<mangix>
Wow. TPM chips are expensive now. Wonder if I should sell mine.
rotanid has quit [Ping timeout: 480 seconds]
mkresin has quit [Ping timeout: 480 seconds]
danitool has joined #openwrt-devel
<Tapper>
Just out of intrest if some one was to donate 1000000 $ to OpenWrt what could it achieve? Don't get your hopes up I am skint lol I just wanted to know how far 1000000$ would go.
<Tapper>
and how it would change the face of OpenWrt.
<Tapper>
how or if*
SamantazFox is now known as Guest889
SamantazFox has joined #openwrt-devel
Guest889 has quit [Read error: Connection reset by peer]
<aiyion>
in 802.11s the `mesh id` is an easy way to distinguish differen mesh networks. Using iw scan on my pc does provide me with the id, if I scan for networks.
<aiyion>
I'm just looking into openwrts lua bindings for iwinfo, but just cannot find, which function I'd call to get the mesh id of a scanned network.
<aiyion>
May somebody give me a hint?
<aiyion>
Or is it just not implemented (yet)?
<aiyion>
I'd expect the binding I'm looking for in https://git.openwrt.org/?p=project/iwinfo.git;a=blob;f=iwinfo_lua.c;h=e49e454d3a81e12a58e5bb546b8ad63ecc59be3b;hb=HEAD
<jow>
karlp: markweston: user data pointers have been rejected in the past already
<jow>
preffered api usage is container_of
* ldir
waves at jow
<jow>
'lo
<ldir>
you're a ubus adult are you not?
<jow>
depends
<ldir>
lol
<ldir>
ubus name spaces are intended to be unique ie, per server right?
<jow>
correct, yes
<jow>
there can't be multiple servers offering the same namespace
<ldir>
cool - my understanding is correct :-)
<jow>
dnsmasq multi instance foo?
* ldir
ducks, hides, whimpers
<ldir>
yes
<ldir>
exactly
<jow>
what exactly is dnsmasq offering via ubus?
<jow>
and wouldn't that (whatever it is) break if you start suffixing it to make it unique per instance?
<jow>
break as in whatever consumes it not finding it anymore because it has a random suffix
<hexa->
blogic_: udhcpsnoop looks interesting, is there also a service that consumes/parses these events?
<ldir>
jow: to be blunt, whatever 'we' as openwrt have passed to simon for it to support via ubus
<ldir>
at the moment, some stats/metrics
<jow>
ah, so "nothing"
<jow>
I mean no netifd, proto handlers, adblock packages whatever depending on it
<ldir>
not AFAIK
<jow>
then it should get suffixed, but only if multiple instances are used
<ldir>
which is what is now happening.
<jow>
preferably in a deterministic manner, so that eventual consumers can infer the expected namespace from the running instance they want to target
<ldir>
but it is a hard startup error that if we say use 'foo' namespace and that namespace is already occupied then dnsmasq should basically scream loudly and exit.
<ldir>
Julian Kornberger added ubus support back in 2018
* ldir
is looking at the football and very, very confused
<blogic_>
hexa-: ucentral-events
<blogic_>
its part of the cloud mgmt stack that I am working on
<hexa->
thanks, I am browsing the TIP repo and couldn't find a dependency link, so had to ask :)
<blogic_>
hexa-: you want to be on the uCentral-trunk branch
<philipp64|work>
Do we support 802.11-2016 (i.e. .11mc) and what hardware (if any) is required for it? I know you need ns or better frame timestamping at the MAC layer…
ecloud has quit [Ping timeout: 480 seconds]
ecloud has joined #openwrt-devel
rejoicetreat has quit [Remote host closed the connection]
rejoicetreat has joined #openwrt-devel
rejoicetreat has quit [Remote host closed the connection]
rejoicetreat has joined #openwrt-devel
goliath has quit [Quit: SIGSEGV]
rejoicetreat has quit [Remote host closed the connection]
rejoicetreat has joined #openwrt-devel
rejoicetreat has quit [Remote host closed the connection]
rmilecki has quit [Ping timeout: 480 seconds]
bluew has quit [Read error: Connection reset by peer]
bluew has joined #openwrt-devel
<karlp>
jow: what's the reasoning? I can think of arguments like, "let's not put any friendly code in the library, make the user do it instead, that will make the library 'light'" and "good enough for kernel, must be good enough for everyone" but is there some good reason I'm missing beyond those?
