s-paul...@ti.com writes:
> Patch adds support for DaVinci SPI driver. This was tested on
> the DM355 EVM using the EEPROM present on the EVM. The SPI EEPROM driver
> for ATMEL(at25) EEPROM chips was used as the client driver.
>
> Signed-off-by: Sandeep Paulraj <s-paul...@ti.com>
Dave is the SPI expert (and maintainer) so I'll defer to him here, but
here's my $0.02.
I'd rather see the shared platform_data struct (struct
davinci_spi_platform_data, and it's related enum, CS etc.) defined in
arch code (mach/spi.h) which is then included from
linux/spi/davinci_spi.h.
Aside from it being a cleaner separation IMHO, it also makes the
upstream submission easier since the platform code can go upstream
first without having a build-time dependency on the driver.
Also minor nit below...
[...]
> # SPI protocol drivers (device/link on bus)
> diff --git a/drivers/spi/davinci_spi.c b/drivers/spi/davinci_spi.c
> new file mode 100644
> index 0000000..173439d
> --- /dev/null
> +++ b/drivers/spi/davinci_spi.c
> +static int davinci_spi_bufs_prep(struct spi_device *spi,
> + struct davinci_spi *davinci_spi,
> + struct davinci_spi_config *spi_cfg)
> +{
> + u32 sPIPC0;
Why the mixed-case name? Just use spipco, or better yet something
more human readable like pin_control.
Kevin
_______________________________________________
Davinci-linux-open-source mailing list
Davinci-linux-open-source@linux.davincidsp.com
http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source
