[RFC PATCH 1/2] mmc: core: sdio: move mmc_sdio_init_card to a separate function

[RFC PATCH 1/2] mmc: core: sdio: move mmc_sdio_init_card to a separate function

[Kishon Vijay Abraham I]

No functional change. Move mmc_sdio_init_card to a separate
function (_mmc_sdio_power_restore) and invoke it from
mmc_sdio_power_restore. This is in preparation to invoke
_mmc_sdio_power_restore from mmc_sdio_runtime_resume so that
mmc_claim_host can be invoked before mmc_power_up.

Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
---
 drivers/mmc/core/sdio.c |   22 ++++++++++++++--------
 1 file changed, 14 insertions(+), 8 deletions(-)

diff --git a/drivers/mmc/core/sdio.c b/drivers/mmc/core/sdio.c
index b91abed..0cbbfb8 100644
--- a/drivers/mmc/core/sdio.c
+++ b/drivers/mmc/core/sdio.c
@@ -970,15 +970,10 @@ static int mmc_sdio_resume(struct mmc_host *host)
 	return err;
 }
 
-static int mmc_sdio_power_restore(struct mmc_host *host)
+static int _mmc_sdio_power_restore(struct mmc_host *host)
 {
 	int ret;
 
-	BUG_ON(!host);
-	BUG_ON(!host->card);
-
-	mmc_claim_host(host);
-
 	/*
 	 * Reset the card by performing the same steps that are taken by
 	 * mmc_rescan_try_freq() and mmc_attach_sdio() during a "normal" probe.
@@ -1002,14 +997,25 @@ static int mmc_sdio_power_restore(struct mmc_host *host)
 
 	ret = mmc_send_io_op_cond(host, 0, NULL);
 	if (ret)
-		goto out;
+		return ret;
 
 	ret = mmc_sdio_init_card(host, host->card->ocr, host->card,
 				mmc_card_keep_power(host));
 	if (!ret && host->sdio_irqs)
 		mmc_signal_sdio_irq(host);
 
-out:
+	return 0;
+}
+
+static int mmc_sdio_power_restore(struct mmc_host *host)
+{
+	int ret;
+
+	BUG_ON(!host);
+	BUG_ON(!host->card);
+
+	mmc_claim_host(host);
+	ret = _mmc_sdio_power_restore(host);
 	mmc_release_host(host);
 
 	return ret;
-- 
1.7.9.5

[RFC PATCH 2/2] mmc: core: sdio: claim host before power up or power off

[Kishon Vijay Abraham I]

mmc_sdio_runtime_resume and mmc_sdio_runtime_suspend does power up and
power off respectively but does so without claiming the host. Among other
things mmc_claim_host inovkes pm_runtime_get_sync to enable the clocks.
Invoke mmc_claim_host before mmc_power_up and mmc_power_off in
mmc_sdio_runtime_resume and mmc_sdio_runtime_suspend respectively.

This is required since certain platforms (like TI SoCs) access
the controller registers during power up.

Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
---
 drivers/mmc/core/sdio.c |   10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/drivers/mmc/core/sdio.c b/drivers/mmc/core/sdio.c
index 0cbbfb8..40795b2 100644
--- a/drivers/mmc/core/sdio.c
+++ b/drivers/mmc/core/sdio.c
@@ -1023,16 +1023,24 @@ static int mmc_sdio_power_restore(struct mmc_host *host)
 
 static int mmc_sdio_runtime_suspend(struct mmc_host *host)
 {
+	mmc_claim_host(host);
 	/* No references to the card, cut the power to it. */
 	mmc_power_off(host);
+	mmc_release_host(host);
 	return 0;
 }
 
 static int mmc_sdio_runtime_resume(struct mmc_host *host)
 {
+	int ret;
+
+	mmc_claim_host(host);
 	/* Restore power and re-initialize. */
 	mmc_power_up(host, host->card->ocr);
-	return mmc_sdio_power_restore(host);
+	ret = _mmc_sdio_power_restore(host);
+	mmc_release_host(host);
+
+	return ret;
 }
 
 static int mmc_sdio_reset(struct mmc_host *host)
-- 
1.7.9.5

Re: [RFC PATCH 2/2] mmc: core: sdio: claim host before power up or power off

[Ulf Hansson]

On 24 August 2015 at 12:15, Kishon Vijay Abraham I <kishon@ti.com> wrote:
> mmc_sdio_runtime_resume and mmc_sdio_runtime_suspend does power up and
> power off respectively but does so without claiming the host. Among other
> things mmc_claim_host inovkes pm_runtime_get_sync to enable the clocks.
> Invoke mmc_claim_host before mmc_power_up and mmc_power_off in
> mmc_sdio_runtime_resume and mmc_sdio_runtime_suspend respectively.
>
> This is required since certain platforms (like TI SoCs) access
> the controller registers during power up.

I realize the issue. Actually I have been working on similar patch
which take a bit wider approach. That's because there are more cases
where the host needs to be claimed, when it isn't.

I just posted a patch, please have a look and see if it makes sense to
you, you are on cc.
"[PATCH] mmc: core: Keep host claimed while invoking mmc_power_off|up()"

Kind regards
Uffe

>
> Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
> ---
>  drivers/mmc/core/sdio.c |   10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/mmc/core/sdio.c b/drivers/mmc/core/sdio.c
> index 0cbbfb8..40795b2 100644
> --- a/drivers/mmc/core/sdio.c
> +++ b/drivers/mmc/core/sdio.c
> @@ -1023,16 +1023,24 @@ static int mmc_sdio_power_restore(struct mmc_host *host)
>
>  static int mmc_sdio_runtime_suspend(struct mmc_host *host)
>  {
> +       mmc_claim_host(host);
>         /* No references to the card, cut the power to it. */
>         mmc_power_off(host);
> +       mmc_release_host(host);
>         return 0;
>  }
>
>  static int mmc_sdio_runtime_resume(struct mmc_host *host)
>  {
> +       int ret;
> +
> +       mmc_claim_host(host);
>         /* Restore power and re-initialize. */
>         mmc_power_up(host, host->card->ocr);
> -       return mmc_sdio_power_restore(host);
> +       ret = _mmc_sdio_power_restore(host);
> +       mmc_release_host(host);
> +
> +       return ret;
>  }
>
>  static int mmc_sdio_reset(struct mmc_host *host)
> --
> 1.7.9.5
>

Re: [RFC PATCH 2/2] mmc: core: sdio: claim host before power up or power off

[Kishon Vijay Abraham I]

Hi,

On Friday 11 September 2015 06:15 PM, Ulf Hansson wrote:
> On 24 August 2015 at 12:15, Kishon Vijay Abraham I <kishon@ti.com> wrote:
>> mmc_sdio_runtime_resume and mmc_sdio_runtime_suspend does power up and
>> power off respectively but does so without claiming the host. Among other
>> things mmc_claim_host inovkes pm_runtime_get_sync to enable the clocks.
>> Invoke mmc_claim_host before mmc_power_up and mmc_power_off in
>> mmc_sdio_runtime_resume and mmc_sdio_runtime_suspend respectively.
>>
>> This is required since certain platforms (like TI SoCs) access
>> the controller registers during power up.
> 
> I realize the issue. Actually I have been working on similar patch
> which take a bit wider approach. That's because there are more cases
> where the host needs to be claimed, when it isn't.
> 
> I just posted a patch, please have a look and see if it makes sense to
> you, you are on cc.
> "[PATCH] mmc: core: Keep host claimed while invoking mmc_power_off|up()"

Sure Uffe.

-Kishon

> 
> Kind regards
> Uffe
> 
>>
>> Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
>> ---
>>  drivers/mmc/core/sdio.c |   10 +++++++++-
>>  1 file changed, 9 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/mmc/core/sdio.c b/drivers/mmc/core/sdio.c
>> index 0cbbfb8..40795b2 100644
>> --- a/drivers/mmc/core/sdio.c
>> +++ b/drivers/mmc/core/sdio.c
>> @@ -1023,16 +1023,24 @@ static int mmc_sdio_power_restore(struct mmc_host *host)
>>
>>  static int mmc_sdio_runtime_suspend(struct mmc_host *host)
>>  {
>> +       mmc_claim_host(host);
>>         /* No references to the card, cut the power to it. */
>>         mmc_power_off(host);
>> +       mmc_release_host(host);
>>         return 0;
>>  }
>>
>>  static int mmc_sdio_runtime_resume(struct mmc_host *host)
>>  {
>> +       int ret;
>> +
>> +       mmc_claim_host(host);
>>         /* Restore power and re-initialize. */
>>         mmc_power_up(host, host->card->ocr);
>> -       return mmc_sdio_power_restore(host);
>> +       ret = _mmc_sdio_power_restore(host);
>> +       mmc_release_host(host);
>> +
>> +       return ret;
>>  }
>>
>>  static int mmc_sdio_reset(struct mmc_host *host)
>> --
>> 1.7.9.5
>>