SamantazFox is now known as Guest934
SamantazFox has joined #openwrt-devel
takimata has quit [Quit: wat?]
<mangix>
rsalvaterra: turns out it's some frenzy after microsoft announced Windows 11 requires TPM. They don't understand that Intel and AMD already implement in in either CPU or chipset. No need for an external one.
<karlp>
iaui, they only implement the sufficient version in particularly recent cpus, and only option for older is a hw module
<philipp64|work>
mangix: well, it depends on the chipset… some of the older Atom64’s don’t.
<philipp64|work>
And it needs to be 1.2 or 2.0, doesn’t it?
Guest934 has quit [Ping timeout: 481 seconds]
<mangix>
philipp64|work: 11 apparently is 2018 CPUs and up
<mangix>
all of those have TPM
<slh>
wait and see, I don't really don't see a cut-off at 2018 to fly for them
<hauke>
I would be suprised if Windows 11 will really demand a TPM
<mangix>
they backed it off to 2017 CPUs looks like
<hauke>
some features will not work and to say it is windows 11 compatible you need it, but I assume it will work
<hauke>
Microsoft would loose too many customers
<mangix>
but again, 2017 CPUs and up all have TPM
<hauke>
there are so many older PCs out there
<hauke>
and without the newest windows you can not so easy use all the Micosoft cloud products, which is very important to them
<mangix>
all I know is, TPM chips are sold out everywhere
takimata has joined #openwrt-devel
<philipp64|work>
I think “chips are sold out everywhere” is probably also true… even uControllers for new cars and refrigerators…
<mangix>
Anyway it's all moot. Windows 10 is supported until 2025
<mangix>
philipp64|work: right but it happened after the 11 announcement
<hauke>
the retail market for TPM chips is probably pretty small compared to the number of chips Dell or Lenovo and so on buy
<hauke>
you can buy Infineon stocks if you think there will be a huge increase in TPM chip demand
<mangix>
right. OEMs don't have this issue
<philipp64|work>
Actually a lot of IoT’s are incorporating TPM. My last project at Gigamon was bringing up secure boot on their switch platforms…
<hauke>
philipp64|work: do they use TPMs for that? Is this not supported by the SoC?
<philipp64|work>
if you mean off-chip TPM’s, yes. The Infineons.
<hauke>
philipp64|work: the recent Lantiq/Intel/MXL SoCs can check a RSA signature of the first thing they load
<hauke>
the check is done by code in the bootrom
<hauke>
and then each step can check the signature of the next before executing it
<philipp64|work>
The Rangeley chip doesn’t have TPM, as I remember.
<philipp64|work>
hauke: right, it’s an incremental measurement.
<hauke>
the Lantiq/Intel/MXL chips are not following the TPM standard
<hauke>
and all this is in the normal SoC, but most devices do not use it
<philipp64|work>
They also used Denverton, and that didn’t have it either.
<philipp64|work>
hauke: “do not use it” … “it” being what?
<hauke>
I was more talking about the MIPS SoCs
<hauke>
most routers I am aware of do not check the signature, they just boot everything
<philipp64|work>
I’d like to see Linux mandate signed kernels…
<mangix>
???
<hauke>
will never happen
<hauke>
;-)
<hauke>
I have to go to sleep
<mangix>
I thought linux was an OS by developers for developers
<mangix>
*kernel
danitool has quit [Remote host closed the connection]