On 20/12/2018 at 14:14, Axel Lin wrote:
> mcp16502_suspend/resume_noirq is only used by SET_NOIRQ_SYSTEM_SLEEP_PM_OPS
> when CONFIG_PM_SLEEP is defined.
> So use #ifdef CONFIG_PM_SLEEP instead CONFIG_SUSPEND guard.
>
> Signed-off-by: Axel Lin <axel....@ingics.com>
I'm not against this patch and don't know the habits about this... one
remark below...
> ---
> This patch replaces below patch I sent today.
> regulator: mcp16502: Fix build error when !CONFIG_SUSPEND && CONFIG_PM_SLEEP
> Changed the subject line because after commit 308144ce8e51 ("regulator:
> mcp16502: code cleanup")
> it's not related to build error now.
>
> drivers/regulator/mcp16502.c | 7 ++-----
> 1 file changed, 2 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/regulator/mcp16502.c b/drivers/regulator/mcp16502.c
> index 809a664733a4..cb11fb206899 100644
> --- a/drivers/regulator/mcp16502.c
> +++ b/drivers/regulator/mcp16502.c
> @@ -497,7 +497,7 @@ static int mcp16502_probe(struct i2c_client *client,
> return 0;
> }
>
> -#ifdef CONFIG_SUSPEND
> +#ifdef CONFIG_PM_SLEEP
Why not use CONFIG_PM just like ...
> static int mcp16502_suspend_noirq(struct device *dev)
> {
> struct i2c_client *client = to_i2c_client(dev);
> @@ -517,10 +517,7 @@ static int mcp16502_resume_noirq(struct device *dev)
>
> return 0;
> }
> -#else /* !CONFIG_SUSPEND */
> -#define mcp16502_suspend_noirq NULL
> -#define mcp16502_resume_noirq NULL
> -#endif /* !CONFIG_SUSPEND */
> +#endif
>
> #ifdef CONFIG_PM
... here.... or unify both directives to "CONFIG_PM_SLEEP" ?
Otherwise, no problem in adding my:
Acked-by: Nicolas Ferre <nicolas.fe...@microchip.com>
(as Andrei is not with Microchip anymore).
> static const struct dev_pm_ops mcp16502_pm_ops = {
>
--
Nicolas Ferre
