Re: 73-usb-net-by-mac.rules is no longer used in Bullseye for USB ethernet devices?
On Tue, 2022-02-22 at 22:32 -0600, Flacusbigotis wrote:
[...]
>
> Feb 22 17:26:11 server1 kernel: [ 205.693604] ax88179_178a 5-1:1.0
> enx001122334455: Failed to read reg index 0x0000: -22
>
> And as you can see in those logs there is an issue with xhci_hcd on
> this
> card and later the USB ethernet driver complains too and says it's
> going to
> randomize the MAC! None of these appear in the logs for the Buster
> OS
> Live USB. So, at least now I know why the MAC changes randomly in
> Bullseye and I also now know that this is a bug that was introduced
> by
> Debian Bullseye.
The change to randomise MAC [1] was introduced in the kernel at the end
of 2019. Note, that the old code didn't check for an error reading the
MAC address from the device and would use a garbage MAC address from
the data that happened to be in memory at the time. The new code (after
a fix [2]) does detect error indirectly.
So, it could be your device never always produced an error reading a
MAC address from it, and that went unnoticed until the recent kernel
changes. Or, it could be buggy kernel drivers can't access the device
properly any more.
Interesting, there's another change [3] to the driver that seems
related to writing a MAC address back to the device, but it also seems
to indicate you may need to use the 'ethtool' program, whatever that
is.
Sorry, I know nothing about all this, I just got curious about your
problem and looked at the source code.
[1] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/drivers/net/usb/ax88179_178a.c?h=v5.10.101&id=9fb137aef34e4eedaa23307d309b0ebe8358fea1
[2] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/drivers/net/usb/ax88179_178a.c?h=v5.10.101&id=2e68890993d0940cedc1f0369282256854293a93
[3] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/drivers/net/usb/ax88179_178a.c?h=v5.10.101&id=78734404ef9c133eac70339415c8028dbe19109a
--
Tixy
Reply to: