Tapper has quit [Ping timeout: 480 seconds]
nlowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
schwicht_ has joined #openwrt-devel
schwicht has quit [Read error: Connection reset by peer]
goliath has quit [Quit: SIGSEGV]
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
<KGB-1> https://tests.reproducible-builds.org/openwrt/openwrt_lantiq.html has been updated. (98.2% images and 97.4% packages reproducible in our current test framework.)
Grommish_ is now known as Grommish
Grommish has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
Grommish has joined #openwrt-devel
philipp64|work has quit [Quit: philipp64|work]
philipp64|work has joined #openwrt-devel
JohnA has quit [Ping timeout: 480 seconds]
indy has quit [Read error: Connection reset by peer]
indy has joined #openwrt-devel
philipp64 has joined #openwrt-devel
philipp64|work has quit [Quit: philipp64|work]
<philipp64> Anyone know of a grep-like command for matching IPv4 subnets?
<Grommish> SInce the subnet will always start with 255.xxx.xxx.xxx, might work
<SwedeMike> philipp64: are you looking to find all addresses that matches a certain subnet?
<SwedeMike> Grommish: are you thinking of netmask?
<Grommish> Subnet is Netmask in expanded form.. and you can't subnet a class A at the first octect, can you?
<SwedeMike> philipp64: I am not aware of any such command line tool, you'd probably have to write a short program using one of the libraries such as https://docs.python.org/3/library/ipaddress.html
<Grommish> So, wouldn't it be always a 255. to match?
<SwedeMike> Grommish: no, it wouldn't always be 255.
<SwedeMike> 10.10.10.8/30 has no 255 in it
<Grommish> With 2 assignable addresses
<Grommish> its 255.255.255.252
<Grommish> for a /30
Tapper has joined #openwrt-devel
<Grommish> No?
<SwedeMike> Grommish: that's not a subnet, that's a netmask.
<SwedeMike> 10.10.10.8/30 is the subnet, 255.255.255.252 is the netmask
<Grommish> Yes, but a subnet just is shorthand AFAIK.. https://dnsmadeeasy.com/support/subnet/
<Grommish> Its a collision domain
<Grommish> set by the netmask?
<Grommish> Or am I just totally confused?
<SwedeMike> yes, but 255.255.255.252 isn't called "a subnet".
<SwedeMike> it's called netmask
<SwedeMike> it defines the size of the subnetwork, same as /30
<Grommish> Right.. Gotcha
nitroshift has joined #openwrt-devel
Tapper1 has joined #openwrt-devel
Tapper has quit [Read error: Connection reset by peer]
Tapper1 is now known as tapper
greearb has quit [Remote host closed the connection]
greearb has joined #openwrt-devel
Zaba2 has joined #openwrt-devel
<owrt-snap-builds> Build [#124](https://buildbot.openwrt.org/master/images/#builders/64/builds/124) of `realtek/generic` failed.
<blogic> svanheule: short question, i thought that ethtool -m worked once upon a time on SFP ports
<blogic> svanheule: trying it now it does not work anymore
<blogic> do you know if it did once work ?
<svanheule> blogic: I haven't tried using SFP, or ethtool -m, with OpenWrt
<blogic> svanheule: ok
<blogic> svanheule: will try an old image, thanks 1
<svanheule> blogic: recently asked bmork if he was having any issues with I2C via the rtl8231, but he said his module were polling just fine
<blogic> mine too
<blogic> the irq gave me issues
<blogic> so i turned of gpio-irq
<svanheule> ok
<blogic> will send you a detailed report next couple of days
<blogic> basically it tries to use gpio0
<svanheule> did you also make sure system led is disabled? that's on gpio0 on the realtek target
<blogic> it works
<blogic> but when I unplug the sfp it gets an irq and dthen oopses
Borromini has joined #openwrt-devel
<stintel> blogic: is the ethtool -m question related to realtek or in general ?
