[PATCH] mmc: card: Do not scan boot partitions

[PATCH] mmc: card: Do not scan boot partitions

[Johan Rudholm]

Do not scan boot partitions for "soft" partitions, since the boot
partitions are supposed to contain boot code. Silences the following
message during boot:

 mmcblkXbootY: unknown partition table

Signed-off-by: Johan Rudholm <johan.rudholm@stericsson.com>
---
 drivers/mmc/card/block.c |    2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/mmc/card/block.c b/drivers/mmc/card/block.c
index 7e3f453..efdea68 100644
--- a/drivers/mmc/card/block.c
+++ b/drivers/mmc/card/block.c
@@ -1515,6 +1515,8 @@ static struct mmc_blk_data *mmc_blk_alloc_req(struct mmc_card *card,
 	md->disk->queue = md->queue.queue;
 	md->disk->driverfs_dev = parent;
 	set_disk_ro(md->disk, md->read_only || default_ro);
+	if (area_type & MMC_BLK_DATA_AREA_BOOT)
+		md->disk->flags = GENHD_FL_NO_PART_SCAN;
 
 	/*
 	 * As discussed on lkml, GENHD_FL_REMOVABLE should:
-- 
1.7.10

Re: [PATCH] mmc: card: Do not scan boot partitions

[Chris Ball]

Hi Johan,

On Tue, Jun 12 2012, Johan Rudholm wrote:
> Do not scan boot partitions for "soft" partitions, since the boot
> partitions are supposed to contain boot code. Silences the following
> message during boot:
>
>  mmcblkXbootY: unknown partition table
>
> Signed-off-by: Johan Rudholm <johan.rudholm@stericsson.com>
> ---
>  drivers/mmc/card/block.c |    2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/drivers/mmc/card/block.c b/drivers/mmc/card/block.c
> index 7e3f453..efdea68 100644
> --- a/drivers/mmc/card/block.c
> +++ b/drivers/mmc/card/block.c
> @@ -1515,6 +1515,8 @@ static struct mmc_blk_data *mmc_blk_alloc_req(struct mmc_card *card,
>  	md->disk->queue = md->queue.queue;
>  	md->disk->driverfs_dev = parent;
>  	set_disk_ro(md->disk, md->read_only || default_ro);
> +	if (area_type & MMC_BLK_DATA_AREA_BOOT)
> +		md->disk->flags = GENHD_FL_NO_PART_SCAN;
>  
>  	/*
>  	 * As discussed on lkml, GENHD_FL_REMOVABLE should:

Would it be impossible to add a partition table to mmcblkXbootY,
or just unlikely?

Thanks,

- Chris.
-- 
Chris Ball   <cjb@laptop.org>   <http://printf.net/>
One Laptop Per Child

Re: [PATCH] mmc: card: Do not scan boot partitions

[Johan Rudholm]

Hi Chris,

2012/6/20 Chris Ball <cjb@laptop.org>:
> Hi Johan,
>
> On Tue, Jun 12 2012, Johan Rudholm wrote:
>> Do not scan boot partitions for "soft" partitions, since the boot
>> partitions are supposed to contain boot code. Silences the following
>> message during boot:
>>
>>  mmcblkXbootY: unknown partition table
>>
>> Signed-off-by: Johan Rudholm <johan.rudholm@stericsson.com>
>> ---
>>  drivers/mmc/card/block.c |    2 ++
>>  1 file changed, 2 insertions(+)
>>
>> diff --git a/drivers/mmc/card/block.c b/drivers/mmc/card/block.c
>> index 7e3f453..efdea68 100644
>> --- a/drivers/mmc/card/block.c
>> +++ b/drivers/mmc/card/block.c
>> @@ -1515,6 +1515,8 @@ static struct mmc_blk_data *mmc_blk_alloc_req(struct mmc_card *card,
>>       md->disk->queue = md->queue.queue;
>>       md->disk->driverfs_dev = parent;
>>       set_disk_ro(md->disk, md->read_only || default_ro);
>> +     if (area_type & MMC_BLK_DATA_AREA_BOOT)
>> +             md->disk->flags = GENHD_FL_NO_PART_SCAN;
>>
>>       /*
>>        * As discussed on lkml, GENHD_FL_REMOVABLE should:
>
> Would it be impossible to add a partition table to mmcblkXbootY,
> or just unlikely?

I believe this is just an unlikely scenario since the boot partitions
are supposed to contain boot code and also because of their relatively
small size (a few MBs), but there is no technical problem. Perhaps it
should be left up to the user? RPMB partitions are an example of
partitions that would actually generate an error on partition scanning
(because they require authentication), but this is not the case for
boot partitions.

Kind regards, Johan

Re: [PATCH] mmc: card: Do not scan boot partitions

[Ulf Hansson]

