[PATCH 2/3] net: fec: implement dtb property phy-reset-durationy

[Sascha Hauer]

On Tue, Nov 17, 2015 at 03:38:43PM +0100, Stefan Christ wrote:
> Hi,
> 
> > > Yeah, that would be the obvious and plausible thing to do, but the driver in
> > > the linux kernel does it this way and it's documented in the device tree
> > > bindings.  So I'm not sure if we can change the default behavior here.
> > 
> > Yes, we can ;)
> > 
> > The rationale behind this code may be: The reset duration is
> > unreasonably high (>1s), so that it to a sane default (1ms).
> > 
> > I suggest to drop the check completely. If someone ever starts barebox
> > with an insanely high value (maybe he thought the value is in us) then
> > he won't notice the mistake if barebox silently falls back to 1ms as
> > duration. If instead barebox waits for multiple seconds then he will
> > notice and it the device tree which is what we want.
> 
> Ok. So what about printing a warn message that case? Then everybody should
> notice.

No, I don't think that's necessary. Such an overlong delay should be
noticeable already and give the right hint. Let's save the binary space.

> 
> When I send a v2 patch, should the description in the device tree binding be
> updated, too?

No, the bindings docs are directly taken from the kernel and changes
will be lost the next I update them.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |