**** BEGIN LOGGING AT Sun Sep 27 03:02:45 2020
Sep 27 04:40:03 <lmore377_> im porting to a device and it has funky led controls and i dont know how to name the gpio pins for it. basically theres a tiny microcontroller that controls a red/blue led based on how 3 different gpio pins are switched (here's an example: https://pastebin.com/raw/iPZhuXp0). Right now I have them named rt4230w:ledctrl:1,rt4230w:ledctrl:2, and rt4230w:ledctrl:3 in the dts but I know that this doesn't follow the normal naming
Sep 27 04:40:04 <lmore377_> scheme. what exactly should I name them?
Sep 27 04:41:31 <Grommish> lmore377_: If you can, just hang out here and someone will answer.. adrianschmutzler should be able to help, but he isn't here right now (keep an eye out for him)
Sep 27 04:42:04 <Grommish> lmore377_: karlp would probably have suggestions, too
Sep 27 04:43:06 <lmore377_> adrianschmutzler is actually the one that pointed that out in my pull request https://github.com/openwrt/openwrt/pull/3115
Sep 27 04:46:00 <Grommish> Right.. he's pendantic about naming conventions, but it helps with the long-term organization
Sep 27 04:51:57 <Grommish> I'd offer suggestions, but my device doesn't use dtb/dts, so I don't even know the format.. i never had to learn
Sep 27 04:52:51 <Grommish> But, outside the dts/dtb, you could call the GPIOs directly, if you know what the gpio pin numbers are
Sep 27 04:54:21 <Grommish> lmore377_:  https://pastebin.com/06f2VHEP is how i do it locally
Sep 27 04:55:26 <Grommish> but I'm sure there  is a nifty shortcut easy way since the leds are mapped in the device tree
Sep 27 05:03:12 <lmore377_> i hope so. this is what the led section of the dts loos like https://pastebin.com/LbngbxhR
Sep 27 06:51:36 <grift> aparcar[m]: i think so yes
Sep 27 06:52:27 <grift> aparcar[m]: would have a look into packaging a bit
Sep 27 06:59:26 <mangix> aparcar[m]: huh?
Sep 27 07:15:51 <Grommish> mangix: did your popcorn event actually pull in a random outside party?  Impressive...
Sep 27 07:16:05 <Grommish> mangix: you do not fail to entertain ;p
Sep 27 07:29:18 <grift> aparcar[m]: i suppose i would have to atleast depend on a secilc host package since that is one of the few selinux user space packages not available yet
Sep 27 07:47:28 <aparcar[m]> Grommish: what is a random outside party?
Sep 27 07:48:03 <Grommish> aparcar[m]: https://github.com/openwrt/packages/issues/12812#issuecomment-699476038
Sep 27 07:48:11 <aparcar[m]> grift: packaging is a pure joyride, please try to package both and let me know if you need help
Sep 27 07:49:03 <aparcar[m]> Grommish: thanks. Maybe he can also help with the redesign of openwrt.org
Sep 27 07:49:54 <Grommish> haha I mean.. As far as I can tell, that guy just stumbled by and was like.. To the rescue? Dunno
Sep 27 07:50:00 <aparcar[m]> mangix: so even with NLS vim builds fine, even though you stated it should break. Is vim fixed or is the CI test not tough enough?
Sep 27 07:55:09 <aparcar[m]> grift: Grommish do you use docker?
Sep 27 07:55:31 <aparcar[m]> Grommish: your msgfmt hint was pretty close to the solution it works now
Sep 27 07:55:41 <Grommish> No.. I could, i just never had a need.. when I did vx, i used esxi
Sep 27 07:56:04 <Grommish> Want me to load it?