<blogic> stintel: yeah
<blogic> to the sfp ports
<stintel> ok I don't have any realtek with SFP, only octeon and there it didn't work
<blogic> ok, maybe i tested it on my pc witht he 10g sfps
<svanheule> blogic: I tried to enable a i2c-gpio bus (over the mdio-gpio bus to the RTL8231) to the SFPs on my Cisco SG220, but nothing shows up on 0x50/0x51 when I run i2cdetect
<svanheule> although I haven't been able to probe _any_ device via I2C on that machine, using the pinouts described in their source archive
<svanheule> also not the lm75a temperature sensors
danitool has joined #openwrt-devel
rejoicetreat has joined #openwrt-devel
goliath has joined #openwrt-devel
swegener has quit [Quit: leaving]
tapper has quit [Ping timeout: 480 seconds]
swegener has joined #openwrt-devel
swegener has quit []
swegener has joined #openwrt-devel
tapper has joined #openwrt-devel
Zaba2 has quit [Quit: Leaving.]
<rsalvaterra> I've been meaning to ask this for a long time, but I keep forgetting (because, well, it works)… Is specifying the IP address of an interface in CIDR notation (omitting option netmask) a supported feature, albeit undocumented, or does it just work by chance? I'm almost sure it's the former, but I'd like an authoritative answer… :)
rsalvate_ has joined #openwrt-devel
rsalvaterra has quit [Ping timeout: 480 seconds]
<jow> it is supported and documented
<jow> in fact ipaddr/netmask is legacy
<jow> when using cidr notation you can also specify multiple addresses or ranges using uci list notation
<jow> list ipaddr 192.168.1.1/24; list ipaddr 192.168.1.2; list ipaddr 10.25.0.1/24
<jow> all in one proto static interface
<jow> same for ip6addr
goliath has quit [Ping timeout: 480 seconds]
<Borromini> isn't CIDR kind of enforced by LuCI now?
goliath has joined #openwrt-devel
<rsalvate_> jow: Good to know, but there's no info about it on /etc/config/network in the wiki I could find.
rsalvate_ is now known as rsalvaterra
<rsalvaterra> jow: I knew about the list of multiple IP addresses, since I had to configure a connection with IP aliasing years ago: https://paste.debian.net/1199999/ :)
barhom has joined #openwrt-devel
mustafacan has quit [Remote host closed the connection]
<barhom> Trying to start using the physical buttons. Reading the docs it says there are two ways hotplug.d and procd. And it mentions that hotplug is deprecated? So I should not be using /etc/hotplug.d/buttons ? Docs are a bit confusing.
<barhom> The deprecated notice was several years ago
feriman has joined #openwrt-devel
gch981213 has quit [Quit: leaving]
<wb9688> Hotplug2 is deprecated in favor of procd, which reads the /etc/hotplug.d directory
lucenera has quit [Quit: The Lounge - https://thelounge.chat]
lucenera has joined #openwrt-devel
<Borromini> there is some backward compatibility :)
tapper has quit [Remote host closed the connection]
tapper has joined #openwrt-devel
<barhom> ok perfect, so /etc/hotplug.d is the new format using procd ;)
<KGB-2> https://tests.reproducible-builds.org/openwrt/openwrt_bcm47xx.html has been updated. (100.0% images and 95.1% packages reproducible in our current test framework.)
rejoicetreat has quit [Ping timeout: 480 seconds]
tapper has quit [Read error: Connection reset by peer]
tapper has joined #openwrt-devel
nitroshift has quit [Quit: Gone that way --->]
* russell-- has an ancient soekris net4826 he's been nursing along, which is hanging on a warm boot: http://paste.debian.net/1200002/ ... if i power cycle it, it boots successfully.
pmelange has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
<jow> russell--: gives me a deja-vu feeling
<jow> afair that issue existed in the past as well
<russell--> yeah, i haven't been paying close attention ... it's probably been there for a while
<russell--> maybe tomorrow i'll try to bracket it with some old builds
<tmn505> this might be CONFIG_X86_REBOOTFIXUPS not enabled, but seing as it's hanging around clocksource maybe You'll need clocksource=scx200_hrt in kernel command line.
