**** BEGIN LOGGING AT Sun Mar 14 02:59:56 2021
Mar 14 03:03:51 <mangix> no idea
Mar 14 03:04:07 <Grommish> Yeah, I'll continue to dig :) Thanks!
Mar 14 03:23:02 <owrt-snap-builds> Build [#3](https://buildbot.openwrt.org/master/images/#builders/16/builds/3) of `ath79/mikrotik` completed successfully.
Mar 14 03:27:02 <owrt-2102-builds> Build [#2](https://buildbot.openwrt.org/openwrt-21.02/images/#builders/27/builds/2) of `apm821xx/nand` completed successfully.
Mar 14 03:42:22 <Grommish> Oh ho!  I've now got an eth5 showing up
Mar 14 03:43:29 <Grommish> I mean, doesn't do anything, but it is showing up at least
Mar 14 03:47:23 <owrt-snap-builds> Build [#2](https://buildbot.openwrt.org/master/images/#builders/68/builds/2) of `at91/sama5` completed successfully.
Mar 14 04:08:19 <owrt-snap-builds> Build [#2](https://buildbot.openwrt.org/master/images/#builders/39/builds/2) of `ath79/nand` completed successfully.
Mar 14 05:37:57 <owrt-snap-builds> Build [#2](https://buildbot.openwrt.org/master/images/#builders/48/builds/2) of `bcm47xx/generic` completed successfully.
Mar 14 05:55:25 >KGB-0< https://tests.reproducible-builds.org/openwrt/openwrt_lantiq.html has been updated. (98.2% images and 98.3% packages reproducible in our current test framework.)
Mar 14 07:00:09 <owrt-2102-builds> Build [#2](https://buildbot.openwrt.org/openwrt-21.02/images/#builders/44/builds/2) of `bcm47xx/legacy` completed successfully.
Mar 14 10:08:31 <rsalvaterra> lipnitsk: Related patch series: https://lore.kernel.org/lkml/1358183124-28461-1-git-send-email-mark.rutland@arm.com/
Mar 14 10:09:41 <rsalvaterra> mangix: Nope. I just reverted to the current in-tree musl for my RM2100… :/
Mar 14 11:58:45 <grift> anything wrong with mailing list?
Mar 14 11:58:51 <grift> Mar 14 12:48:43 agnus postfix/smtp[1328787]: connect to util-01.infra.openwrt.org[2001:470:1f0b:a42::1]:25: Connection refused
Mar 14 12:00:15 <grift> k never mind
Mar 14 12:01:07 <stintel> grift: ML is not on util-01
Mar 14 12:01:31 <grift> i guess dns issue then
Mar 14 12:01:42 <grift> but seems to have been solved as my mail arrived
Mar 14 12:03:00 <stintel> k
Mar 14 15:27:17 <stintel> so I'm looking into the wireguard issue I'm having again, decided to try it on my backup router, it shows the same problem as my main router. interface comes up during boot, but if I then do ifdown wg0; ifup wg0, it's fubar
Mar 14 15:27:19 <stintel> https://gist.github.com/stintel/5c34aca662344084e7b61df1c7c6b64d
Mar 14 15:36:42 <mangix> stintel: try using LuCI
Mar 14 15:37:09 <owrt-snap-builds> Build [#4](https://buildbot.openwrt.org/master/images/#builders/61/builds/4) of `arc770/generic` failed.
Mar 14 15:48:38 <stintel> mangix: how is that going to help anything?
Mar 14 15:48:55 <stintel> mangix: the connection works fine on boot
Mar 14 16:18:38 <grift> stintel unable to produce that here fwiw
Mar 14 17:03:37 <owrt-snap-builds> Build [#6](https://buildbot.openwrt.org/master/images/#builders/26/builds/6) of `apm821xx/sata` failed.
Mar 14 17:06:09 <ldir> guidosarducci: Hi Tony, when I get home from work and after I've successfully installed your kernel switch on my APU 2 I'll merge your PR.
