[i2c] [PATCH 6/6] Blackfin I2C TWI driver: ADSP-BF52x arch/mach support

[Jean Delvare]

On Tue, 25 Mar 2008 11:55:14 +0800, Bryan Wu wrote:
> On Tue, Mar 25, 2008 at 5:33 AM, Jean Delvare <khali at linux-fr.org> wrote:
> >
> > On Fri, 14 Mar 2008 00:22:40 -0700, Bryan Wu wrote:
> >  > From: Michael Hennerich <michael.hennerich at analog.com>
> >  >
> >  > Signed-off-by: Michael Hennerich <michael.hennerich at analog.com>
> >  > Signed-off-by: Bryan Wu <bryan.wu at analog.com>
> >  > Signed-off-by: Bryan Wu <cooloney at kernel.org>
> >  > ---
> >  >  drivers/i2c/busses/Kconfig |    6 +++---
> >  >  1 files changed, 3 insertions(+), 3 deletions(-)
> >  >
> >  > diff --git a/drivers/i2c/busses/Kconfig b/drivers/i2c/busses/Kconfig
> >  > index 476b0bb..a1351ce 100644
> >  > --- a/drivers/i2c/busses/Kconfig
> >  > +++ b/drivers/i2c/busses/Kconfig
> >  > @@ -100,10 +100,10 @@ config I2C_AU1550
> >  >
> >  >  config I2C_BLACKFIN_TWI
> >  >       tristate "Blackfin TWI I2C support"
> >  > -     depends on BF534 || BF536 || BF537 || BF54x
> >  > +     depends on BF534 || BF536 || BF537 || BF54x || BF522 || BF525 || BF527
> >  >       help
> >  > -       This is the TWI I2C device driver for Blackfin 534/536/537/54x.
> >  > -       This driver can also be built as a module.  If so, the module
> >  > +       This is the TWI I2C device driver for Blackfin 522/525/527/534/536/537/54x.
> >  > +       This driver can also be built as a module. If so, the module
> >  >         will be called i2c-bfin-twi.
> >
> >  Now that the i2c-bfin-twi driver no longer requires special defines in
> >  the arch/mach header files, wouldn't it be more simple to just make it
> >  depend on BLACKFIN? Otherwise we'll have to update the "depends on"
> >  statement over and over again with every new machine.
> >
> 
> Unfortunately, 2 important Blackfin processor family BF533 and  BF561
> do not integrate on-chip i2c twi controller.
> They can use this driver. I plan to define HAVE_BFIN_I2C_TWI in each

I guess you meant "they can NOT use this driver"?

> Blackfin family, then Blackfin I2C TWI driver just "depends on
> BLACKFIN && HAVE_BFIN_I2C_TWI". We don't change this configuration
> again when new machine born.

Well, there's nothing wrong with letting people build this driver even
if the platform doesn't have any supported device. As long as it
doesn't cause a build failure, it's alright. So you could simply use
"depends on BLACKFIN" (which is technically correct) and just document
which machines have the device (as you already did).

But if you prefer to introduce HAVE_BFIN_I2C_TWI, that's fine with me,
it's really as you prefer.

Out of curiosity, isn't it possible to build a single kernel that would
support more than just one Blackfin machine? I wonder if this is
something people could be interested in - I have seen such requirements
for other platforms already.

-- 
Jean Delvare