Re: i2c-eg20t: regression since i2c_add_numbered_adapter change

[Feng Tang <feng.tang-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>]

Hi Alexander,

On Wed, 22 Aug 2012 09:57:07 +0200
Alexander Stein <alexander.stein-93q1YBGzJSMe9JSWTWOYM3xStJ4P+DSV@public.gmane.org> wrote:

> Hello,
> 
> On Wednesday 22 August 2012 15:29:18, Feng Tang wrote:
> > On Wed, 22 Aug 2012 08:30:35 +0200
> > Alexander Stein <alexander.stein-93q1YBGzJSMe9JSWTWOYM3xStJ4P+DSV@public.gmane.org> wrote:
> > 
> > > Hello,
> > > 
> > > I just noticed the 3.4 linux kernel fails to sucessfully probe the i2c-eg20t 
> > > driver. I returns with EBUSY error. It worked on the 3.0 kernel. To my view it 
> > > is caused the commit 07e8a51ff68353e01d795cceafbac9f54c49132b ( i2c-eg20t: use 
> > > i2c_add_numbered_adapter to get a fixed bus number).
> > > The reason it actually fails is that the i2c-isch driver is registered 
> > > beforehand which gets bus number 0. But this one is the bus number the eg20t 
> > > driver wants to register.
> > 
> > Make sense.
> > 
> > > A possibility is that if i2c_add_numbered_adapter failed with EBUSY just use 
> > > i2c_add_adapter to get at least the driver working, but with a non-fixed bus 
> > > number. Opinions?
> > 
> > Or can we give it a fixed offset, like let the i2c_eg20t controller bus number
> > start with 4? I don't expect there will be more than 4 other i2c controllers 
> > on EG20T compatible platforms.
> 
> Why use a fixed one? Give the driver (and maybe every i2c bus driver) a parameter which sets the base bus number it should use.
> E.g. i2c-eg20t.base-bus-num=2 so it will register the bus numbers starting from 2. If this parameter is unset. It would use the first free one, thus simply using i2c_add_adapter.

The reason we need a fixed number is it is easier for platform code
which needs to register dozens of i2c devices to different controllers
with i2c_register_board_info, and they need provide a bus number for
each i2c device, this _binding_ info is not detectable but have to
be fixed.

Yes, your module parameter "base-bus-num" sounds like a plan too,
at the cost of some extra setting in the kernel cmdline. And I'm fine
with all solutions as long as I can get a _fixed_ bus number so that
I can easily write the platform config code (I guess this is same
for device tree, and even ACPI 5.0)

Thanks,
Feng