<nick[m]4> how can I add "libm" as dependency to a package? gpsd fails under glibc
<russell--> tmn505: clocksource=scx200_hrt didn't help
<tmn505> thanks, I'll test k5.10 later today on my WRAP boards, this might be related to changes in config.
<russell--> i'm building with CONFIG_X86_REBOOTFIXUPS right now
<tmn505> btw, which subtarget are You using, geode or legacy?
<jow> nick[m]4: do you get an error "Package xxxx is missing dependencies for the following libraries: libm.so" or something else?
<jow> nick[m]4: because libm.so is part of "libc" with glibc and every non-kmod package in openwrt buildroot implicitly depends on +libc
<jow> nick[m]4: ah okay, so its not a packaging issue
<nick[m]4> jow: hmmm... maybe it is sysroot="$(STAGING_DIR)" ?
<jow> scons... shudder
<nick[m]4> jow: yeah excatly. if I manually patch the scons file to link "-lm". it says it can not find it
<nick[m]4> if I change the sysroot, it can compile because it uses my host lib
<nick[m]4> as I believe
<jow> my first fix attempt would be adding this: https://pastebin.com/XGs5Lkxt
<jow> somewhere before line 85
<jow> or laternatively change SCONS_VARS += LINKFLAGS="$(TARGET_LDFLAGS)"
<jow> to SCONS_VARS += LINKFLAGS="$(TARGET_LDFLAGS) $(if $(CONFIG_USE_GLIBC),-lm)"
<jow> if that does not work, you need to dig deeper and figure out how one can pass additional link libraries to scons
<nick[m]4> thanks. I alrady tried similar but I will try again. it is that scons is searching if libm is existing. otherwise it will just drop libm
<jow> yeah but its yet another stupid nih build system that utterly fails at cross compilation
<jow> it probably probes your host libm.so, failes to link it in its test case and concludes that there is none
<nick[m]4> when I manually patch the scons file to use libm
<nick[m]4> it says it can not find it
<nick[m]4> "cannot find /lib/libm.so.6"
<jow> *facepalm*
<nick[m]4> jow: me, or the program? xD
<jow> the program
<nick[m]4> okay ^^
<jow> it could be that --sysroot interacts somehow with the patched specs in OpenWrt's cross-gcc
<nick[m]4> hmmmm
<nick[m]4> so I have to look into the scons again?
<jow> I believe openwrt does not build its gcc with sysroot support
<nick[m]4> I could also place the lib in scons as libpath
<nick[m]4> but neheb included the sysroot stuff?
<jow> well, maybe simply revert that commit
<jow> it claimed to fix some setups but clearly broke others
<jow> so a revert seems to be the proper course of action, until a better fix can be found
<nick[m]4> but then I break other targets again?
<nick[m]4> I believe then I remove the sysroot stuff, it will just use my host libs?
<jow> not sure tbh, I'd expect it to not be able to link if it would really use your host libs
<nick[m]4> but I build x86_64 on a x86 system
<jow> hm, true
<nick[m]4> I just wanted to fix olsrd with glibc... now I stucked fixing gpsd... why did I take over the maintainership :'(
<jow> I don't understand where it is taking /usr/lib/libm.so.6 from
<jow> it must somehow come from gcc itself
<nick[m]4> "/lib/libm.so.6" would be correct, if it looks at the staging tree
<nick[m]4> staging_dir
<nick[m]4> maybe the paths need to be extended?
<nick[m]4> "$(STAGING_DIR)/lib/libm.so.6"
<jow> can you paste the complete output of "make package/gpsd/{clean,compile} V=sc" ?
<jow> *pastebin
<nick[m]4> yeah sure
<nick[m]4> I will clean all my changes first
<nick[m]4> or do you want with patched lm?
<jow> with the patche
<jow> -e
<nick[m]4> okay
rejoicetreat has joined #openwrt-devel
<nick[m]4> maybe prefix is wrong?
<nick[m]4> and should also be STAGING_DIR?
<jow> I don't think so
T-Bone has joined #openwrt-devel
T-Bone is now known as f00b4r0
<jow> what is reported by find /tmp/openwrt/staging_dir/toolchain-x86_64_gcc-8.4.0_glibc/ -name "libm.so*" ?
