Re: [PATCH v3 02/21] spi: dw: Add DWC SSI capability

[Serge Semin <Sergey.Semin@xxxxxxxxxxxxxxxxxxxx>]

On Fri, Oct 02, 2020 at 09:26:07PM +0300, Andy Shevchenko wrote:
> On Fri, Oct 2, 2020 at 8:18 PM Serge Semin
> <Sergey.Semin@xxxxxxxxxxxxxxxxxxxx> wrote:
> >
> > On Fri, Oct 02, 2020 at 01:19:29PM +0300, Andy Shevchenko wrote:
> > > On Fri, Oct 02, 2020 at 01:28:10AM +0300, Serge Semin wrote:
> > > > Currently DWC SSI core is supported by means of setting up the
> > > > core-specific update_cr0() callback. It isn't suitable for multiple
> > > > reasons. First of all having exported several methods doing the same thing
> > > > but for different chips makes the code harder to maintain. Secondly the
> > > > spi-dw-core driver exports the methods, then the spi-dw-mmio driver sets
> > > > the private data callback with one of them so to be called by the core
> > > > driver again. That makes the code logic too complicated. Thirdly using
> > > > callbacks for just updating the CR0 register is problematic, since in case
> > > > if the register needed to be updated from different parts of the code,
> > > > we'd have to create another callback (for instance the SPI device-specific
> > > > parameters don't need to be calculated each time the SPI transfer is
> > > > submitted, so it's better to pre-calculate the CR0 data at the SPI-device
> > > > setup stage).
> > > >
> > > > So keeping all the above in mind let's discard the update_cr0() callbacks,
> > > > define a generic and static dw_spi_update_cr0() method and create the
> > > > DW_SPI_CAP_DWC_SSI capability, which when enabled would activate the
> > > > alternative CR0 register layout.
> > > >
> > > > While at it add the comments to the code path of the normal DW APB SSI
> > > > controller setup to make the dw_spi_update_cr0() method looking coherent.
> > >
> >
> > > What the point to increase indentation level and produce additional churn?
> > > Can't you simply leave functions, unexport them, and call in one conditional of
> > > whatever new function is called?
> >
> > I forgot to mention that in the commit log, there is another reason why it's
> > better to create a generic dw_spi_update_cr0() instead of doing what you suggest.
> > As it will be seen from the following up patches, the dw_spi_update_cr0() function
> > (to be more precise it's successor, but anyway) will be used from the SPI memory
> > ops implementation. So if-else-ing here and there isn't a good idea for
> > maintainability. For the same reason of the maintainability it's better to have a
> > generic method which reflects all the config peculiarities, so in case of any
> > changes they would be not be forgotten to be introduced for both DWC SSI and DW
> > APB SSI parts of the setup procedures. As I see it that overbeats the additional
> > indentation level drawback.
> 

> What I meant is to leave functions as is and call them under conditional
> 
> if ()
>  call one
> else
>  call another

Yeah, I understood what you meant. What you suggest would be a better solution
if I needed to call the dw_spi_update_cr0() method just from a single place of
the driver (but in that case I wouldn't need to replace the callback-based
approach with the Capabilities-based one at all). The thing is that the
dw_spi_update_cr0() will be also called from the SPI memory exec_op() callback
(see patch "[PATCH v3 17/21] spi: dw: Add memory operations support" and the
method dw_spi_update_config() invocation) in the same way as it is called from
the SPI core transfer-one callback. Following your suggestion I would have to
implement the same "if () call one else call another" pattern there too. Copying
it here and there would be a weak design from the maintainability point of view
and from the coding style too. Much better is to create a generic
dw_spi_update_cr0() (later in this patchset it will be renamed to
dw_spi_update_config()...), which would work for both DWC SSI and DW APB SSI by
embedding the "if (is_CAP) call one else call another" into the method itself as
I suggested in this patch.

-Sergey

> 
> 
> -- 
> With Best Regards,
> Andy Shevchenko