Re: [PATCH 2/2] spi: Add Baikal-T1 System Boot SPI Controller driver

[Serge Semin <Sergey.Semin@xxxxxxxxxxxxxxxxxxxx>]

On Fri, May 08, 2020 at 01:26:52PM +0300, Andy Shevchenko wrote:
> On Fri, May 8, 2020 at 1:15 PM Serge Semin
> <Sergey.Semin@xxxxxxxxxxxxxxxxxxxx> wrote:
> >
> > On Fri, May 08, 2020 at 01:03:11PM +0300, Andy Shevchenko wrote:
> > > On Fri, May 8, 2020 at 12:37 PM Serge Semin
> > > <Sergey.Semin@xxxxxxxxxxxxxxxxxxxx> wrote:
> > > >
> > > > This SPI-controller is a part of the Baikal-T1 System Controller and
> > > > is based on the DW APB SSI IP-core, but with very limited resources:
> > > > no IRQ, no DMA, only a single native chip-select and just 8 bytes Tx/Rx
> > > > FIFO available. In order to provide a transparent initial boot code
> > > > execution this controller is also utilized by an vendor-specific block,
> > > > which provides an CS0 SPI flash direct mapping interface. Since both
> > > > direct mapping and SPI controller normal utilization are mutual exclusive
> > > > only a one of these interfaces can be used to access an external SPI
> > > > slave device. Taking into account the peculiarities of the controller
> > > > registers and physically mapped SPI flash access, very limited resources
> > > > and seeing the normal usecase of the controller is to access an external
> > > > SPI-nor flash, we decided to create a dedicated SPI driver for it.
> > >
> > > It seems a lot of code.
> > > Why can't you use spi-dw-mmio.c et al.?
> >
> > I said above why. Even though the registers set is similar It's too specific
> > to be integrated into the generic DW SSI driver.
> 
> At least you may do at the beginning is to reuse header spi-dw.h and
> put your stuff under
> spi-dw-baikal.c or so. Then, look at the spi-dw.c and check what can
> be reused (I think a lot).

Nah, It's not a lot, barely a few things. What is useful is the registers map
(though it has to be shifted and most of the registers are unused), IO operations
(two small read-write methods with questionable design) and possibly a few lines
of config setting procedure.

You think I didn't had in mind to integrate this controller support into
the generic driver or resuse something from there? If it was worth a try I would
have done it.

-Sergey

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