On 7 April 2015 at 20:49, Simon Glass sjg@chromium.org wrote:
On 7 April 2015 at 20:46, Simon Glass sjg@chromium.org wrote:
Hi Masahiro,
On 5 April 2015 at 21:22, Masahiro Yamada yamada.masahiro@socionext.com wrote:
Hi Pavel,
2015-04-02 21:47 GMT+09:00 Pavel Machek pavel@denx.de:
Hi!
> Signed-off-by: Masahiro Yamada yamada.masahiro@socionext.com > --- > > arch/arm/Kconfig | 6 ++++++ > configs/socfpga_arria5_defconfig | 3 --- > configs/socfpga_cyclone5_defconfig | 3 --- > configs/socfpga_socrates_defconfig | 3 --- > 4 files changed, 6 insertions(+), 9 deletions(-) > > diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig > index b25a4a9..6ae48c6 100644 > --- a/arch/arm/Kconfig > +++ b/arch/arm/Kconfig > @@ -618,11 +618,17 @@ config TARGET_SOCFPGA_ARRIA5 > bool "Support socfpga_arria5" > select CPU_V7 > select SUPPORT_SPL > + select DM > + select DM_SPI_FLASH > + select DM_SPI
Are you sure?
Yes, I'm sure.
config DM_SPI_FLASH bool "Enable Driver Model for SPI flash" depends on DM && SPI help
DM_SPI_FLASH depends on SPI, you can't just select it...
This dependency is wrong. Before this patch, it must be fixed.
Please see 01/12 of this series: http://patchwork.ozlabs.org/patch/456458/
But I'm not sure.
With this, you made CONFIG_DM_SPI_FLASH mandatory... even for users that have no SPI flash. There must be better solution...?
In my opinion,
CONFIG_DM_* should be user-unconfigurable.
CONFIG_DM_* should not enable/disable a feature, but should define how it is implemented.
For example,
!defined(CONFIG_SPI) --> SPI is disabled
defined(CONFIG_SPI) && defined(CONFIG_DM_SPI) --> Driver Model SPI is enabled
defined(CONFIG_SPI) && !defined(CONFIG_DM_SPI) --> ad-hoc SPI is enabled
Likewise, I think CONFIG_SPI_FLASH should decide if the system has SPI flash.
This sounds right to me. In fact it is important, since we want to remove the CONFIG_DM options eventually. There are only there as a migration aid.
Acked-by: Simon Glass sjg@chromium.org
Applied to u-boot-dm/next, thanks!