Sep 27 07:57:33 <aparcar[m]> no need
Sep 27 07:57:58 <aparcar[m]> there are soonish (2h) some SELinux imagebuilder
Sep 27 07:58:34 <Grommish> ok :)  If you need to be build out again, lemme know
Sep 27 07:58:37 <Grommish> I certainly don't mind
Sep 27 07:58:55 <aparcar[m]> Yea I got something for you in a few minutes
Sep 27 07:59:08 <aparcar[m]> actually I want this computer you're using, 8 min for a full builds sound good
Sep 27 07:59:16 <aparcar[m]> please send me the specs via PM
Sep 27 07:59:25 <Grommish> non-toolchain recompile builds
Sep 27 07:59:33 <Grommish> toolchain takes longer
Sep 27 07:59:45 <aparcar[m]> oh uh I break the toolchain where I can
Sep 27 07:59:58 <Grommish> Yeah, but it isn't hours
Sep 27 08:00:04 <Grommish> I think its like 45 mins or so
Sep 27 08:00:08 <Grommish> I'll time it next time
Sep 27 08:05:58 <aparcar[m]> mhh
Sep 27 08:06:44 <mangix> aparcar[m]: beats me. i doubt it though
Sep 27 08:07:16 <aparcar[m]> what does beats me mean in this context?
Sep 27 08:07:29 <mangix> i don't know
Sep 27 08:07:56 <mangix> yeah...looking at the history, it should still be failing
Sep 27 08:08:11 <aparcar[m]> snap, so how do I have to run it?
Sep 27 08:08:51 <mangix> it's a compile time failure
Sep 27 08:09:01 <mangix> related to the host tools actually
Sep 27 08:09:26 <aparcar[m]> mangix: uhm well I got the SDK with NLS enabled. Isn't that enough?
Sep 27 08:09:29 <mangix> wait a minute....I may or may not have fixed the offending host uytility
Sep 27 08:09:46 <mangix> I'd have to pull up my VM and see if it still fails.
Sep 27 08:11:14 <aparcar[m]> please dp
Sep 27 08:11:17 <aparcar[m]> do
Sep 27 08:14:57 <mangix> alright, building
Sep 27 08:15:16 <mangix> Unfortunately, I have a slow CPU
Sep 27 08:17:01 <aparcar[m]> mangix: sudo upgrade engine
Sep 27 08:17:42 <aparcar[m]> grift: I'm lost, what exactly is security_context_t? Looking at the selinux source it's just a string. So is it the labels separated via :?
Sep 27 08:19:44 <mangix> aparcar[m]: note that I'm building all packages and see which fails first
Sep 27 08:21:11 <aparcar[m]> you should get some sponsoring going and get some fancy hexa core engine
Sep 27 08:21:28 <grift> aparcar[m]: not sure what you mean but a security context to the kernel is just a string
Sep 27 08:21:59 <aparcar[m]> oh then this is so much easier than I though
Sep 27 08:22:02 <aparcar[m]> t
Sep 27 08:22:04 <aparcar[m]> thanks
Sep 27 08:30:29 <mangix> aparcar[m]: where does this nls thing come from?
Sep 27 08:31:01 <aparcar[m]> form my CI
Sep 27 08:31:15 <mangix> no i mean where does the sdk come from?
Sep 27 08:31:27 <aparcar[m]> freshly build
Sep 27 08:31:30 <aparcar[m]> in my CI
Sep 27 08:31:42 <aparcar[m]> you want a link?
Sep 27 08:32:23 <mangix> nah that's fine
Sep 27 08:32:37 <mangix> i was wondering if it comes from downloads.openwrt.org
Sep 27 08:33:00 <aparcar[m]> no upstream doesn't offer that currently afaik
Sep 27 08:34:18 <mangix> nothing failing locally yet...
Sep 27 08:34:20 <mangix> weird
Sep 27 08:34:36 <aparcar[m]> ha!
Sep 27 08:34:50 <aparcar[m]> please find a package that fails to test the CI afterwards
Sep 27 08:37:21 <mangix> ooo cppcon has new videos
Sep 27 08:38:39 <aparcar[m]> grift: could you please give me an example context for e.g. odhcpd?
Sep 27 08:42:05 <grift> u:r:odhcpd.subj
Sep 27 08:44:26 <mangix> aparcar[m]: vim was the first to fail
Sep 27 08:45:10 <aparcar[m]> can you try the sdk and let me know what could be the difference?
Sep 27 08:45:18 <aparcar[m]> mangix: do you need to add any extra option?
Sep 27 08:45:23 <aparcar[m]> grift: sweet thanks
Sep 27 08:45:48 * aparcar[m] sent a long message:  < https://matrix.org/_matrix/media/r0/download/matrix.org/DDddChlomhYyIqwhUJaqGUiE/message.txt >
Sep 27 08:45:52 <aparcar[m]> not sure what is going on here...
