 12 10:03:55 Expect a compression *loss*, not gain. The point is to reduce the number o compression algorithms in the kernel to the least possible (ideally one which is good for everything, and the best candidate at this time is zstd). Mar 12 10:04:13 ok Mar 12 10:04:36 Compression algorithms are relatively large beasts. Having just one is beneficial. Mar 12 10:05:14 ok Mar 12 10:05:31 does openwrt include multiple by default? sorry for the basic questions. Mar 12 10:06:23 Yes, it does, especially for jffs2. It even includes lzma, which is not even part of the kernel proper. Just take a look at 530-jffs2_make_lzma_available.patch. Mar 12 10:06:34 ok Mar 12 10:07:01 It's all the sorts of insane I can imagine, and most likely lot of them I can't. Mar 12 10:08:36 Granted, it's an *old* patch… forward-ported since the dawn of days, I wager. Mar 12 10:10:25 lzma has been around for a while Mar 12 10:11:39 The algorithm, yes. But there's no upstream kernel support for lzma (and it doesn't make sense, nowadays). Mar 12 10:12:04 as in newer and more efficient algorithms have popped up? Mar 12 10:12:33 Borromini: Exactly. Case in point, zstd. :) Mar 12 10:12:35 guys, i can overrule dts properties right? if a dtsi has wifi0,0 blabla i can do &wifi0,0 in my top dts to override those settings? Mar 12 10:21:27 Borromini: that should work. Mar 12 10:24:02 The solution is to replace jffs2 with f2fs. Mar 12 10:24:11 * mangix hides Mar 12 10:25:05 =) Mar 12 10:25:16 I believe that also has problems with small sizes Mar 12 10:25:31 mangix: Dude! :P Mar 12 10:25:33 svanheule: thanks. would this work? https://paste.debian.net/plain/1188995 Mar 12 10:26:01 thing is the eeprom property used not to be in the dtsi. Mar 12 10:26:13 erm sorry the @ should be an & Mar 12 10:26:19 Borromini: no, that won't work (even with &) Mar 12 10:26:33 & can be used with a label Mar 12 10:26:46 i.e. "node_label: node {...}" Mar 12 10:27:24 alternatively, you can specify the full path: /pci_whatever/wifi@0,0 {...} Mar 12 10:27:28 For those who might think f2fs is a good solution, remember f2fs assumes wear leveling at the lower layer. And don't get me started on the (space) overhead. :P Mar 12 10:27:47 svanheule: ok so i'd need to strip it from the dtsi again then Mar 12 10:28:58 Borromini: is this a new DTSI, or an existing one? Mar 12 10:29:30 rsalvaterra: mtd-utils might need patching as well Mar 12 10:29:34 svanheule: it's an existing one, mt7621_netgear_sercomm_chj.dtsi Mar 12 10:29:55 It's somewhat strange that LZMA is used Mar 12 10:30:27 mangix: s/strange/insane/ Mar 12 10:30:38 svanheule: commit 0cf889db00 has shuffled some stuff around. Mar 12 10:31:08 rsalvaterra: how so? Mar 12 10:31:27 Borromini: in the DTS, specify the node in the same way as here https://github.com/openwrt/openwrt/blob/master/target/linux/ramips/dts/mt7621_netgear_sercomm_chj.dtsi#L80 Mar 12 10:31:28 svanheule: there the wifi nodes do overrule the eeprom property, so i'm a bit confused Mar 12 10:31:37 mangix: Have you seen the size of the lzma patch? Mar 12 10:31:57 Borromini: so "&pcieX { wifi@0,0 { ... } }" Mar 12 10:32:02 I have not. Not on a computer ATM Mar 12 10:32:15 svanheule: ok, so i overrule the complete node in the top dts? Mar 12 10:33:01 mangix: It's over 140 kiB. In a *single* patch. Mar 12 10:33:15 Borromini: well, I think that's just specifiying where the node is located in the tree, so that (property) overrides happen in the right place Mar 12 10:33:46 rsalvaterra: got a link? Mar 12 10:34:34 mangix: https://git.openwrt.org/?p=openwrt/openwrt.git;a=blob;f=target/linux/generic/pending-5.10/530-jffs2_make_lzma_available.patch;h=1bccb30a69f69ff8e2e70a747b1f3b25d1079c8f;hb=HEAD Mar 12 10:35:13 Borromini: mt7621.dtsi is where those pcieN labels are defined, that netgear DTSI is already complementing/overriding it. Mar 12 10:35:18 svanheule: