Re: [PATCH net-next 0/2] net: phy: sfp: Add single-byte SMBus SFP access

[Maxime Chevallier]

On Mon, 24 Feb 2025 17:24:07 +0100
Maxime Chevallier <maxime.chevallier@xxxxxxxxxxx> wrote:

> On Mon, 24 Feb 2025 13:48:19 +0000
> "Russell King (Oracle)" <linux@xxxxxxxxxxxxxxx> wrote:
> 
> > On Mon, Feb 24, 2025 at 02:31:42PM +0100, Andrew Lunn wrote:  
> > > > What do you think will be the effect of such a warning?  Who is the
> > > > target audience?    
> > > 
> > > It will act as a disclaimer. The kernel is doing its best with broken
> > > hardware, but don't blame the kernel when it does not work
> > > correctly....    
> > 
> > Indeed.
> >   
> > > > You can obviously add it, and I don't really care.  But I believe the
> > > > result will be an endless stream of end users worrying about this scary
> > > > warning and wanting to know what they can do about it.  What will be
> > > > your answer?    
> > > 
> > > I agree that the wording needs to be though about. Maybe something
> > > like:
> > > 
> > > This hardware is broken by design, and there is nothing the kernel, or
> > > the community can do about it. The kernel will try its best, but some
> > > standard SFP features are disabled, and the features which are
> > > implemented may not work correctly because of the design errors. Use
> > > with caution, and don't blame the kernel when it all goes horribly
> > > wrong.    
> > 
> > I was hoping for something shorter, but I think it needs to be expansive
> > so that users can fully understand. Another idea based on your
> > suggestion above:
> > 
> > "Please note:
> > This hardware is broken by design. There is nothing that the kernel or
> > community can do to fix it. The kernel will try best efforts, but some
> > features are disabled, other features may be unreliable or sporadically
> > fail. Use with caution. Please verify any problems on hardware that
> > supports multi-byte I2C transactions."
> >   
> 
> I think what's missing in this message is some indication about what is
> actually wrong with the hardware, so :

I realise that I have formulated the sentence above a bit strongly, of
course this is a suggestion :)

> "Please note:
> This SFP cage is accessed via an SMBus only capable of single byte
> transactions. Some features are disabled, other may be unreliable or
> sporadically fail. Use with caution. There is nothing that the kernel
> or community can do to fix it, the kernel will try best efforts. Please
> verify any problems on hardware that supports multi-byte I2C transactions."

Maxime