Sep 27 08:47:03 <mangix> msgfmt: Cannot convert from "ISO-8859-1" to "UTF-8". msgfmt relies on iconv(). This version was built without iconv().
Sep 27 08:47:19 <mangix> gettext-full problem
Sep 27 08:48:24 <Grommish>  Does the package have $(ICONV_DEPENDS) in DEPENDS?
Sep 27 08:48:29 <Grommish> Just out of curiousity
Sep 27 08:49:03 <aparcar[m]> mangix: so that's missing in the sdk?
Sep 27 08:53:49 <mangix> Guess so
Sep 27 08:53:53 <mangix> hence it succeeds
Sep 27 08:54:11 <mangix> Grommish: not the issue. The host binary is bad
Sep 27 08:54:17 <Grommish> Ah
Sep 27 08:54:37 <Grommish> Wait, how does that work though
Sep 27 08:54:56 <Grommish> I built out aparcar[m] selinux-nls branch without issues earlier hehe
Sep 27 08:55:04 <aparcar[m]> mangix: grift can you help me with <artificial>:(.text+0x4424): undefined reference to `setexeccon'
Sep 27 08:55:16 <mangix> I'm talking about vim, not selinux
Sep 27 08:55:18 <aparcar[m]> I'm including selinux.h but it doesn't want to link things
Sep 27 08:55:24 <Grommish> what's wrong with vim?
Sep 27 08:55:27 <aparcar[m]> mangix: but you're a smart C person
Sep 27 08:55:43 <aparcar[m]> It's pretty much the third time I touch a C program
Sep 27 08:55:53 <mangix> I don't really code C.
Sep 27 08:56:07 <aparcar[m]> but you fix all kinds of crazy C issues
Sep 27 08:56:42 <Grommish> I need to find someone who is amazing with gdb *sigh* I don't know if I can dig deep enough into asm to identify the issues i'm having
Sep 27 08:57:06 <mangix> well, C++ and C go hand in hard
Sep 27 08:57:08 <mangix> *hand'
Sep 27 08:57:40 <mangix> Grommish: was it the float thing?
Sep 27 08:57:44 <Grommish> yes
Sep 27 08:57:47 <Grommish> Appears to be a musl bug
Sep 27 08:57:53 <mangix> do tell
Sep 27 08:58:08 <Grommish> FILE f* is passing a negative float of like 3.xxxxe-43 hehe
Sep 27 08:58:22 <Grommish> https://forum.suricata.io/t/suricata6-0-0-beta1-on-openwrt-illegal-instruction-error/572
Sep 27 08:58:41 <Grommish> but, the target is using -msoft-float
Sep 27 08:58:59 <Grommish> so it shouldn't be passing sdc1 at all from what I understand
Sep 27 08:59:10 <Grommish> yet, there it is.. and no one can figure out where it's coming from
Sep 27 08:59:31 <mangix> weird
Sep 27 08:59:42 <mangix> have you tried a glibc build?
Sep 27 08:59:53 <Grommish> It failed to build out at all
Sep 27 09:00:10 <Grommish> I'm using 5.4 with 10.2 and 2.35 binutils
Sep 27 09:00:41 <Grommish> and I simply don't know enough to find info that might be helpful.. PaulFertser was trying hehe
Sep 27 09:01:14 <mangix> which version of musl are you using?
Sep 27 09:01:22 <Grommish> I thought about moving to musl1.2.x, but it's whatever is defaulted
Sep 27 09:01:36 <Grommish> I never hard the 1.2.0 issues everyone else did..no wifi
Sep 27 09:01:39 <Grommish> err had
Sep 27 09:01:49 <mangix> Hauke fixed that
Sep 27 09:01:59 <mangix> I currently run it on my tp-link router
Sep 27 09:02:29 <Grommish>  PKG_VERSION:=1.1.24
Sep 27 09:02:51 <mangix> Grommish: https://git.musl-libc.org/cgit/musl/commit/?id=bff78954995b115e469aadb7636357798978fffd
Sep 27 09:02:55 <mangix> :)
Sep 27 09:03:10 <mangix> aparcar[m]: try testing with gphoto2
Sep 27 09:03:15 <mangix> it's also failing here
Sep 27 09:03:29 <grift> aparcar[m]: i am not sure  about that could it be that youre including procd or busybox with and without selinux support?
