<tbilan>
Tusker: I did the --force. I think it worked
<Tusker>
tbilan: OK, that sounds good, is everything working ?
Tapper has quit [Ping timeout: 480 seconds]
Weasel___ has quit [Quit: Leaving]
Weasel___ has joined #openwrt-devel
<tbilan>
Tusker: stupid question... how can I tell if I'm running the sysupgrade.bin or initramfs.bin image?
<tbilan>
Tusker: I want to see if the --force took the upgrade or if it just reverted back to initramfs
<Tusker>
tbilan: if you power off and power on, if it boots into OpenWRT, then it's written it
<grid>
'dmesg|grep initram'
<grid>
if it saves settings between reboots, then it's probably not initramfs. even if you specified an overlay, the overlay settings wouldn't persist reboots
<grid>
fstab overlay settings rather
<tbilan>
dmesg | grep initram comes back empty
<tbilan>
it's saving settings
<tbilan>
so cool, it's running in normal mode then
<tbilan>
now I just need to figure out why only the bg radio is showing up in wireless. I checked the kernel wireless modules and bcm43 is enabled and both my bg and ac chips are bcm43 devices (bcm43520 and 43428)
<grid>
dmesg might give clues
<tbilan>
dug through that but only 1 802.11 line showed up. I'm looking to get pcitools on this
<Tusker>
your ac chip is supported by bcm43 ?
<tbilan>
Trying to find if the kernel module has support for 43428 but it's vague on make menuconfig
<tbilan>
getting the mr32 working is definitely an uphill battle, lol
<tbilan>
Table of Hardware: WLAN driver says it's "
<tbilan>
"partly supported'... whatever that means
<Tusker>
looking at https://git.openwrt.org/?p=openwrt/openwrt.git;a=commit;h=0e4092cbade0a601a2e924932c8b900aef5a11a4 - looks like only 1 card is supported
<Tusker>
it does look like the 43428 is the only one that comes up, I don't see the others
<tbilan>
I did some more digging and it looks like the 43520 is the ac chip
Rentong has quit [Ping timeout: 480 seconds]
Tusker has quit [Quit: Reconnecting]
Tusker has joined #openwrt-devel
<Tusker>
sorry, was testing my sons cat6 cabling skills and talked to him too long, and forgot to plug back myself in
<tbilan>
I have 2 boys. When my older kid was 12 we had a terminating RJ45 lesson. I walked up and cut his ethernet cord so he was properly motivated to learn how to put an end on correctly.
<tbilan>
I found that the 43520 chip is actually the ac chip
<Tusker>
i think it might be easier to grab some atheros boards and transplant than to get those brcm working
<Tusker>
need to log off, need to take my sons to the park before they go stir crazy
Tusker has quit [Quit: Time wasted on IRC: 2 hours 37 minutes 39 seconds]
<mangix>
`tbilan: b43 only supports g
<slh>
let's rephrase that as, it doesn't support HT/ VHT operations (son <=54 MBit/s), basic 5 GHz support exists (not that one would want to use it, due to the lack of performance and stability in combination with newer chipsets)
<slh>
s/son/so/
Tusker has joined #openwrt-devel
<tbilan>
bummer
<Tusker>
when I am trawling for APs on ebay... if it has Broadcom chips, I don't even bother trying
<mangix>
wonder when brcm will get DSA :B
<tbilan>
this mr32 has been collecting dust on a shelf for about 3 years so it was a good exercise to get it doing SOMETHING at least :)
<slh>
ipq40xx (as well as ipq806x and ipq807x) would be closer to my interest, at least those SOCs are actually useful (DSA for ipq806x already exists, it just needs to be merged)
<tbilan>
I got an MR18 working last week and it runs awesome on openwrt
<Tusker>
slh: yeah, I cracked open an old MR33 that was gathering dust, ipq40xx, but I need to lift the NAND and write a different uboot on to it before I can get openwrt going
<slh>
sadly these meraki devices aren't really available around here (well, they are, but not in larger quantities/ for reasonable prices)
<Tusker>
mr33 are $99 AUD unused on ebay here
<Tusker>
i paid AU$108 for the one I have, shipped it from the UK a year or so ago
minimal has quit []
<tbilan>
I have a Z1 and MX60 I'm going to try probably in a week or two
<slh>
for those prices I'd expect more cooperative hardware (yeah, I know, prices in .au are much worse in general)
<Tusker>
well, I paid AU$30 for the extreme network 3935i, which is much more beefy
<Tusker>
i suppose it's unusable hardware for most people, hence the price
<mangix>
b53 is a DSA driver...
<mangix>
oh I see. There are two drivers. The upstream one and a local swconfig one
<tbilan>
so who makes the best chipset?
<mangix>
tbilan: what does best mean?
<mangix>
what i'm getting at is, everything has tradeoffs
<mangix>
I just realized the issue with bcm43xx and DSA. No DTS support.
<rsalvaterra>
blocktrron_: Ok, doing a sanity build just to make sure…
<m>
I don’t know what to do right. There are two nearly identical devices. Do I need to combine them under one name manufacturer ODM / OEM? Beeline Smart Box TURBO+(128M RAM, ZigBee yes), SerComm S3 (256M RAM, ZigBee No).
fda has joined #openwrt-devel
rejoicetreat has quit [Ping timeout: 480 seconds]
fda- has joined #openwrt-devel
fda has quit [Ping timeout: 480 seconds]
<Slimey>
whats the difference between dts, dtb and dtsi files
<PaulFertser>
Slimey: dtb is device tree in binary form, without comments etc
fda- has quit [Remote host closed the connection]
<lemmi>
dts is what you write, dtsi is what you write and is usually included (as in C include) for things that multiple boards share, dtb is the compiled thing
<PaulFertser>
Slimey: dtsi are dt source include files. dts is just source
mangix has quit [Read error: Connection reset by peer]
mangix has joined #openwrt-devel
fda has joined #openwrt-devel
<PaulFertser>
Slimey: sources are preprocessed with regular "cpp".