Mar 14 17:33:48 <stintel> grift: I have 2 other devices that don't show this problem either
Mar 14 17:33:58 <stintel> I'm starting to suspect it might be related to using macvlan
Mar 14 17:34:43 <stintel> my ISP offers me 2 static DHCP leases, I'm using macvlan so I can float them between both routers
Mar 14 17:36:07 <stintel> the setup is explained here if anyone cares: https://stijn.tintel.eu/blog/2017/09/28/building-redundant-router-setup-open-source-software-part-2
Mar 14 17:36:15 <stintel> well the macvlan specific part
Mar 14 17:43:49 <lipnitsk> rsalvaterra: thanks! Do you know which drivers in ramips builds use those timer broadcasts? If I try to upstream that change I'd like to understand
Mar 14 18:02:24 <rsalvaterra> lipnitsk: No idea. But do note that MIPS implements tick_broadcast(), so it doesn't matter, even if no driver uses them: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/arch/mips/kernel/smp.c?h=linux-5.4.y#n697
Mar 14 18:11:37 <rsalvaterra> lipnitsk: This is what I did in my tree: https://github.com/rsalvaterra/openwrt/commit/fc857f3589bff9d4ed8ed91c589c6ecfe15afc53
Mar 14 18:13:20 <lipnitsk> Yeah that is a good way to do it
Mar 14 18:13:31 <lipnitsk> Feel free to resubmit or drive those changes
Mar 14 18:13:51 <lipnitsk> rsalvaterra: ^
Mar 14 18:14:39 <lipnitsk> There may be other obsolete patches but a little at a time is probably easier to get reviewed and integrated/tested
Mar 14 18:23:07 <rsalvaterra> lipnitsk: According to blogic, the problem WEAK_REORDERING_BEYOND_LLSC fixed (paper over?) manifested itself in HIGHMEM systems. Unfortunately, my RM2100 only has 128 MiB of RAM. That said, it's completely stable without selecting WEAK_REORDERING_BEYOND_LLSC.
Mar 14 18:23:19 <rsalvaterra> *papered over
Mar 14 18:24:39 <rsalvaterra> If someone with a HIGHMEM MT7621 system could test, it would be great.
Mar 14 18:26:20 <rsalvaterra> It would probably be something with 512 MiB of RAM (or maybe even 256 MiB, I don't know).
Mar 14 18:26:39 * Borromini has only 128 MiB MT7621 devices
Mar 14 18:27:12 <rsalvaterra> The EdgeRouter 10X has 512 MiB…
Mar 14 18:27:20 <lipnitsk> My er x has 256 so still no high mem
Mar 14 18:27:41 <rsalvaterra> lipnitsk: Bummer… :/
Mar 14 18:28:12 <lipnitsk> Grommish was working on enabling a 10x device I think he got it booting
Mar 14 18:28:30 <Grommish> Hey
Mar 14 18:28:36 <Grommish> Oh it boots fine
Mar 14 18:28:44 <Grommish> Whatcha need?
Mar 14 18:28:47 <rsalvaterra> Yeah, but he was still fighting the switch configuration.
Mar 14 18:29:02 <rsalvaterra> Grommish: Oh, you did it? :)
Mar 14 18:29:03 <Grommish> Ports 0-4 works fine, its 5-9 that are screwy
Mar 14 18:29:12 <Grommish> althought I"ve got eth5 showing now
Mar 14 18:29:35 <Grommish> [    8.924312] mt7530 mdio-bus:1f eth5 (uninitialized): PHY [mdio-bus:00] driver [Generic PHY] (irq=POLL)
Mar 14 18:30:04 <rsalvaterra> Grommish: If possible, please, give this commit a spin on your ER10X… :) https://github.com/rsalvaterra/openwrt/commit/fc857f3589bff9d4ed8ed91c589c6ecfe15afc53
Mar 14 18:31:17 <rsalvaterra> But before that, can I see the contents if its /proc/meminfo?
