On 22/06/2015 08:44, Jagan Teki wrote:

On 22 June 2015 at 03:10, Angelo Dureghello angelo@sysam.it wrote:

...

Fix compilation warnings for redefined 'clamp' macro and non-uniform clamp macro types.

Signed-off-by: Angelo Dureghello angelo@sysam.it

drivers/spi/cf_qspi.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/spi/cf_qspi.c b/drivers/spi/cf_qspi.c index 834c5bd..e57e63e 100644 --- a/drivers/spi/cf_qspi.c +++ b/drivers/spi/cf_qspi.c @@ -19,7 +19,6 @@

DECLARE_GLOBAL_DATA_PTR;

-#define clamp(x, low, high) (min(max(low, x), high)) #define to_cf_qspi_slave(s) container_of(s, struct cf_qspi_slave, slave)

struct cf_qspi_slave { @@ -120,7 +119,7 @@ struct spi_slave *spi_setup_slave(unsigned int bus, unsigned int cs, dev->qmr = 2u; else /* Get the closest baud rate */ dev->qmr = clamp(((gd->bus_clk >> 2) + max_hz - 1)/max_hz,

•                                   2u, 255u);
  

•                                   2lu, 255lu);
  

Any Tested-by ?

Hi Jagan,

i am supposing there are no issues here, since 2 and 255 are inside the 8-bit range, and they was unsigned int instead of unsigned long so the change gives no effect to the clamp.

I am still waiting the hardware here for proper mcf / qspi-testing. Will test it then.

Best regards, Angelo Durgehello

Hi Jagan,

On 29/06/2015 22:00, Jagan Teki wrote:

On 22 June 2015 at 12:49, Angelo Dureghello angelo@sysam.it wrote:

...

On 22/06/2015 08:44, Jagan Teki wrote:

...

On 22 June 2015 at 03:10, Angelo Dureghello angelo@sysam.it wrote:

...

Fix compilation warnings for redefined 'clamp' macro and non-uniform clamp macro types.

Signed-off-by: Angelo Dureghello angelo@sysam.it

drivers/spi/cf_qspi.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/spi/cf_qspi.c b/drivers/spi/cf_qspi.c index 834c5bd..e57e63e 100644 --- a/drivers/spi/cf_qspi.c +++ b/drivers/spi/cf_qspi.c @@ -19,7 +19,6 @@

DECLARE_GLOBAL_DATA_PTR;

-#define clamp(x, low, high) (min(max(low, x), high)) #define to_cf_qspi_slave(s) container_of(s, struct cf_qspi_slave, slave)

struct cf_qspi_slave { @@ -120,7 +119,7 @@ struct spi_slave *spi_setup_slave(unsigned int bus, unsigned int cs, dev->qmr = 2u; else /* Get the closest baud rate */ dev->qmr = clamp(((gd->bus_clk >> 2) + max_hz - 1)/max_hz,

•                                   2u, 255u);
  

•                                   2lu, 255lu);
  

replacing u with lu - any reason?

Compiling this file you will get compilation warnings. The reason is that the "unnecessary type check" throws the warnings since clamp (and so included min max) uses different types.

u is unsigned int lu uniforms types to gd->bus_clk, that is unsigned long.

Best regards Angelo Dureghello

...

...

Any Tested-by ?

Hi Jagan,

i am supposing there are no issues here, since 2 and 255 are inside the 8-bit range, and they was unsigned int instead of unsigned long so the change gives no effect to the clamp.

I am still waiting the hardware here for proper mcf / qspi-testing. Will test it then.

thanks!