Re: [PATCH 0/6] mfd: TPS65910 bug fixes and enhancements

[Graeme Gregory <gg@slimlogic.co.uk>]

On 18/10/11 19:26, Kyle Manna wrote:
> This series of patches fixes a number of small problems with the TPS65910 that
> were discovered on a custom board.
>
> Additionally, the last patch enhances the platform to driver interface using the
> TWL4030 series of regulators as an example.
>
> Kyle Manna (6):
>    mfd: TPS65910: Handle non-existent devices
>    mfd: TPS65910: Add I2C slave address macros
>    mfd: TPS65910: Fix typo that clobbers genirq
>    mfd: TPS65910: Move linux/gpio.h include to header
>    mfd: TPS65910: Fix tps65910_set_voltage
>    mfd: TPS65910: Improve regulator init data
>
>   drivers/mfd/tps65910.c                 |   23 ++-
>   drivers/regulator/tps65910-regulator.c |  246 +++++++++++++++++++++++++-------
>   include/linux/mfd/tps65910.h           |   36 +++++-
>   3 files changed, 247 insertions(+), 58 deletions(-)
>
In my opinion patches 5 and 6 are regressions.

set_voltage_sel was added to core api to allow the voltage selection 
code that a lot of drivers implemented identically to be abstracted to 
the regulator core.

The twl4030 driver method of defining platform data has left us with a 
very inflexible way of adding functionality when a new variant of the 
chip is released. This was remarked on by a number of kernel developers 
when I upstreamed the twl6025 support. I feel copying the twl driver is 
the wrong direction to go in in this regard.

Graeme