This refactor has the secondary benefit of stopping set rx_mode
requests from building up as only the most recent request (before the
work has gotten a chance to run) will be confirmed/executed. Does this
sound good enough to justify the refactor for drivers which do the I/O
under a driver specific spin lock?

On Tue, 6 Jan 2026 at 16:53, Xuan Zhuo [off-list ref] wrote:

If this is a common requirement, it would be better to provide more examples of
driver modifications.

Thanks.

I am planning on modifying these drivers which I can test with
QEMU/vng. Does this list sound good?

e1000
8139cp

(I will do these if the secondary benefit sounds useful)

vmxnet3
pcnet32