Mar 14 18:31:42 <Grommish> sure..
Mar 14 18:32:12 <Grommish> rsalvaterra: v
Mar 14 18:32:14 <Grommish> er https://gist.github.com/Grommish/9792056aff6375eea5a7be23ab1e3243
Mar 14 18:32:41 <Grommish> Ok.. I"m having a moment here..  best way to apply this patch? patch -p1 ?
Mar 14 18:33:05 <rsalvaterra> We have {High,Low}/{Total,Free}, great!
Mar 14 18:33:07 <Grommish> because you're altering the target/xxx so I can't just plop it into the patches-5.10 dir :)
Mar 14 18:33:31 <rsalvaterra> Oh, I usually just download the commit as a patch and apply with git am. :)
Mar 14 18:33:45 <rsalvaterra> Like this: https://github.com/rsalvaterra/openwrt/commit/fc857f3589bff9d4ed8ed91c589c6ecfe15afc53.patch
Mar 14 18:33:46 <Grommish> I pulled it as a .patch
Mar 14 18:33:53 <rsalvaterra> Right!
Mar 14 18:34:09 <Grommish> Oh.  Dirty index.. on sec
Mar 14 18:34:17 <rsalvaterra> It's a patch which changes patches. Patches all the way down. :)
Mar 14 18:36:10 <Grommish> Ah I see why
Mar 14 18:36:12 <Grommish> IU had to rebase first
Mar 14 18:36:33 <Grommish> Cleaning and building..  few minuntes
Mar 14 18:45:27 <Grommish> Gah.. the rebase means toolchain rebuild haha almost done..
Mar 14 18:45:52 <rsalvaterra> Grommish: Oh, dear… sorry about that… :(
Mar 14 18:46:09 <Grommish> Oh, it's fine.. Just takes a few mins more
Mar 14 18:46:26 <Grommish> I just hadn't realized it
Mar 14 18:46:56 <rsalvaterra> In my case, a full build from scratch takes 150+ minutes. :P
Mar 14 18:46:59 <Grommish> I really want to get this stupid realtek switch working
Mar 14 18:47:45 <Grommish> 5.12 has a staging driver for rtl8367c but I'm not sure if that would work
Mar 14 18:48:04 <rsalvaterra> But I'll probably move to my i7-4770R soon… that will make things a bit faster.
Mar 14 18:48:13 <Tapper> Grommish Hi dude.
Mar 14 18:48:19 <Tapper> Grommish Hows things?
Mar 14 18:48:21 <Grommish> Tapper: Hey mate!
Mar 14 18:48:39 <Grommish> rsalvaterra: My baremetal build system is only a i7-3xxxx series
Mar 14 18:48:43 <Tapper> Still haveing truble with that driver for the second switch?
Mar 14 18:49:01 <Grommish> Tapper: Oh yeah, of course :) Wouldn't be me without issues
Mar 14 18:49:14 <Tapper> Grommish lol
Mar 14 18:49:37 <Tapper> Can you find an other router that has the same switch in it?
Mar 14 18:50:00 <Tapper> One were you can look at the DTS or how the driver works?
Mar 14 18:50:05 <Grommish> WSL2 I'm still enamered with though.  There are some, but the way it connects is super wierd
Mar 14 18:50:27 <Grommish> I think it takes the second switch as a whole unit, rather than 5 additional ports..
Mar 14 18:51:07 <Tapper> Dam OEMs and there lack of open code
Mar 14 18:51:31 <Grommish> It's open.. I have the kernel code for the device, but it's super old
Mar 14 18:51:56 <Tapper> all most as bad as having no code then I beat?
