[PATCH linux dev-4.7 v2] drivers/fsi: Adjust slave ID based on address

[Eddie James]

Looks good to me.

Reviewed-by: Eddie James <eajames at us.ibm.com>

On 02/09/2017 02:17 PM, Christopher Bostic wrote:
> In order to access slave address ranges > 0x1FFFFF the slave ID
> must be adjusted accordingly in the command encoding:
>
> ID 0: 0x00000000 - 0x001FFFFF
> ID 1: 0x00200000 - 0x003FFFFF
> ID 2: 0x00400000 - 0x005FFFFF
> ID 3: 0x00600000 - 0x007FFFFF
>
> Signed-off-by: Christopher Bostic <cbostic at linux.vnet.ibm.com>
>
> ---
>
> Changes in V2
>      - Remove alignment masking as per recommendation of Alistair Popple
>        <alistair at popple.id.au>
> ---
>   drivers/fsi/fsi-master-gpio.c | 6 +++++-
>   1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/fsi/fsi-master-gpio.c b/drivers/fsi/fsi-master-gpio.c
> index 73d9985..91bdbf2 100644
> --- a/drivers/fsi/fsi-master-gpio.c
> +++ b/drivers/fsi/fsi-master-gpio.c
> @@ -40,6 +40,9 @@
>   #define	FSI_GPIO_CMD_DFLT_LEN	28
>   #define	FSI_GPIO_CMD_CRC_SHIFT	60
>
> +#define	FSI_SLAVE_SHIFT		21
> +#define	FSI_SLAVE_MASK		0x3
> +
>   /* Bus errors */
>   #define	FSI_GPIO_ERR_BUSY	1	/* Slave stuck in busy state */
>   #define	FSI_GPIO_RESP_ERRA	2	/* Any (misc) Error */
> @@ -305,7 +308,8 @@ static void build_abs_ar_command(struct fsi_gpio_msg *cmd, uint64_t mode,
>   	cmd->bits = FSI_GPIO_CMD_DFLT_LEN;
>   	cmd->msg = FSI_GPIO_CMD_DEFAULT;
>   	cmd->msg |= mode;
> -	cmd->msg &= ~FSI_GPIO_CMD_SLAVE_MASK;
> +	if (!slave)
> +		slave = (addr >> FSI_SLAVE_SHIFT) & FSI_SLAVE_MASK;
>   	cmd->msg |= (((uint64_t)slave) << FSI_GPIO_CMD_SLV_SHIFT);
>   	addr &= FSI_GPIO_CMD_ADDR_MASK;
>   	cmd->msg |= (((uint64_t)addr) << FSI_GPIO_CMD_ADDR_SHIFT);