yeah Mar 12 10:35:27 so i'm overriding the override huh :) Mar 12 10:36:11 rsalvaterra: huh. Doesn't the kernel already include LZMA? Mar 12 10:36:19 mangix: No. Mar 12 10:36:34 weird Mar 12 10:36:46 It includes xz, not lzma. Mar 12 10:37:04 Well then Mar 12 10:37:24 xz == LZMA2 Mar 12 10:37:26 And I think it's only for kernel/initramfs decompression. Mar 12 10:37:33 Sorry, no compression. Mar 12 10:37:53 interesting... Mar 12 10:38:15 https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/lib/xz/Kconfig Mar 12 10:39:07 svanheule: so this would be the right thing(tm) to do? https://paste.debian.net/1188997/ Mar 12 10:39:23 Well, zstd is definitely the way forward Mar 12 10:39:36 Lower CPU and RAM Mar 12 10:39:58 mangix: Yep, that's why I went with zstd. :) Mar 12 10:42:29 Borromini: Hm, but what about if you are working under NDAs and such? Mar 12 10:43:01 (oh this one should be to "thread" about working as kernel dev, but can not link any work) Mar 12 10:43:15 Pepe: concerning the e-mail/real name stuff? i don't think that's what Festivenari is doing. he has a pile of old Broadcom stuff he's playing with Mar 12 10:43:17 Borromini: AFAICT only the mediatek,mtd-eeprom properties are different, so you don't have to specify the other ones. But then information is really scattered acros files, so reviewers may not like that Mar 12 10:43:57 svanheule: ok. as you can see from commit 0cf889db00 that's how it used to be (just the eeprom property). but if this works for this guy i'll send in the patch with the full nodes Mar 12 10:43:59 Borromini: yep, you're correct. I do have a pile of broadcom-based routers (luckily a few are already supported by OpenWRT) Mar 12 10:44:23 a few extra lines that help readability won't annoy most people i reckon, like you say Mar 12 10:45:18 lipnitsk: That's a patch from LEDE. Have you tested MT7621 without WEAK_REORDERING_BEYOND_LLSC in a recent kernel? Mar 12 10:47:09 Borromini: I must admit that it is a shame that broadcom is not willing to open-source their drivers Mar 12 10:47:21 I have to wonder how many vulnerabilities there are in their drivers Mar 12 10:49:54 Festivenari: it is, but most big companies seem to think security by obscurity is a good thing. Mar 12 11:32:07 lipnitsk: This might be related: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=42344113ba7a1ed7b5654cd5270af0d5698d8521 Mar 12 11:41:51 svanheule: thanks for your help. another happy mt7621 customer :^) Mar 12 11:46:41 Borromini: np :-) Mar 12 12:00:34 I'm having a pretty strangy problem with a custom build of openwrt Mar 12 12:00:43 I have xfrm-interface selected Mar 12 12:01:03 johnf@johnf-desktop:~/tmp/build/openwrt-tl842/openwrt$ grep -i xfrm .config Mar 12 12:01:06 CONFIG_PACKAGE_kmod-xfrm-interface=m Mar 12 12:01:42 but when I look for the resulting file it doesn't exist: Mar 12 12:01:44 find ./ -name xfrm-interface.ko Mar 12 12:01:47 this returns nothing Mar 12 12:02:06 (even if I use an _ and not a dash) Mar 12 12:02:17 I get a .ipkg file, but it contains no binary Mar 12 12:02:24 I see it's building correctly on snapshot Mar 12 12:02:40 how can I see why this module isn't getting built? Mar 12 12:02:47 johnf: did you install the package? Mar 12 12:02:57 =m isn't included in the image =y is Mar 12 12:03:34 johnf: maybe it's one of those kernel things that can only be included statically Mar 12 12:03:40 it's not Mar 12 12:03:56 https://gist.github.com/stintel/5c9c78e177dc372003071c7ecbe60764 Mar 12 12:05:21 nah, if you grab the lastest matching kernel from snapshot it's clean Mar 12 12:05:35 ./lib/modules/5.4.102/xfrm_interface.ko Mar 12 12:05:47 it contains that once you extract it's data file Mar 12 12:05:58 this is working on other builds that I have Mar 12 12:07:03 johnf: make V=s on the package in your buildroot? Mar 12 12:07:19 you can do that for a kmod? Mar 12 