ChanServ changed the topic of #libreelec to: [ LibreELEC Support Channel ~ current release: (Leia) 9.2.6 ~ No discussion/support for piracy addons/services ~ Log: https://freenode.irclog.whitequark.org/libreelec ]
Turmoil has joined #libreelec
turm01l has quit [Ping timeout: 260 seconds]
Turmoil is now known as turm01l
tsal has quit [Ping timeout: 260 seconds]
tsal has joined #libreelec
oblikoamorale has quit [Ping timeout: 240 seconds]
oblikoamorale has joined #libreelec
_fraggle_ has quit [Remote host closed the connection]
buzzmarshall has quit [Remote host closed the connection]
pyrodex has quit [Quit: What up my glip glops!]
pyrodex has joined #libreelec
chewitt has quit [Quit: Zzz..]
chewitt has joined #libreelec
<rdorsch>
I see ethernet connections with frequently changing IDs there: # ls -l /storage/.cache/connman/ => http://ix.io/2Vc1
<rdorsch>
Is that expected?
gouchi has joined #libreelec
<chewitt>
rdorsch if u-boot doesn't provide a fixed mac address for the device the kernel drivers will normally auto-generate one
gouchi has quit [Client Quit]
<chewitt>
this means mac changes on each boot
<rdorsch>
can that be the root cause that the interface forgets its setting almost constantly?
<chewitt>
if everything is set via DHCP it mostly goes unnoticed
<rdorsch>
But I essentially need to disable the interface, since I run vlan 1, 2, and 7
<rdorsch>
Can I enter u-boot and set a mac address?
<chewitt>
connman stores 'service' profiles with a name that includes the base64 encoded MAC
<rdorsch>
I debian I can use: setenv ethaddr d0:63:b4:00:2b:ac
<chewitt>
this uses CPU serial to derive/set a persistent MAC regardless of what u-boot does/doesn't set
<chewitt>
but see if the u-boot setenv trick works first
<chewitt>
or perhaps append ethaddr=d0:63:b4:00:2b:ac into boot params and see what happens
<rdorsch>
I had a quick look at ethmactool-config
<rdorsch>
At least /proc/cpuinfo produces a serial of all 0: http://ix.io/2Vc5
<rdorsch>
Do you know how I can enter the u-boot shell? In Debian I need just to hit a key during a countdown IIRC.
blackest_mamba has quit [Ping timeout: 268 seconds]
blackest_mamba has joined #libreelec
Gittun has joined #libreelec
<chewitt>
normally it's "press any key" at boot
<chewitt>
I didn't pay much attention to the u-boot config we're using :)
<chewitt>
in ye olden days most boards set a delay
<chewitt>
these days most have a delay of zero, but any key-press (or space/enter) drops you to the u-boot shell
<chewitt>
looking at the output, just adding ethmactool to the default image should work
<chewitt>
or maybe not ..
<chewitt>
i'm in a conf call and not reading the script right, ignore..
lolek has joined #libreelec
<rdorsch>
I tried to get into u-boot a few times, but no success.
<rdorsch>
for the ethmactool: wouldn't it be an option to check for 0x00000000 at then either read the mac from a mac-file of if the mac-file does not exist, use the autogenerated one and store it to the mac-file.
<chewitt>
to a native English speaker "blowing fuses" sounds rather scary, but..
turm01l has quit [Ping timeout: 260 seconds]
<chewitt>
so easiest route forwards is extending the ethmactool scripts to detect fsl compatible and support reading a mac from /storage/.cache/ethmac or similar
<rdorsch>
yes, I fully agree.
<rdorsch>
"Blowing fuses is an irreversible act. If you set a bit from '0' to '1' you can not set it back to '0'." does not seem to be appropriate, if there is the option with the ethmactool.
turm01l has joined #libreelec
Fenster` has quit [Ping timeout: 268 seconds]
emOne has joined #libreelec
ghostcube has joined #libreelec
turm01l has quit [Ping timeout: 260 seconds]
turm01l has joined #libreelec
Turmoil has joined #libreelec
<chewitt>
vpeter: how did LE on BSP kernel solve the MAC issue in the past? (or didn't it?)
turm01l has quit [Read error: Connection reset by peer]
Turmoil is now known as turm01l
Fenster has joined #libreelec
asdf28 has joined #libreelec
Fenster` has joined #libreelec
<rdorsch>
I just noticed there is a .cache/systemd-machine-id which at least survived one reboot. Not sure how unique it is though: I have here d5c5be398b3ed655fadf856260196fee
Fenster has quit [Ping timeout: 240 seconds]
lolek has quit [Ping timeout: 260 seconds]
<chewitt>
it should be unique
<chewitt>
or unique enough .. it's a random generation at first run
<rdorsch>
that is part of ethermac-tool, correct? where is the function called?
<rdorsch>
I cannot exchange it anyways in the squashfs, correct?
asdf28 has quit [Ping timeout: 260 seconds]
k3yp has quit [Ping timeout: 246 seconds]
<rdorsch>
Wouldn't you want to add a default file generation in the else branch: http://ix.io/2Vax
<rdorsch>
Not sure if eth0 is always correct.
<chewitt>
systemd-machine-id is created very early on first run as there are numerous things in systemd that depend upon it
<chewitt>
so while it's bad scripting .. it's fairly safe to assume it's always present, as ethmactool depends on network setup which is much later in the boot process
<rdorsch>
chewitt: the method you implemented for the mac address produces a non-zero and constant mac
buzzmarshall has joined #libreelec
gouchi has quit [Remote host closed the connection]
<rdorsch>
chewitt: what differences do you have wrt. the nighly builds. I noticed that your build is still called 9.95.1 whereas the nighlies are already called 10
RaphGro has quit [Quit: Please remember your own message. It'll be read as soon as possible.]
Tobbi has quit [Ping timeout: 260 seconds]
lolek has quit [Quit: Leaving.]
Zoolook has quit [Ping timeout: 240 seconds]
Kostenko has joined #libreelec
Kostenko has quit [Client Quit]
Gittun has quit [Quit: UPP]
EuroTrash has quit [Remote host closed the connection]
EuroTrash has joined #libreelec
asdf28 has quit [Ping timeout: 268 seconds]
kriger has joined #libreelec
LjL has quit [Read error: Connection reset by peer]
LjL has joined #libreelec
psymin has quit [Quit: Leaving]
ghostcube has quit [Quit: Verlassend]
kriger has quit [Read error: Connection timed out]