<nick[m]4> /tmp/openwrt/staging_dir/toolchain-x86_64_gcc-8.4.0_glibc/lib/libm.so.6
<nick[m]4> /tmp/openwrt/staging_dir/toolchain-x86_64_gcc-8.4.0_glibc/lib/libm.so
<nick[m]4> ahhh maybe now I have to add lm again to linkingflags?
<jow> well I think part of the problem is that the libm is not part of the sysroot
<jow> --sysroot points to staging_dir while libm etc. are in toolchain_dir
<jow> maybe it is already enough if you switch the sysroot from $(STAGING_DIR) to $(TOOLCHAIN_DIR)
<jow> staging dir search paths are added through -I and -L, so that shouldn't be an issue
<plntyk2> nick[m]4, dunno but did you check if buildroot scons patch for gpsd works ? https://git.buildroot.net/buildroot/log/package/gpsd/0001-Fix-errors-during-gpsd-3.20-cross-compilation.patch
schwicht_ has quit [Read error: Connection reset by peer]
schwicht has joined #openwrt-devel
<nick[m]4> jow: that works! :D shit I though staging_dir is the toolchain dir xD
<jow> no it isn't, a common misconception
<jow> since on openwrt one toolchain may be shared by several targets, they're separate
JohnA has joined #openwrt-devel
JohnA has left #openwrt-devel [Leaving]
JohnA has joined #openwrt-devel
<nick[m]4> jow: thanks! made a PR
<f00b4r0> jow: fiber coming to slashdirt buildbot in the coming weeks. Expect them back online this summer hopefully :)
<nick[m]4> xD
<nick[m]4> shit
<jow> nick[m]4: yeah, that's due to -L/home/build/openwrt/staging_dir/hostpkg/lib -L/home/build/openwrt/staging_dir/host/lib
<jow> no idea what adds this and why
<jow> passing host lib dirs to across build is clearly no good idea
<nick[m]4> yep
<nick[m]4> maybe the "env.Prepend(LIBPATH=[os.path.realpath(os.curdir)])" ?
<nick[m]4> I will try deleting that
<jow> sorry, phone call
_whitelogger has joined #openwrt-devel
<nick[m]4> added that now as default
<nick[m]4> let see if it will help
<JohnA> wrt3200acm + pppoe, is this a bug? is there a work aound?
<nick[m]4> thanks for your support
Borromini has quit [Ping timeout: 480 seconds]
<jow> nick[m]4: I think the wrong .../host/lib and .../hostpkg/lib paths come from scons itself
<jow> these paths likely got builtin as default paths when scons/host got build
<jow> and now it appends these as -L paths to each target build
_lore_- has joined #openwrt-devel
_lore_ has quit [Ping timeout: 480 seconds]
_lore_- is now known as _lore_
pmelange1 has joined #openwrt-devel
pmelange has joined #openwrt-devel
pmelange1 has quit [Ping timeout: 480 seconds]
<nick[m]4> jow: hmmmm and how can we fix it? ^^
nlowe has joined #openwrt-devel
<nick[m]4> in the scons.mk is a variable called "EXTRA_LDFFLAGS"
<nlowe> https://github.com/openwrt/openwrt/pull/4228 has been opened by huaracheguarache for fragattacks fixes for ath10k-ct
<nick[m]4> I will try overriding LDFLAGS by scons
<nlowe> Suggest this goes both in to mainline and ultimately 21.02.0
greearb_ has joined #openwrt-devel
pavlix has joined #openwrt-devel
<pavlix> Hello. What exactly is procd_add_validation for in service_triggers? It doesn't seem to affect reload_config.
<pavlix> Also `ubus call service validate` only shows data from some of the services.
pmelange has left #openwrt-devel [#openwrt-devel]
pmelange has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
pmelange has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
_lore_ has quit [Ping timeout: 480 seconds]
_lore_ has joined #openwrt-devel
<nick[m]4> jow: I finallky fixed it! :D
f5 has quit [Remote host closed the connection]
f5 has joined #openwrt-devel
f5 has quit [Remote host closed the connection]
f5 has joined #openwrt-devel
Luke-Jr has quit [Ping timeout: 480 seconds]
<philipp64> SwedeMike: yes, only addresses matching CIDR blocks or subnets.
rsalvate_ has joined #openwrt-devel
rsalvate_ is now known as rsalvaterra_
philipp64 is now known as philipp64|laptop
philipp64|laptop is now known as group
group is now known as philipp64|laptop
philipp64|laptop is now known as philipp64
philipp64 is now known as philipp64|laptop
rsalvaterra has quit [Ping timeout: 480 seconds]
<tmn505> russell--: mine WRAP bootloops with k5.10
Luke-Jr has joined #openwrt-devel
Tapper has quit [Remote host closed the connection]
Tapper has joined #openwrt-devel
<jow> nick[m]4: how?
<nick[m]4> setting cflags
<nick[m]4> however, the other scon stuff is probably still wrong
<jow> nick[m]4: ah yes. Still I think the scons build itself should be fixed to not use the wrong default flags. But this is obviously far out of scope of waht you actually wanted to do, polish olsrd :)
<nick[m]4> xD
<jow> but as long as providing CFLAGS completely overrides the bogus builtin default, the solution is adequate for gpsd
<johnf> I'm working on support for the gl-xe300
<johnf> my image files work through the gl-inet uboot http interface
<johnf> but not using sysupgrade in OS
<johnf> I think the error that matters is tar: invalid tar magic
<johnf> which suggests that I'm using the wrong compression maybe, but I don't understand how that could be
<johnf> I'd really appreciate any help anyone can offer, I want to submit a PR but no sysupgrade support isn't really where I want it to be when I do :(
Zaba2 has joined #openwrt-devel
Borromini has joined #openwrt-devel
<tmn505> russell--: it boots fine the legacy subtarget image with clocksource=scx200_hrt applied to kernel. So the issue is, when kernel configs were prepared for 5.10, guidosarducci used generic subtarget to refresh, instead of lowest denominator which is legacy subtarget.
<tmn505> That's why geode kernel inherits CONFIG_M686 from target config instead of CONFIG_M586MMX. So now geode is i686.
philipp64|work has joined #openwrt-devel
<jow> tmn505: do you have a list of kconfig symbols we should enable/change/disable for legacy?
pekster has quit [Quit: musical hosts]
<jow> not sure if I still own my wrap board, might that I've thrown it into the trash before the move to the new appartment
pekster has joined #openwrt-devel
<tmn505> jow: legacy is fine, it was fixed later
<tmn505> the issue is with geode target
<tmn505> the symbols which should be changed are CONFIG_X86_MINIMUM_CPU_FAMILY=6 to 5 and CONFIG_M686 to CONFIG_M586MMX.
<tmn505> I'll dig what else changed and prepare patches for that issue.
netprince has quit [Remote host closed the connection]
<johnf> nand_upgrade_tar /tmp/gl-xe300.bin
<johnf> ok, so it's deciding to do a tar upgrade
<johnf> but the image file is actually not a tar file, but a uboot image
<johnf> openwrt-ath79-nand-glinet_gl-x300b-nor-squashfs-sysupgrade.bin: u-boot legacy uImage, MIPS OpenWrt Linux-5.4.117, Linux/MIPS, OS Kernel Image (lzma), 2252807 bytes, Fri May 21 09:31:46 2021, Load Address: 0x80060000, Entry Point: 0x80060000, Header CRC: 0x45AD609B, Data CRC: 0x7F502497
Luke-Jr has quit [Ping timeout: 480 seconds]
<tmn505> jow: russell--: I migth have pointed fingers to soon, it seems that geode cpu family is set properly in config. Will investigate what else chaged.
