Re: [PATCH 0/2] i2c: Don't assume bus nr 0 if none was specified

[Karol Lewandowski <k.lewandowsk@samsung.com>]

On 16.03.2012 13:19, Karol Lewandowski wrote:

> Hi,
> 
> i2c controller drivers used to assume bus number 0 when none (-1) was specified.
> 
> This worked on non-device tree systems, where one could explicitly specify
> bus number via platform data.  On DT-enabled systems bus number is always -1.
> 
> Some drivers assume bus number 0 when -1 is specified.  This patchset kills
> this logic and switches to dynamic bus allocation (default when -1 is provided[1]).


[ I must have lost actual problem description while rewording
  message itself... ]

Problem arises when multiple drivers (or multiple instances
of one driver) try to assume the same fixed bus number (0).

This causes simply causes i2c_add_numbered_bus() to fail.
Leaving -1 works perfectly, as registration function switches
to dynamic id registration.

> 
> [ I've found out this when I've tried to instantiate more than one i2c-gpio
>   instance on DT-enabled system. I've, hopefully, fixed few other drivers too. ]
> 
> [1] Introduced by Grant Likely in 488bf314b ("i2c: Allow i2c_add_numbered_adapter()
>     to assign a bus id")
> 
> Karol Lewandowski (2):
>   i2c-pxa: Drop leftover comment
>   i2c: Dynamically assign adapter id if it wasn't explictly specified
> 
>  drivers/i2c/busses/i2c-gpio.c         |    7 +------
>  drivers/i2c/busses/i2c-octeon.c       |    2 +-
>  drivers/i2c/busses/i2c-pca-platform.c |    2 +-
>  drivers/i2c/busses/i2c-pxa.c          |    5 -----
>  drivers/i2c/busses/i2c-versatile.c    |    9 ++-------
>  5 files changed, 5 insertions(+), 20 deletions(-)
> 



-- 
Karol Lewandowski | Samsung Poland R&D Center | Linux/Platform