[U-Boot] [v3 PATCH 02/12] arm, at91: add spi dataflash support for the taurus board

[Andreas Bießmann]

Dear Jagan Teki,

On 31.10.14 13:31, Heiko Schocher wrote:
> Hello Jagan,
> 
> Am 31.10.2014 11:48, schrieb Jagan Teki:
>> On 31 October 2014 13:00, Heiko Schocher<hs at denx.de>  wrote:
>>> Signed-off-by: Heiko Schocher<hs at denx.de>
>>> Reviewed-by: Bo Shen<voice.shen at atmel.com>
>>>
>>> ---
>>>
>>> Changes in v3:
>>> add Reviewed-by from Bo Shen
>>>
>>> Changes in v2:
>>> - add comment from Jagan Teki:
>>>    - remove spi_init_f() from board file
>>>    - remove CONFIG_SYS_SPI_WRITE_TOUT from board config file
>>>      instead define a default in the spi driver ->  new patch for v2
>>>
>>>   board/siemens/taurus/taurus.c | 18 ++++++++++++++++++
>>>   include/configs/taurus.h      | 10 ++++++++++
>>>   2 files changed, 28 insertions(+)
>>>
>>> diff --git a/board/siemens/taurus/taurus.c
>>> b/board/siemens/taurus/taurus.c
>>> index 673b302..76609c7 100644
>>> --- a/board/siemens/taurus/taurus.c
>>> +++ b/board/siemens/taurus/taurus.c
>>> @@ -22,6 +22,8 @@
>>>   #include<asm/arch/gpio.h>
>>>   #include<asm/arch/at91sam9_sdramc.h>
>>>   #include<atmel_mci.h>
>>> +#include<asm/arch/at91_spi.h>
>>> +#include<spi.h>
>>>
>>>   #include<net.h>
>>>   #include<netdev.h>
>>> @@ -127,6 +129,21 @@ int board_early_init_f(void)
>>>          return 0;
>>>   }
>>>
>>> +int spi_cs_is_valid(unsigned int bus, unsigned int cs)
>>> +{
>>> +       return bus == 0&&  cs == 0;
>>> +}
>>> +
>>> +void spi_cs_activate(struct spi_slave *slave)
>>> +{
>>> +       at91_set_gpio_value(TAURUS_SPI_CS_PIN, 0);
>>> +}
>>> +
>>> +void spi_cs_deactivate(struct spi_slave *slave)
>>> +{
>>> +       at91_set_gpio_value(TAURUS_SPI_CS_PIN, 1);
>>> +}
>>> +
>>
>> Please add all these stuff in drivers/spi/atmel_spi.c
>>
>> Did you see any issues?
> 
> Hmm.. this is board specific ... or?

Heiko is right, this is board specific and follows the current way to
implement spi for the atmel driver.
But Jagan is also right, we should move this code to the driver code
cause in fact it is parametrisation of generic code. Unfortunately this
change will blow up this series. I think we should consider this when
rewriting the atmel spi driver for dm, which will be done soon (but I
think not in this merge window). Jagan, is this Ok for you?

Best regards

Andreas Bießmann