On Sat, Oct 14, 2023 at 10:18:58PM +0000, Benno Lossin wrote:

On 14.10.23 23:55, Andrew Lunn wrote:

quoted

quoted

quoted

The PHY driver asks the state from the abstractions then the
abstractions ask the state from PHYLIB. So when the abstractions get a
bad value from PHYLIB, the abstractions must return something to the
PHY driver. As I wrote, the abstractions return a random value or an
error. In either way, probably the system cannot continue.

Sure then let the system BUG if it cannot continue. I think that
allowing UB is worse than BUGing.

There is nothing a PHY driver can do which justifies calling BUG().

I was mostly replying to Tomonori's comment "In either way, probably
the system cannot continue.". I think that defaulting the value to
`PHY_ERROR` when it is out of range to be a lot better way of handling
this.

You could actually call phy_error(phydev); That will set the state to
PHY_ERROR and transition the state machine to put the link down.

https://elixir.bootlin.com/linux/latest/source/drivers/net/phy/phy.c#L1213

Its not documented for this use case, its more intended for the
hardware has a problem, and stopped talking to us. But if phylib
itself is messed up, it would seem like a reasonable way to try to
recover. It will dump the stack as well.

	 Andrew