If you have to respin - there are 4 patches in the series.
Sashikos point at bugs in the EAGAIN flow handling in
ndo_set_rx_mode_async. Can you double check that?
Yes, I need to generate a new spin to resolve the EAGAIN issue and
address the other comments.
https://sashiko.dev/#/patchset/20260724142954.3101980-2-vikas.gupta@broadcom.com
https://netdev-ai.bots.linux.dev/sashiko/#/patchset/20260724142954.3101980-2-vikas.gupta@broadcom.com
And the concern about IRQs being in use on shutdown perhaps.
Since the IRQ issue stems from the new shared bng_re RoCE driver, and
it lacks a bng_re_shutdown
implementation to free the IRQs, should we implement this in the
current series, especially given that
the bng_re driver implementation is currently minimal?
Thanks,
Vikas