Re: [PATCH] mmc: core: Attach PM domain prior probing of SDIO func driver

[Dmitry Torokhov <dmitry.torokhov@gmail.com>]

Hi Ulf,

On Mon, Jun 01, 2015 at 12:18:25PM +0200, Ulf Hansson wrote:
> Other subsystem buses attach PM domains during probe, but prior calling
> the driver's ->probe() method. During the removal phase, detaching the PM
> domain will be done after invoking the driver's ->remove() callback.
> 
> Convert the SDIO bus to follow this behavior and add error handling.
> 
> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
> ---
> 
> A similar patch as $subject patch has been posted and discussed earlier.
> 
> According to Aaron Lu, who added the initial support for the ACPI PM domain to
> the SDIO bus, this change is safe.
> http://www.spinics.net/lists/linux-mmc/msg28946.html
> 
> ---
>  drivers/mmc/core/sdio_bus.c | 12 ++++++++----
>  1 file changed, 8 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/mmc/core/sdio_bus.c b/drivers/mmc/core/sdio_bus.c
> index bee02e6..7e327a6 100644
> --- a/drivers/mmc/core/sdio_bus.c
> +++ b/drivers/mmc/core/sdio_bus.c
> @@ -137,6 +137,10 @@ static int sdio_bus_probe(struct device *dev)
>  	if (!id)
>  		return -ENODEV;
>  
> +	ret = dev_pm_domain_attach(dev, false);
> +	if (ret == -EPROBE_DEFER)
> +		return ret;

Why do we handle only -EPROBE_DEFER? What about other errors? Maybe
dev_pm_domain_attach() is too chatty?

> +
>  	/* Unbound SDIO functions are always suspended.
>  	 * During probe, the function is set active and the usage count
>  	 * is incremented.  If the driver supports runtime PM,
> @@ -166,6 +170,7 @@ static int sdio_bus_probe(struct device *dev)
>  disable_runtimepm:
>  	if (func->card->host->caps & MMC_CAP_POWER_OFF_CARD)
>  		pm_runtime_put_noidle(dev);
> +	dev_pm_domain_detach(dev, false);

If dev_pm_domain_attach() returned -EEXIST (which means that someone
else attached the device to the domain) should we be detaching it?

I realize that the other buses do the same thing, so these questions are
general.

Thanks.

-- 
Dmitry