Mar 14 18:52:04 <Tapper> bet*
Mar 14 18:52:27 <Grommish> If I want the 3.x kernel, it would be fine :D
Mar 14 18:53:06 <Tapper> lol
Mar 14 18:54:08 <Grommish> rsalvaterra: Booted.. https://gist.github.com/Grommish/288b19600ecbaff4b4ac24834086321a
Mar 14 18:54:15 <rsalvaterra> Grommish: The factory kernel of my RM2100 is also 3.x… :P
Mar 14 18:55:03 <rsalvaterra> Awesome! \o/
Mar 14 18:55:20 <rsalvaterra> Keep it running, though. Need to make sure it's stable. :)
Mar 14 18:55:40 <rsalvaterra> If you don't see random hangs, it's fine.
Mar 14 18:56:51 <Grommish> No network, but I don't know if that's my issue.. or yours
Mar 14 18:57:25 <rsalvaterra> Grommish: This should be totally unrelated to networking…
Mar 14 18:57:25 <Grommish> Lets try this again
Mar 14 19:02:05 <Grommish> Well..  No idea why the network is being screwy, but, hasn't hung yet anyway
Mar 14 19:04:36 <rsalvaterra> Grommish: I was looking at the ER10X… If I understand correctly, the thing has two independent 4-port switches, each connected to an internal Ethernet interface, right?
Mar 14 19:04:47 <Grommish> Let me revert the dts changes and see if that effects anything
Mar 14 19:05:00 <rsalvaterra> (Four external ports, that is.)
Mar 14 19:05:21 <Grommish> From what I can gather, it has Ports 0-4 on the MT7621 board, and then the RTL8367RB switch that connects to the CPU port
Mar 14 19:05:30 <Grommish> which are ports 5-9
Mar 14 19:06:09 <Grommish> Rene__ explained it like this: MT7621 GMAC0 -> RGMII BUS -> MT7530 PORT 6 -> MT7530 SWITCH -> MT7530 PORT 5 -> RGMII2 BUS -> RTL8367RB PORT x -> RTL8367RB switch -> RTL8367RB other ports
Mar 14 19:06:55 <rsalvaterra> WTF…!
Mar 14 19:07:21 <Grommish> POrt 6 is the DSA CPU port
Mar 14 19:08:57 <Grommish> Ok.. So.. the dts changes I made screwed the network up.. I reverted that and am testing just your patch
Mar 14 19:09:45 <Grommish> rsalvaterra: https://gist.github.com/Grommish/ddc6bfac69b0cf078ee548169d928c0c with working network.. I can just let it sit for a bit if you want
Mar 14 19:12:01 <rsalvaterra> Grommish: I just needed to see your meminfo to make sure your system is using HIGHMEM. But yes, let it sit (it should make your system marginally faster).
Mar 14 19:12:45 <Grommish> rsalvaterra: Nice! :)
Mar 14 19:13:55 <Grommish> rsalvaterra: The Archer C6 v1 uses the RTL switch, but doesn't seem to be something I can move over (probably a lack of knowledge on my end)
Mar 14 19:17:48 <Grommish> rsalvaterra: blogic and nbd are all over the code source, but I'm not sure they have the time or patience to help with this hehe
Mar 14 19:19:26 <rsalvaterra> Heh… nbd is mostly busy with mac80211 and mt76 work, both upstream and in OpenWrt… :)
Mar 14 19:20:12 <Grommish> rsalvaterra: *nod*
Mar 14 19:23:38 <rsalvaterra> Speaking of upstream, I am yet to receive any comments on my jffs2 zstd support patch… :P
Mar 14 19:24:35 <rsalvaterra> … but I sent it on Friday, I'd be surprised if someone looked at it during the weekend. :)
Mar 14 19:25:02 <shibboleth> did anyone comment on why it was decided that you now need javascript enabled to simply read forum posts?
Mar 14 19:25:29 <shibboleth> jow said it was likely based on UA, but whoever is in charge of the forum should really reconsider this
Mar 14 19:28:48 <plntyk> UA = ?
Mar 14 19:29:31 <Grommish> User Interface?