Sep 27 09:03:33 <Grommish> mangix: holy hell.. Much appreciated
Sep 27 09:03:50 <grift> setexeccon should be a libselinux function
Sep 27 09:04:14 <mangix> Grommish: for future reference, I went to musl's git and searched for float in the log msg
Sep 27 09:04:19 <mangix> that seemed relevant
Sep 27 09:04:31 <mangix> that commit is part of 1.2.0
Sep 27 09:04:32 <aparcar[m]> yea I include selinux/selinux.h
Sep 27 09:04:46 <Grommish> Appreciate the info..  I am still learning where everything simply is :)
Sep 27 09:04:46 <aparcar[m]> http://sprunge.us/dQgOfw
Sep 27 09:07:13 <mangix> Grommish: if 1.2.1 fixes the crash, I'd mention it in the musl update PR so that the OpenWrt caball merges it
Sep 27 09:07:47 <Grommish> I'm actually going to backport it to 1.1.24 for testing
Sep 27 09:08:10 <grift> aparcar[m]: i am not big on c maybe this helps: https://github.com/SELinuxProject/selinux/blob/master/libselinux/src/setexecfilecon.c
Sep 27 09:09:09 <aparcar[m]> including something called _internal.h sounds wrong
Sep 27 09:09:22 <aparcar[m]> the function is found in selinux.h, i just don't understand what the error really means
Sep 27 09:10:22 <grift> i guess maybe its not meant to be used externally, but i am out of my field now
Sep 27 09:10:30 <aparcar[m]> ok
Sep 27 09:10:41 <grift> theres an example in that c file for rpm. where it uses setexecfilecon()
Sep 27 09:10:52 <grift> maybe that is public
Sep 27 09:12:09 <Grommish> Ok.. I don't know much in the way of coding, but, adding the header file isn't enough
Sep 27 09:12:10 <grift> but dont hard reference types as is done there because thats bad
Sep 27 09:12:29 <Grommish> the header defines it, but doesn't instantiate it or something
Sep 27 09:12:38 <Grommish> that has to be done in the .c before you use it
Sep 27 09:14:05 <Grommish> really, only other helpful thing is that <> are kernel includes, "" are local-directory based
Sep 27 09:16:02 <aparcar[m]> https://git.openwrt.org/?p=project/procd.git;a=blob;f=initd/init.c;h=2eb6ead2c0f466a7d91eb813f7de0cbdaedba42d;hb=HEAD
Sep 27 09:16:11 <aparcar[m]> it's already part of procd
Sep 27 09:16:19 <aparcar[m]> and at this point it works
Sep 27 09:16:22 <aparcar[m]> but not for instance
Sep 27 09:18:20 <aparcar[m]> jow: ideas?
Sep 27 09:18:51 <grift> aside: if you want to get ideas these are for example some of the functions systemd uses https://github.com/systemd/systemd/blob/e813a74ae8fa0d824f1a86b789f9fb585cb2d6ed/src/basic/selinux-util.c
Sep 27 09:23:12 <aparcar[m]> I gotta sleep
Sep 27 09:23:18 <aparcar[m]> big waves tomorrow morning
Sep 27 09:23:33 <aparcar[m]> I'll hope dangole_ can give some insights later
Sep 27 09:23:34 <aparcar[m]> byee
Sep 27 09:23:47 <aparcar[m]> mangix: of and if you have an idea on how to break vim via the NLS sdk please let me know
Sep 27 09:24:29 <mangix> none
Sep 27 09:24:38 <mangix> i don't care for the package honestly
Sep 27 09:24:45 <mangix> try gphoto2
Sep 27 09:24:48 <mangix> it also fails
Sep 27 09:49:53 <russell--> i'm seeing a bunch of dlsym messages related to fakeroot during a ramips build, what's up with that?
Sep 27 09:50:49 <russell--> e.g.: logs/package/system/fstools/compile.txt:dlsym(acl_set_file): /home/openwrt/src/mth/staging_dir/host/lib/libfakeroot.so: undefined symbol: acl_set_file
**** BEGIN LOGGING AT Wed Sep 30 01:30:48 2020
Sep 30 01:45:06 <mangix> dunno
Sep 30 01:45:16 <mangix> maybe jow has an answer
**** ENDING LOGGING AT Wed Sep 30 02:59:57 2020