12:07:37 ok, I made the kernel clean and I'm doing a V=s Mar 12 12:07:46 teeing into a soon to be large log file Mar 12 12:08:12 I didn't think I could compile a kmod package like that, I thought that was more packaging the result of the full kernel build Mar 12 12:08:35 I love compiling on my personal computer Mar 12 12:08:44 it's like, 15 times faster than the build server Mar 12 12:08:49 :P Mar 12 12:09:02 make package/kmod-xfrm-interface/compile should just work i think Mar 12 12:09:25 the Ryzen 7 3700X + NVMe is alot faster than a really stale Xeon and overburdened shared storage ;) Mar 12 12:09:35 hehe Mar 12 12:09:41 Borromini: wow, ok, wish I'd known Mar 12 12:09:51 i'd like a 5800X but my 1800X still works, so... :P Mar 12 12:10:10 I'm also good for now with dual E5-2673 v4 :P Mar 12 12:10:16 although the clock speed is a tad bit low Mar 12 12:10:19 well, I got this and it replace a i7-4771 Mar 12 12:10:28 I like buying higher end procs Mar 12 12:10:31 but I run them forever Mar 12 12:10:41 wow, very very nice stintel Mar 12 12:10:48 aah. Intel's '4 cores oughta be enough for anybody' schtick Mar 12 12:11:12 johnf: 2nd hand HP Z840, 40 cores + 256GB RAM, twas kind of a steal Mar 12 12:12:04 wow Mar 12 12:12:04 3k EUR ex VAT Mar 12 12:12:07 :P Mar 12 12:12:51 and since I had just messed up the socket on the mobo of my i7 3930k box I kind of needed a replacement instantly Mar 12 12:13:02 stintel: No POWER9, though. :P Mar 12 12:13:20 so I got in the car and picked it up - friend bought it 2nd hand for AI/ML stuff but the clock speed was to low for his workload Mar 12 12:13:38 rsalvaterra: yeah, the problem with that talos II is the shipping Mar 12 12:13:56 + VAT + import duties Mar 12 12:14:04 stintel: No, the problem is the base price. Shipping is a whole other problem. :P Mar 12 12:14:38 nothing wrong with the base price, try getting similarly spec'd workstation from HP or Dell or so Mar 12 12:14:53 there is absolutely nothing wrong with the base price imo Mar 12 12:15:23 it's simply not for home/personal use Mar 12 12:15:33 get a blackbird then Mar 12 12:16:10 The Blackbird is RAM-channel-castrated. Mar 12 12:16:33 still should be good enough for home/personal/hobby use ;) Mar 12 12:16:41 but yes, it's true, it's the reason I don't consider it Mar 12 12:16:49 stintel: I recently got a dual E5-2630v4 with SSD storage at work, I should totally try building OpenWrt on that machine... Mar 12 12:17:26 stintel: Yeah, me neither… it's a shame, as I *really* like the POWER arch… :( Mar 12 12:18:13 rsalvaterra: I would love to have the Talos II but the premium to get in EU + shitty performance compared to recent x86 ... Mar 12 12:18:24 if I'd make twice what I'm making now, nobrainer Mar 12 12:19:03 but it'd cost >10% of my yearly turnover, it's a bit hard to sell, even if it's to myself Mar 12 12:19:20 Unfortunatelly, you can get a top Threadripper for the price of a dual-CPU Talos II… and it will run rings around it. Mar 12 12:19:39 also by now it's old tech Mar 12 12:19:50 I'll reconsider when Raptor offers a Power10 Mar 12 12:19:55 hmm, ok Mar 12 12:20:02 https://openwrt.org/docs/guide-developer/single.package#kernel_modules Mar 12 12:20:13 so it looks like you can't build a single kmod like that Mar 12 12:21:16 my target/linux/compile contains no references for xfrm_interface Mar 12 12:21:24 johnf: sorry Mar 12 12:21:35 make package/kernel/xfrm-interface/{clean,compile} V=s Mar 12 12:21:36 try that Mar 12 12:21:47 ok, sure, just a moment Mar 12 12:22:19 although there's no subdir in package/kernel so that might not work either Mar 12 12:22:20 make[1]: *** No rule to make target 'package/kernel/xfrm-interface/clean'. Stop. Mar 12 12:22:25 nope Mar 12 12:22:55 ok, so I could give you the full build logs probably Mar 12 12:22:56 CC [M] net/ipv4/xfrm4_tunnel.o Mar 12 12:23:08 I do see other