Mar 14 19:29:36 <Grommish> I usually see it as UX though
Mar 14 19:30:22 <shibboleth> useragent
Mar 14 19:31:11 <shibboleth> and since firefox quantum you can no longer spoof it (i know, that feature just HAD to go), meaning owrt forums now need js to simply view text
Mar 14 19:37:34 <owrt-snap-builds> Build [#7](https://buildbot.openwrt.org/master/images/#builders/26/builds/7) of `apm821xx/sata` completed successfully.
Mar 14 19:38:28 <stintel> shibboleth: afaik the person maintaining the forum is not on IRC
Mar 14 19:38:44 <shibboleth> anyone care to drop'em a line?
Mar 14 19:39:20 <stintel> send him an email ?
Mar 14 19:40:08 <shibboleth> i suppose it'll carry more weight coming from you? i spoke to jow, apparently it's working in lynx/w3m (due to ua)
Mar 14 19:40:34 <plntyk> hm... firefox 86 with addon https://addons.mozilla.org/en-US/firefox/addon/user-agent-string-switcher/ and browser ID of Fennec and abou:config deactivated JS seem to show forum entries
Mar 14 19:40:43 <plntyk> but some other strings do not work
Mar 14 19:40:50 <stintel> shibboleth: I don't have a problem with the forum, so I don't see why I should be the one sending a mail about yours
Mar 14 19:41:34 <shibboleth> plntyk, yeah, but that doesn't really solve the issue, merely offsets it. you're trading no js for a third-party extension
Mar 14 19:41:51 <shibboleth> re: great suspender-brouhaha
Mar 14 19:48:26 <plntyk> palemoon without JS seems fine too
Mar 14 19:50:25 <shibboleth> so does lynx and w3m. merely offsets the issue
Mar 14 19:54:19 <shibboleth> but re palemon ir proves that js isn't really required for it to work in ff, it's just disqus being di**s
Mar 14 19:54:50 <Borromini> discourse?
Mar 14 19:56:18 <shibboleth> isn't the forum based in disquis?
Mar 14 19:57:49 <plntyk> shibboleth, is attaching a /print at the end of a valid forum url working ?
Mar 14 19:57:49 <Grommish> It's all Hauke's fault ;p
Mar 14 19:58:19 <shibboleth> plntyk, how about them apples. thanks
Mar 14 19:58:34 <Borromini> shibboleth: it's discourse. disqus is some software to enable comments on blogs etc
Mar 14 19:58:40 <shibboleth> right
Mar 14 19:59:20 <plntyk> shibboleth, i found it when browsing through "https://meta.discourse.org/t/make-discourse-play-nice-with-the-wayback-machine/34579/38"
Mar 14 19:59:35 <shibboleth> anyway, requiring js for something that doesn't need js is "dark pattern"-y and i didn't expect owrt forum to join in
Mar 14 19:59:49 <plntyk> there the "crawler" pure html "1996" output is mentioned and after it somebody said /print
Mar 14 19:59:53 <plntyk> view
Mar 14 20:00:47 <plntyk> "JS only" websites - open source to inspect source code and see nothing ^^
Mar 14 20:10:48 <shibboleth> same with gitlab
Mar 14 20:11:57 <shibboleth> yeah, github keep chipping away at stuff, but at least it offers basic view-functionality without the need for js
Mar 14 20:27:22 <Grommish> I lived thru the original HTML 1.2  I'll take a little scripting to make a site more useable
Mar 14 20:27:55 <Grommish> That Koala still haunts my nightmares
Mar 14 20:29:00 <aparcar[m]> does anyone here build openwrt on the new m1 apple chip?
Mar 14 20:46:56 <grift> Grommish: "a little scripting" is not an option
Mar 14 21:27:30 <mangix> rsalvaterra: sergio has a highmem device. email him.
