lshw give me

but ifconfig seem correct btw

Been following this thread because it is of interest to me.
eg: I have wireless turned off in my fibre modem but was considering turning it on only for very specific purposes.

So ...
What is left if the only protection left seems to be WPA2?

Best,

A.

I've actually looked into firewall filtering by MAC before, for a fairly large public Wi-Fi network. My conclusion, unfortunately, is that these days it is essentially useless, because (1) it is broadcast in the clear, and (2) people can and do eavesdrop on your MAC addresses and spoof them. It's as easy as changing a system setting. Which means (3) filtering by MAC is basically worthless. Worse yet, (4) many devices actually stop working properly if you force them not to randomize their MACs. If they even let you turn it off in the first place. Some devices no longer even offer the option.

Given this state of affairs, it's hardly surprising that the firmware driver will automatically do MAC randomization upon startup. Long gone are the good ole days where you can count on MACs being a reliable device identifier.

I suppose it could store it textually?
would be 37363a30633a64313a34363a66373a3839 in hex

Good idea but no luck. I think we've excluded that the initial mac address is hardcoded in the firmware.

Hi pcalvert. That's a good thought, but not applicable to my setup. The settings you are referring to are wifi.scan-rand-mac-address and wifi.mac-address-randomization. I use MAC address filtering in my router, so definitely don't want my devices to connect using randomized MAC addresses. I can confirm that NetworkManager is not using any of these settings:

$ cd /etc/NetworkManager/
$ sudo grep -r 'mac-address' *
[no hits]

Plus, I dual boot Devuan and Tiny Core Linux. In the latter, MAC address of my wireless card is the same as what Devuan is showing me (i.e., 54:2a:a2:6a:xx:xx).

Ralph, I think your initial theory is sounds right: There must be a hardware MAC address, which the driver or firmware changes to what I think of as my "actual" MAC address.  A small amount of time (~2 seconds) elapses between wlan0 being brought up and the driver/firmware changing the MAC address.

EDIT: Ralph, if you run my trivial script in reply #2 as a startup job, can you reproduce my observation that MAC address changes? (Although exactly what happens--and when--likely depends on the specific hardware, driver, and firmware.)

EDIT: would be odd to do

$ xxd -p mt7662.bin | tr -d '\n' | grep 760cd146f789
[no hits]
$ xxd -p mt7662_rom_patch.bin | tr -d '\n' | grep 760cd146f789
[no hits]

Maybe I'm not looking the right way.

Perhaps the MAC is changed by firmware?

Hi ralph.ronnquist. I'd love to get to the bottom of this if you can help me. I tried removing the firmware files to see what would happen. If I boot without the firmware, I get this (which is more or less what I expected):

$ cat /sys/class/net/wlan0/address
cat: /sys/class/net/wlan0/address: No such file or directory

What would be the best way to test your theory?

Although in my tests on Devuan, the above command gives a nonsense address if run in the first two seconds after booting. Does anyone know why? If I run this little script at boot:

#!/bin/sh
while true; do
    cat /sys/class/net/wlan0/address >>/home/bruno/Deskop/log.txt
    sleep 1
done

The first two lines in the log file show the unexpected 76:0c:d1:46:f7:89. Lines three onwards show the device's true mac address, which is 54:2a:a2:6a:xx:xx (redacted for privacy).

I've always wondered why wlan0's mac address in /sys/class/net/wlan0/address is only accurate approximately 2 seconds after boot. Does anyone have any idea?

in Devuan 4 (that run as a Virtual Machine)

in the xfce GUI Network connection, under the tab Wi-Fi I can see Device: wlan0 (a MAC address )

with nmcli

So I was wondering which one give the true "physical"/OEM MAC address

Thanks