xfrm packages getting built Mar 12 12:23:36 but nothing for xfrm_interface Mar 12 12:23:42 this is very confusing Mar 12 12:24:11 svanheule: that reminds me I should really get my OpenWrt build + deploy pipeline back up and running Mar 12 12:24:57 I had staged deployments for my Raspberry Pi Zero W's - 1 testing, if it came back after sysupgrade, all others were auto sysupgraded too Mar 12 12:25:06 can do the same for my ERL as I have two Mar 12 12:26:03 (and only one of them running in production) Mar 12 12:26:27 there's a kernel config, that get's generated from the settings in .config Mar 12 12:26:30 right? Mar 12 12:26:35 yes Mar 12 12:26:50 tmp/.kconfig-target-subtarget or something Mar 12 12:27:18 tmp/.kconfig-target_subtarget Mar 12 12:27:58 CONFIG_XFRM=y Mar 12 12:27:59 # CONFIG_XFRM_INTERFACE is not set Mar 12 12:28:05 ok, so this is the root cause of my issue Mar 12 12:28:20 now look, I'm not going to lie to you, I may have done a few things to this kernel Mar 12 12:28:23 to free up ram Mar 12 12:28:25 johnf: maybe rm -rf tmp/ and build again, just to be sure it's not a temp glitch Mar 12 12:28:46 building Mar 12 12:28:51 stintel: would you mind sharing the code for that? Mar 12 12:29:21 really appreciate the help stintel, Borromini Mar 12 12:29:26 Borromini: it was OVH CDS based and I decided it's too hard to maintain my own CDS Mar 12 12:30:23 other option is to use public gitlab and local runner but I don't trust cloud-based things enough to add an SSH key that allows root login to my OpenWrt devices Mar 12 12:30:26 even if it's only on the runner Mar 12 12:30:45 the runner can be controlled by the cloud, so there's a possibility that can be exploited Mar 12 12:31:23 so a local CI/CD thing is the only acceptable thing for me Mar 12 12:31:41 also I really don't like gitlab, that was my reason for looking at alternatives in the first place Mar 12 12:31:49 * stintel making it hard on himself since 1984 Mar 12 12:32:30 lol Mar 12 12:32:38 ok, it's still building, but I took a peek in tmp Mar 12 12:32:46 same behaviour, that module isn't selected Mar 12 12:34:04 johnf: then I'd suggest to look at the dependencies in the kernel source Mar 12 12:34:16 hmm Mar 12 12:34:16 │ Selects: PACKAGE_kmod-ipsec6 [=n] && PACKAGE_kmod-ipsec4 [=m] │ Mar 12 12:34:25 that's from the menuconfig info Mar 12 12:34:37 stintel: I like GitLab… :P Mar 12 12:34:38 depends on XFRM && IPV6 Mar 12 12:34:43 kmod-ipsec6 isn't selected and I have disabled IPv6 Mar 12 12:34:47 did you disable IPv6? :) Mar 12 12:34:47 fuck Mar 12 12:34:54 yes, of course Mar 12 12:34:56 it's so big Mar 12 12:35:02 yeah, don't do that, it's 2021. there's no valid reason to do so Mar 12 12:35:14 32MB of RAM says there's a reason Mar 12 12:35:25 but I do see what you're saying Mar 12 12:35:42 johnf: No, there isn't. I have IPv6 enabled on a 32 MiB device (AirGrid M2). Mar 12 12:36:06 another option is: drop IPsec and use wireguard Mar 12 12:36:21 I'm seriously considering that, unfortunately I'm having major issues with wireguard on openwrt Mar 12 12:36:28 didn't get to debugging that yet Mar 12 12:36:49 when I can get that fixed I'm even going to stop maintaining strongswan in the packages feed Mar 12 12:37:47 I have like 2000 routers running variants of this image Mar 12 12:37:56 I'm not moving to wireguard any time soon Mar 12 12:37:59 ok, that might be difficult :) Mar 12 12:38:03 though I have used it for some other stuff quite well Mar 12 12:38:19 rsalvaterra: not with strongswan, quagga/frr and some other stuff Mar 12 12:38:27 I have only 4 OpenWrt devices with IPsec, and even that was enough to hold off on migrating to wg Mar 12 12:39:10 hmm Mar 12 12:39:21 spent a lot of time optimizing that setup, and don't fix if not broken etc Mar 12 12:39:34 is there a way I can update that kconfig file without actually building? Mar 12 