On Tue, Sep 22, 2026 at 08:16:16AM +0200, Greg Kroah-Hartman wrote:
> On Tue, Sep 22, 2026 at 02:10:01PM +0800, Inochi Amaoto wrote:
> > On Tue, Sep 22, 2026 at 07:35:19AM +0200, Greg Kroah-Hartman wrote:
> > > On Tue, Sep 22, 2026 at 10:47:16AM +0800, Inochi Amaoto wrote:
> > > > Currently, the phy subsystem uses ENOSYS for dummy blob functions,
> > > > it does not reflect the function state correctly. As kernel already
> > > > has EOPNOTSUPP for disabled blob function, replace ENOSYS with
> > > > EOPNOTSUPP for a clear return value.
> > > > 
> > > > Signed-off-by: Inochi Amaoto <[email protected]>
> > > > ---
> > > >  include/linux/phy/phy.h | 52 ++++++++++++++++++++---------------------
> > > >  1 file changed, 26 insertions(+), 26 deletions(-)
> > > 
> > > Based on the other patches in this series, this patch now breaks the
> > > users, right?  Shouldn't this all happen at once?
> > > 
> > > thanks,
> > > 
> > > greg k-h
> > 
> > It will break, and the change should happen at one. But I think
> > the affect should be limited as it only affect a special edge
> > case: build kernel with CONFIG_GENERIC_PHY disabled. So the
> > break should be limited.
> > 
> > IIRC I was told to seperate the patch into small part so each
> > maintainer can take its own. Is it fine to squash these patches
> > into one in this a case?
> 
> You can't break bisection of the tree, so if it all has to happen in one
> commit, that's required.
> 

Good to know, I will squash them.

> But really, why is this needed at all?  Who will benefit from this
> change?
> 

This change make the phy subsystem aligned with other subsystem like 
clk, reset... they both use -EOPNOTSUPP. After checking the comment 
in errno.h. I think this is more clear to provide the information 
about unsupported function.

The related discussion is at 
https://lore.kernel.org/linux-phy/20260907114837.2y55l7dfqqrgcka2@skbuf/
After some discussion, Vladimir think it is possible to go, so I
send this patch as an improvement.

Regards,
Inochi

Reply via email to