<Slimey>
thanks i found the other day and couldnt find it again
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
<Slimey>
the manuf supplied them with the source do i need to edit or compare with whats in /sys/
fda- has joined #openwrt-devel
<PaulFertser>
Slimey: to do what?
<Slimey>
so the device tree should be the same?
<PaulFertser>
Slimey: please rephrase, I can't get the question.
<Slimey>
can i use the same dts to attempt openwrt build without editing it
fda has quit [Ping timeout: 480 seconds]
<PaulFertser>
Slimey: you can but usually DT needs to be kinda in sync with the kernel sources. As it's the kernel drivers that handle configuration data from the DT.
<fda->
i built the image with fedora, which has selinux enabled
Borromini has joined #openwrt-devel
<fda->
with freetz we had problems, eg systen-"sed" reported errors in fakeroot which does not support selinux
fda has quit [Ping timeout: 480 seconds]
<fda->
maybe some crosscompiler-uses-host-features?
<fda->
the non-ubi image built on same system worked well
<dangole>
fda-: so may related to jffs2 vs. ubifs overlay on top of an xattr-enabled squashfs... i'll try using exactly that config input and build the image on my host and flash it to my testing device
<dangole>
the config itself looks innocent, i don't think your choice of packages is the cause
<dangole>
meanwhile, if you like to try, it'd be cool if you can build an image using https://chef.libremesh.org/ having exactly your packages selected (you can't customize busybox there though, but that also should not be the cause)
<dangole>
fda-: will take me 20 minutes, buildhost is busy atm, but won't take too long (`nproc` == 16)
<fda->
some additional packages and kernel modules should not harm
<dangole>
fda-: there is a 'customize' field which allows you to add/remove packages from the image
<dangole>
fda-: it does NOT allow you to modify build-time configuration such as busybox options and stuff like that, basically it's just a web-frontend for the OpenWrt ImageBuilder which assembles images from the binary packages on downloads.openwrt.org.
<fda->
oh, i did not recognize "Customize +" is clickable
<dangole>
fda: yes, we got that disabled unless you are building OpenWrt with SELinux. thing is, that also the generated squashfs filesystem shouldn't be xattr-enabled unless you enable SELinux...
<dangole>
but somehow it got xattr in your case and hence the warning about it during boot (and most likely ubifs overlay freaking out because of that, because that also doesn't have xattr enabled by default but only if CONFIG_SELINUX=y
<fda>
so maybe the tool which formats the squashfs
Acinonyx has quit [Ping timeout: 480 seconds]
<dangole>
fda: i just looked into include/image.mk where we do that and realized we just assume that the host won't have security labels when calling mksquashfs...
<dangole>
fda: (look at Image/mkfs/squashfs-common, nothing tells squashfs to explicitely enable or disable xattr, security labels and all that. it's decided by either running it inside fakeroot (and hence with xattr and sec labels) or without, and when not using fakeroot we just assume there are no xattr and selinux labels
<dangole>
fda: we can't really disable XATTR support in squashfs-tools as we do need it when building SELinux-enabled images (which, btw, should work on your system :)
<fda>
selinux causes all the time problems...
<dangole>
fda: we'd have to tell mksquashfs in image.mk: setting `-no-xattrs` for regular builds and `-xattr` for SELinux-enabled builds should do the trick
<dangole>
fda: don't say that. SELinux is great, it just sometimes complicates things a bit...
<dangole>
fda: now flashing the image generated by your config script on my (non-SELinux) arch buildhost
<dangole>
fda: works like a charm. set the root passwor and populated /etc/dropbear/authrized_keys, rebooted, changed network config, rebooted again, all good. so that confirms our suspicion of something related to squashfs and xattr/selinux on your buildhost must have broken the image...
Rentong has joined #openwrt-devel
<fda>
dangole: thx, i will try different things, nur sure what exactly :) when openwrt needa a binary wiht and without selinux, should then be 2 build? depending on settings in .config?
<fda>
can an im age with selinux be build on a system without selinux?
<fda>
should i try with: SQUASHFSOPT += $(if $(CONFIG_SELINUX),-no-xattrs,-xattr) ?
<dangole>
fda: yes, that's how it is intended and we just didn't care :P
<fda>
complete build will take 25-30 minutes ...
<dangole>
fda: thank you for reporting and testing that!
<fda>
oh, if else has to be exchanged
<dangole>
fda: indeed...
Rentong has quit [Ping timeout: 480 seconds]
<fda>
addition question: when is mksquashfs executed? if i uncheck "keep settings" and then flash on luci?
<PaulFertser>
fda: mksquashfs is executed only during building of an image on host.
lucenera has quit [Read error: Connection reset by peer]
lucenera has joined #openwrt-devel
<fda>
thx
lucenera has quit []
danitool has quit [Read error: Connection reset by peer]
lucenera has joined #openwrt-devel
danitool has joined #openwrt-devel
Borromini has quit [Ping timeout: 480 seconds]
glbp has quit [Quit: leaving]
Borromini has joined #openwrt-devel
fda- has joined #openwrt-devel
fda has quit [Ping timeout: 480 seconds]
fda| has joined #openwrt-devel
fda| has quit []
DLange has quit [Quit: oh, look, upgrades...]
goliath has quit [Quit: SIGSEGV]
eck has quit [Quit: PIRCH98:WIN 95/98/WIN NT:1.0 (build 1.0.1.1190)]