12:42:00 johnf: if you throw out PPP you should still have quite a bit of wiggle room even with 4/32 Mar 12 12:42:34 it sounds like actually if you disable ipv6 you shouldn't even be seeing xfrm-interface as selectable, if it needs ipv6 Mar 12 12:42:45 johnf: make kernel_menuconfig Mar 12 12:43:48 need PPP for mobile Mar 12 12:44:01 it's giant too Mar 12 12:44:07 I'm amply aware :/ Mar 12 12:44:34 man, I was so happy when I finally got this thing built Mar 12 12:45:15 johnf: can you do vti instead of xfrm? Mar 12 12:45:41 xfrm is nicer, sure, but maybe it's an acceptable solution Mar 12 12:45:41 already have 300 XFRM devices deployed, and my cisco core needs this XFRM setup Mar 12 12:45:57 it's these ancient routers that are proving particularly challenging Mar 12 12:46:01 and I have rather a lot of them Mar 12 12:46:01 hmmm Mar 12 12:46:13 I believe one side can use XFRM while other side can use VTI no problem Mar 12 12:46:35 actually I'm sure it can, I'm doing it Mar 12 12:46:59 hmm Mar 12 12:47:19 I really want to just get this module built Mar 12 12:47:28 I've already put a lot of effort into this current config Mar 12 12:47:33 don't want to fragment things further Mar 12 12:47:38 but we'll see what options I have Mar 12 12:47:59 where did you get that DEPENDS info above stintel? Mar 12 12:48:02 well I guess you'd have to look into the possibility of dropping the IPv6 requirement for the xfrm_interface module, but that might be even harder :) Mar 12 12:48:08 lol Mar 12 12:48:12 johnf: linux.git net/xfrm/Kconfig Mar 12 12:48:12 that's not in the cards Mar 12 12:48:22 ah Mar 12 12:48:31 ok, so I've brought kernel IPv6 back Mar 12 12:48:35 but not userspace Mar 12 12:48:39 I will add that dependency to the kmod Mar 12 12:48:39 but when I'm building Mar 12 12:48:49 saves other people the trouble in the future :) Mar 12 12:48:53 that module appears to still not be selected Mar 12 12:49:37 XFRM is definately selected, builds as well Mar 12 12:50:31 CONFIG_IPV6=m Mar 12 12:50:39 global v6 is present (as a module, is that a thing?) Mar 12 12:52:34 ah sweet, it is built Mar 12 12:52:43 https://git.openwrt.org/d6845655 Mar 12 12:52:46 and I have an xfrm_interface file Mar 12 12:52:55 oops Mar 12 12:52:58 hmm, no such commit Mar 12 12:53:01 https://git.openwrt.org/f813fdf3 Mar 12 12:53:05 mea culpa Mar 12 12:53:30 untested, someone please do :) Mar 12 12:55:24 with pleasure, give me a few minutes Mar 12 12:55:30 thanks Mar 12 12:56:20 cloning Mar 12 12:56:44 actually, while we're talking about this stuff, I had a question Mar 12 12:56:52 at one point I wanted to make a change to quagga Mar 12 12:57:03 and I couldn't find some of the packages on github Mar 12 12:57:19 I think it was the net repo that comes out of feeds Mar 12 12:57:23 but I could be remembering wrong Mar 12 12:57:36 a few of the packages I'm using have been orphaned and removed Mar 12 12:57:46 I'd like to get some of them back in and support them Mar 12 12:58:38 https://github.com/openwrt-routing/packages.git Mar 12 12:58:42 that's where quagga is Mar 12 12:59:06 as for orphaned packages, feel free to do a PR to add them to the packages feed Mar 12 12:59:23 but you should be willing to maintain them Mar 12 12:59:43 yes, I am Mar 12 13:00:31 very happy to give back in the small ways that I can Mar 12 13:01:39 glad to hear that :) Mar 12 13:04:36 ok, so https://git.openwrt.org/feed/routing.git is then available as openwrt-routing Mar 12 13:04:44 I was unable to make the connection Mar 12 13:04:58 oh, I see Mar 12 13:05:04 openwrt-routing is a top level entity Mar 12 13:05:08 as opposed to like Mar 12 13:05:31 https://github.com/openwrt/packages Mar 12 13:05:52 well that explains why searching all of https://github.com/openwrt/ for quagga didn't return anything even moderately modern Mar 12 13:06:00 =) Mar 12 13:06:03 thanks for that Mar 