Re: [PATCH net-next v2 1/3] net: dsa: microchip: ksz9477: apply a shorter ageing time at once

From: Oleksij Rempel

Date: Tue Nov 17 2026 - 07:00:05 EST


On Mon, Nov 16, 2026 at 11:00:35AM +0000, netdev-bot+sashiko@xxxxxxxxxx wrote:
> Does this flush the dynamic FDB entries of every port on the switch?

Yes. Fast Aging is a control for the whole switch, and so are the Age
Count and Age Period. I checked it on a KSZ9893 with one entry on each
user port: both were gone right after the pulse. v3 says this in the
commit message and the comment.

A flush per port would not help: the stale period applies to the
entries of all ports, so all of them would need the flush.

> The pulse can also fire when the ageing time grows.

Yes, that is on purpose. The commit message has this case: from 4 x 2 s
to 1 x 31 s, an entry stayed about 63 s without the pulse. It happens
only when the Age Count gets smaller. In STP use the count does not get
smaller when the bridge restores the ageing time.

> Is a fixed 2 ms sleep long enough on all chips that use this function?

I can test only the KSZ9893. The comment at the define says so.

pw-bot: cr