Mar 14 21:28:57 <mangix> aparcar[m]: I've built OpenWrt on my NAS running ARM
Mar 14 21:29:13 <mangix> I remember it working fine albeit slow
Mar 14 21:29:31 <aparcar[m]> year I used a rpi as a buildbot worker
Mar 14 21:29:43 <aparcar[m]> but I was wondering if the m1 is fun for development or rather anying
Mar 14 21:30:49 <hurricos> probably not. It performs great for its power but builds happen faster on server-dedicated HW, and ARM has plenty of that (though none of it cheap)
Mar 14 21:31:26 <aparcar[m]> mhh
Mar 14 21:31:52 <hurricos> I noticed up there how folks were taking 150 minutes for builds from scratch and I'm here looking at 26 minutes on 12 cores hyperthreaded Westmere-EP (that's 32nm, positively ancient)
Mar 14 21:32:54 <aparcar[m]> hurricos: I like to develop while running around, so usually 12 cores are a bit heavy
Mar 14 21:33:43 <hurricos> It's a rackserver X)
Mar 14 21:33:44 <aparcar[m]> mangix: Package 'libcurl', required by 'virtual:world', not found
Mar 14 21:33:48 <hurricos> R510
Mar 14 21:33:54 <aparcar[m]> transmission doesn't want to build...
Mar 14 21:41:27 <swalker> updated openwrt/upstream, https://sdwalker.github.io/uscan/index.html
Mar 14 22:02:43 <grift> aparcar[m]: ssh?
Mar 14 22:02:55 <aparcar[m]> grift: what?
Mar 14 22:03:06 <aparcar[m]> oh fair. Well yea I'll just need to find a quick build server I guess
Mar 14 22:03:06 <grift> developing while "walking around"
Mar 14 22:17:01 <grift> it doesnt have to be a "build server" you mentioned something about 12 cores, i think thats a heck of a lot better than building firmware on your phone (and still be able to accept calls whilst doing it)
Mar 14 22:20:04 <aparcar[m]> hah I don't build on my phone but I'll think about it
Mar 14 22:20:23 <aparcar[m]> mangix: what's this libcurl error? I can't test your gettext stuff it it doesn't build at all
Mar 14 22:31:37 <grift> yes
Mar 15 00:09:40 <mangix> aparcar[m]: test json-glib instead. That libcurl error looks bogus.
Mar 15 01:18:03 <Grommish> hurricos: Well, rsalvaterra builds on like a P4 box :) WHich is why it takes him so long
Mar 15 01:23:44 <rsalvaterra> Grommish: Yep! Pentium D 950, until it breaks. :P
Mar 15 01:24:38 <Grommish> I took an old AMD SpamAssassin 1u box and turned it into a BitWarden server
Mar 15 01:24:54 <Grommish> I had to dig deep in the bone pile ot find DDR2 sticks
Mar 15 01:25:23 <rsalvaterra> mangix: Thanks, but Grommish is already running my test patch on his ER10X. :)
Mar 15 01:25:44 <Grommish> rsalvaterra: Its still going btw
Mar 15 01:25:58 <Grommish> So I don't think it'll dump at this point
Mar 15 01:27:33 <rsalvaterra> Grommish: Nice, thanks! Let it simmer for a couple of days, but I'm 95 % confident we're good.
Mar 15 01:28:09 <Grommish> 95% and you've not already send it in three days ago and wondered why it wasn't merged yet? :D ehhe
Mar 15 01:30:39 <rsalvaterra> Oh, I haven't sent it… I just pushed it to my remotes (on GitLab and GitHub).
Mar 15 01:31:55 <rsalvaterra> And I wanted to make sure it's tested on a HIGHMEM system, since blogic said that's where the random hangs were noticed.
Mar 15 01:32:14 <rsalvaterra> Pretty sure it was just a coincidence, but still…
Mar 15 01:53:35 <Grommish> hehe I was joking anyway about the rush of untested stuff being sent in.
Mar 15 02:40:10 <owrt-snap-builds> Build [#10](https://buildbot.openwrt.org/master/images/#builders/26/builds/10) of `apm821xx/sata` completed successfully.
**** ENDING LOGGING AT Mon Mar 15 03:00:14 2021