Re: [PATCH v2 07/14] mfd: zl3073x: Add components versions register defs

[Ivan Vecera]

On 10. 04. 25 7:41 odp., Andrew Lunn wrote:
> > > > +	/* Take device lock */
> > >
> > > What is a device lock? Why do you need to comment standard guards/mutexes?
> >
> > Just to inform code reader, this is a section that accesses device registers
> > that are protected by this zl3073x device lock.
>
> I didn't see a reply to my question about the big picture. Why is the
> regmap lock not sufficient. Why do you need a device lock.
>
> > > > +	scoped_guard(zl3073x, zldev) {
> > > > +		rc = zl3073x_read_id(zldev, &id);
> > > > +		if (rc)
> > > > +			return rc;
> > > > +		rc = zl3073x_read_revision(zldev, &revision);
> > > > +		if (rc)
> > > > +			return rc;
> > > > +		rc = zl3073x_read_fw_ver(zldev, &fw_ver);
> > > > +		if (rc)
> > > > +			return rc;
> > > > +		rc = zl3073x_read_custom_config_ver(zldev, &cfg_ver);
> > > > +		if (rc)
> > > > +			return rc;
> > > > +	}
> > >
> > > Nothing improved here. Andrew comments are still valid and do not send
> > > v3 before the discussion is resolved.
> >
> > I'm accessing device registers here and they are protected by the device
> > lock. I have to take the lock, register access functions expect this by
> > lockdep_assert.
>
> Because lockdep asserts is not a useful answer. Locks are there to
> protect you from something. What are you protecting yourself from? If
> you cannot answer that question, your locking scheme is built on sand
> and very probably broken.
>
>      Andrew

Hi Andrew,
I have described the locking requirements under different message [v1 
05/28]. Could you please take a look?

Thanks,
Ivan