Hi Johan and Chris,

On 20 June 2012 09:27, Johan Rudholm <jrudholm@gmail.com> wrote:
> Hi Chris,
>
> 2012/6/20 Chris Ball <cjb@laptop.org>:
>> Hi Johan,
>>
>> On Tue, Jun 12 2012, Johan Rudholm wrote:
>>> Do not scan boot partitions for "soft" partitions, since the boot
>>> partitions are supposed to contain boot code. Silences the following
>>> message during boot:
>>>
>>>  mmcblkXbootY: unknown partition table
>>>
>>> Signed-off-by: Johan Rudholm <johan.rudholm@stericsson.com>
>>> ---
>>>  drivers/mmc/card/block.c |    2 ++
>>>  1 file changed, 2 insertions(+)
>>>
>>> diff --git a/drivers/mmc/card/block.c b/drivers/mmc/card/block.c
>>> index 7e3f453..efdea68 100644
>>> --- a/drivers/mmc/card/block.c
>>> +++ b/drivers/mmc/card/block.c
>>> @@ -1515,6 +1515,8 @@ static struct mmc_blk_data *mmc_blk_alloc_req(struct mmc_card *card,
>>>       md->disk->queue = md->queue.queue;
>>>       md->disk->driverfs_dev = parent;
>>>       set_disk_ro(md->disk, md->read_only || default_ro);
>>> +     if (area_type & MMC_BLK_DATA_AREA_BOOT)
>>> +             md->disk->flags = GENHD_FL_NO_PART_SCAN;
>>>
>>>       /*
>>>        * As discussed on lkml, GENHD_FL_REMOVABLE should:
>>
>> Would it be impossible to add a partition table to mmcblkXbootY,
>> or just unlikely?
>
> I believe this is just an unlikely scenario since the boot partitions
> are supposed to contain boot code and also because of their relatively
> small size (a few MBs), but there is no technical problem. Perhaps it
> should be left up to the user? RPMB partitions are an example of
> partitions that would actually generate an error on partition scanning
> (because they require authentication), but this is not the case for
> boot partitions.
>

To me this looks like a fair constraint we can put on the boot
partitions. It does not make sense to have a partition table in there.
So I say go ahead with this patch.

Acked-by: Ulf Hansson <ulf.hansson@linaro.org>

Kind regards
Ulf Hansson

Re: [PATCH] mmc: card: Do not scan boot partitions

[Linus Walleij]

On Wed, Jun 20, 2012 at 9:27 AM, Johan Rudholm <jrudholm@gmail.com> wrote:
> 2012/6/20 Chris Ball <cjb@laptop.org>:
>>
>> Would it be impossible to add a partition table to mmcblkXbootY,
>> or just unlikely?
>
> I believe this is just an unlikely scenario since the boot partitions
> are supposed to contain boot code and also because of their relatively
> small size (a few MBs), but there is no technical problem. Perhaps it
> should be left up to the user?

Like we add a Kconfig bool option "support partition scanning on boot
partitions" and then we default that to "n", so systems that really
want to do that can turn it on the day they need it.

Could you augment the patch like that?

Yours,
Linus Walleij

Re: [PATCH] mmc: card: Do not scan boot partitions

[Johan Rudholm]

Hi Linus,

2012/6/20 Linus Walleij <linus.walleij@linaro.org>:
> On Wed, Jun 20, 2012 at 9:27 AM, Johan Rudholm <jrudholm@gmail.com> wrote:
>> 2012/6/20 Chris Ball <cjb@laptop.org>:
>>>
>>> Would it be impossible to add a partition table to mmcblkXbootY,
>>> or just unlikely?
>>
>> I believe this is just an unlikely scenario since the boot partitions
>> are supposed to contain boot code and also because of their relatively
>> small size (a few MBs), but there is no technical problem. Perhaps it
>> should be left up to the user?
>
> Like we add a Kconfig bool option "support partition scanning on boot
> partitions" and then we default that to "n", so systems that really
> want to do that can turn it on the day they need it.
>
> Could you augment the patch like that?

Sure, I'll send an updated version with this functionality!

//Johan

Re: [PATCH] mmc: card: Do not scan boot partitions

[Chris Ball]

Hi,

On Wed, Jun 20 2012, Linus Walleij wrote:
> On Wed, Jun 20, 2012 at 9:27 AM, Johan Rudholm <jrudholm@gmail.com> wrote:
>> 2012/6/20 Chris Ball <cjb@laptop.org>:
>>>
>>> Would it be impossible to add a partition table to mmcblkXbootY,
>>> or just unlikely?
>>
>> I believe this is just an unlikely scenario since the boot partitions
>> are supposed to contain boot code and also because of their relatively
>> small size (a few MBs), but there is no technical problem. Perhaps it
>> should be left up to the user?
>
> Like we add a Kconfig bool option "support partition scanning on boot
> partitions" and then we default that to "n", so systems that really
> want to do that can turn it on the day they need it.

I'm not so happy with adding Kconfig options for configuring driver
behavior -- it's usually a failure to just do the right thing.

If we think that people might want to use a partition table on their
boot partitions then we should just support doing that, as we do now.

In fact, the original patch submission for supporting boot partitions
used a Kconfig entry for turning on boot partition support, but we
decided to ditch the Kconfig entry and made it just happen by default.
I think we should continue that precedent and avoid a new entry here.

Thanks,

- Chris.
-- 
Chris Ball   <cjb@laptop.org>   <http://printf.net/>
One Laptop Per Child

Re: [PATCH] mmc: card: Do not scan boot partitions

[Linus Walleij]

On Wed, Jun 20, 2012 at 7:38 PM, Chris Ball <cjb@laptop.org> wrote:
> On Wed, Jun 20 2012, Linus Walleij wrote:
>>
>> Like we add a Kconfig bool option "support partition scanning on boot
>> partitions" and then we default that to "n", so systems that really
>> want to do that can turn it on the day they need it.
>
> I'm not so happy with adding Kconfig options for configuring driver
> behavior -- it's usually a failure to just do the right thing.

This is not for drivers but for the core. But I get the idea...

> If we think that people might want to use a partition table on their
> boot partitions then we should just support doing that, as we do now.

OK in that case I think it's better to disallow partition tables
on the boot partitions, then the day someone has a valid usecase
we can add it back.

So then the patch is good as it is to me:
Acked-by: Linus Walleij <linus.walleij@linaro.org>

Yours,
Linus Walleij