12 13:06:14 I used to be a quagga user, but I migrated to bird a long time ago Mar 12 13:06:28 I moved to frr with these upgrades Mar 12 13:06:28 now I still have migration to bird2 on my todo list, only partially completed Mar 12 13:06:44 but these ancient tiny ones can't deal with frrs bloat Mar 12 13:06:50 vtysh is insane Mar 12 13:06:54 do you happen to have a frr with full IPv6 BGP feed somewhere? I'd be interested in the mem usage Mar 12 13:06:58 it takes like 2MB of RAM to load Mar 12 13:07:13 hmm, no I don't Mar 12 13:07:21 and I don't have access to a full v6 table Mar 12 13:07:36 I could potentially do something like that with v4 Mar 12 13:07:40 though I winder Mar 12 13:07:41 wonder Mar 12 13:08:07 http://www.routeviews.org/routeviews/index.php/collectors/ Mar 12 13:08:49 root 7284 0.5 0.8 33328 32932 ? S Feb20 168:01 /usr/sbin/bird -f -c /etc/bird.conf -P /var/run/bird.pid Mar 12 13:09:07 32MB including full IPv6 BGP :P Mar 12 13:10:25 lol Mar 12 13:11:59 show me avc denials that illustrate your issue Mar 12 13:12:03 wrong chan Mar 12 13:17:37 this is new - daemon.err modprobe: failed to find a module named act_ipt - and I can't find the cause Mar 12 13:21:55 stintel: builds clean, I have the resulting package: Mar 12 13:21:56 ./targets/ath79/generic/packages/kmod-xfrm-interface_5.4.102-1_mips_24kc.ipk Mar 12 13:22:11 how can I add a tested by or whatever Mar 12 13:22:36 you can write it here and I'll add it Mar 12 13:23:58 Signed-off-by: John Marrett Mar 12 13:24:19 https://github.com/openwrt/openwrt/commit/4531bc4267c7701e3615a806a0fde8212076699a Mar 12 13:24:26 here's one of my ancient commits ;) Mar 12 13:24:39 johnf: cheers! Mar 12 13:24:57 oh wow, my C7 one is there too, I thought that someone else had got the commit in before I did Mar 12 13:24:59 johnf: mind testing if disabling ipv6 in menuconfig hides the kmod-xfrm-interface entirely in menucofnig ? Mar 12 13:25:22 well I could quickly do that myself too Mar 12 13:26:49 no problem man Mar 12 13:26:50 not at all Mar 12 13:26:52 just a second Mar 12 13:27:15 though if it does, it's deceptive Mar 12 13:27:51 ok, so it does, but that's not really the right thing to do Mar 12 13:28:00 the real dependancy is actually on Mar 12 13:29:28 you prefer enabling kmod-xfrm-interface enables IPV6? Mar 12 13:29:34 no no Mar 12 13:29:37 just give me a second Mar 12 13:29:40 or is there a kernel-specific symbol for IPV6 Mar 12 13:29:41 ok :) Mar 12 13:29:43 sorry, trying to find the right way to express this Mar 12 13:29:47 yes, the second one Mar 12 13:29:57 │ Symbol: IPV6 [=n] │ Mar 12 13:30:03 patience is not always my strong suit :P Mar 12 13:30:15 this, in make kernel_menuconfig Mar 12 13:30:18 is what actually is needed Mar 12 13:30:23 in the kernel I just built Mar 12 13:30:23 but that's what I did, no ? Mar 12 13:30:27 I have kernel IPv6 Mar 12 13:30:33 I think you have usermode IPv6 Mar 12 13:30:50 global build settings "Enable IPv6 support in packages" Mar 12 13:31:00 is that what you are looking at, or the kernel setting? Mar 12 13:31:34 johnf: if you check the help for that: │ Enables IPv6 support in kernel (builtin) and packages. │ Mar 12 13:32:32 so, in the build I just did Mar 12 13:32:41 I have that off in make menuconfig Mar 12 13:32:53 but I have turned on, in make kernel_menuconfig, kernel IPv6 support Mar 12 13:32:53 oh Mar 12 13:32:56 this actually works Mar 12 13:33:04 well, I think it does, I'll know for real quite soon Mar 12 13:33:12 I did get a .ko file Mar 12 13:33:18 but I don't know if you can express that dependancy Mar 12 13:33:30 ok, so my change would actually mess up what you're doing Mar 12 13:33:43 I sort of think no one should do what I'm doing Mar 12 13:33:47 :D Mar 12 13:44:55 aha! it's bleeding edge sqm-scripts thing. Mar 12 13:46:37 ldir: The sqm-scripts package should be split… :/ Mar 12 13:46:46