Luke-Jr has joined #openwrt-devel
dedeckeh has joined #openwrt-devel
schwicht has quit [Read error: Connection reset by peer]
philipp64|laptop has quit [Quit: philipp64|laptop]
schwicht has joined #openwrt-devel
rmilecki has joined #openwrt-devel
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
<xdarklight> hauke: I have not added you to https://patchwork.kernel.org/project/netdevbpf/patch/20210604161250.49749-1-lxu@maxlinear.com/#24225991 as your work email is already Cc'ed. just as an FYI if the original mail didn't make it to your inbox
feriman has quit [Ping timeout: 480 seconds]
philipp64 has joined #openwrt-devel
philipp64 has quit []
<nlowe> Geode should be i585?
<nlowe> Geode should be i686 though?
<nlowe> (Ignore typo in first)
<nlowe> All Geodes since and including National Semi's GX2 support CMOV but not NOPL. (Introduced in 2002. These would not break and are i686.)
<nlowe> The AMD branded Geodes (GX and LX) support CMOV but not NOPL. (These would not break.)
<nlowe> Why would you compile for i586?
<aparcar[m]> after a sysupgrade, is the new image writeable before a reboot?
<aparcar[m]> aka is the overlay fs already in place?
rmilecki has quit [Ping timeout: 480 seconds]
nlowe has quit [Quit: Textual IRC Client: www.textualapp.com]
<owrt-snap-builds> Build [#127](https://buildbot.openwrt.org/master/images/#builders/24/builds/127) of `ramips/rt288x` completed successfully.
Borromini has quit [Quit: leaving]
philipp64 has joined #openwrt-devel
philipp64 has quit []
philipp64 has joined #openwrt-devel
dedeckeh has quit [Remote host closed the connection]
rejoicetreat has quit [Read error: Connection reset by peer]
rejoicetreat has joined #openwrt-devel
nlowe has joined #openwrt-devel
<mangix> nick[m]4: I am curious why you're building with glibc
huaracheguarache has joined #openwrt-devel
huaracheguarache has quit []
<nick[m]4> mangix: because some targets use glibc per default
<nick[m]4> ?
<mangix> nick[m]4: that's only ARC
<nick[m]4> mangix: yep. and that the target that is in the github workflow ^^
<mangix> I ask since I never saw gpsd fail on that architecture
<mangix> i don't doubt the PR is good, but typically with failures like these, people use glibc on other architectures where the default is musl.
<nick[m]4> yes. I used it first on x86 with glibc. and then I realized that something is wrong. I'm unsure if arc was wrong before.
Tapper has quit [Remote host closed the connection]
Tapper has joined #openwrt-devel
<mangix> ARC is an architecture on life support
<mangix> in terms of OpenWrt
<nick[m]4> but is it not good, to also support glibc?
<mangix> sure, if someone cares
<mangix> Most of the reason glibc is in the tree AFAIK is because some of the core people want to run fancy stuff like ASAN and UBSAN
<nick[m]4> okay, I did not know that
<mangix> while on the topic of glibc, several hacks in the packages feed probably need to be looked at.
<nick[m]4> or we drop glibc compltetly ...
<mangix> funny you say that. some core developers would be in favor of that. That was a big reason why I got rid of uClibc-ng. The issue then becomes ARC, which has no musl support.
Tapper has quit [Ping timeout: 480 seconds]
<nick[m]4> ... ;) I finally found a glibc workaround for olsrd. however, for that I need to check for glibc in a makefle.
<nick[m]4> why did u get rid of uclibc?
<mangix> nick[m]4: because it was marked as BROKEN and only ARC was using it. ARC has glibc support, which was present in the tree. I didn't introduce glibc.
<mangix> nobody was maintaining it either.
<nick[m]4> mangix: just reading about it. sounds good.
<nick[m]4> that we got rid of it
<nick[m]4> what is actually this arc target? why do we support it? are there some routers using the platform?
<mangix> none
<mangix> it's some russian embedded platform.
<mangix> nvm says American.
floof58 has quit []
floof58 has joined #openwrt-devel
<hauke> xdarklight: thanks for the link, I didn't check my work email in the evening. With the 2.5G support there are many new featuers, I think it is easier to just start with a new driver
rejoicetreat has quit [Ping timeout: 480 seconds]
nlowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]