Patchwork [2/8] mmc: fix max_discard_sectors

login
register
mail settings
Submitter NamJae Jeon
Date April 13, 2013, 1:38 p.m.
Message ID <1365860309-21261-1-git-send-email-linkinjeon@gmail.com>
Download mbox | patch
Permalink /patch/236348/
State New
Headers show

Comments

NamJae Jeon - April 13, 2013, 1:38 p.m.
From: Namjae Jeon <namjae.jeon@samsung.com>

https://lkml.org/lkml/2013/4/1/292
As per above discussion, there is possibility that request's __data_len
field may overflow when max_discard_sectors greater than UINT_MAX >> 9

If multiple discard requests get merged, merged discard request's
size exceeds 4GB, there is possibility that merged discard request's
__data_len field may overflow.

This patch fixes this issue.

Reported-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Vivek Trivedi <t.vivek@samsung.com>
Tested-by: Max Filippov <jcmvbkbc@gmail.com>
---
 drivers/mmc/card/queue.c |    2 +-
 drivers/mmc/core/core.c  |    4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)
Adrian Hunter - April 15, 2013, 7:20 a.m.
On 13/04/13 16:38, Namjae Jeon wrote:
> From: Namjae Jeon <namjae.jeon@samsung.com>
> 
> https://lkml.org/lkml/2013/4/1/292
> As per above discussion, there is possibility that request's __data_len
> field may overflow when max_discard_sectors greater than UINT_MAX >> 9
> 
> If multiple discard requests get merged, merged discard request's
> size exceeds 4GB, there is possibility that merged discard request's
> __data_len field may overflow.
> 
> This patch fixes this issue.
> 
> Reported-by: Max Filippov <jcmvbkbc@gmail.com>
> Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
> Signed-off-by: Vivek Trivedi <t.vivek@samsung.com>
> Tested-by: Max Filippov <jcmvbkbc@gmail.com>
> ---
>  drivers/mmc/card/queue.c |    2 +-
>  drivers/mmc/core/core.c  |    4 ++--
>  2 files changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/mmc/card/queue.c b/drivers/mmc/card/queue.c
> index 9447a0e..54726b7 100644
> --- a/drivers/mmc/card/queue.c
> +++ b/drivers/mmc/card/queue.c
> @@ -166,7 +166,7 @@ static void mmc_queue_setup_discard(struct request_queue *q,
>  		return;
>  
>  	queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, q);
> -	q->limits.max_discard_sectors = max_discard;
> +	blk_queue_max_discard_sectors(q, max_discard);
>  	if (card->erased_byte == 0 && !mmc_can_discard(card))
>  		q->limits.discard_zeroes_data = 1;
>  	q->limits.discard_granularity = card->pref_erase << 9;
> diff --git a/drivers/mmc/core/core.c b/drivers/mmc/core/core.c
> index 7b435a3..6ee530c 100644
> --- a/drivers/mmc/core/core.c
> +++ b/drivers/mmc/core/core.c
> @@ -2058,7 +2058,7 @@ static unsigned int mmc_do_calc_max_discard(struct mmc_card *card,
>  	if (card->erase_shift)
>  		max_qty = UINT_MAX >> card->erase_shift;
>  	else if (mmc_card_sd(card))
> -		max_qty = UINT_MAX;
> +		max_qty = UINT_MAX >> 9;

No.  This function calculates max discard for the card not the block layer.
 Apply the block layer limitation at the block layer interface e.g.
in mmc_queue_setup_discard()

>  	else
>  		max_qty = UINT_MAX / card->erase_size;
>  
> @@ -2100,7 +2100,7 @@ unsigned int mmc_calc_max_discard(struct mmc_card *card)
>  	unsigned int max_discard, max_trim;
>  
>  	if (!host->max_discard_to)
> -		return UINT_MAX;
> +		return UINT_MAX >> 9;

Ditto.

>  
>  	/*
>  	 * Without erase_group_def set, MMC erase timeout depends on clock
>
NamJae Jeon - April 15, 2013, 7:52 a.m.
2013/4/15, Adrian Hunter <adrian.hunter@intel.com>:
> On 13/04/13 16:38, Namjae Jeon wrote:
>> From: Namjae Jeon <namjae.jeon@samsung.com>
>>
>> https://lkml.org/lkml/2013/4/1/292
>> As per above discussion, there is possibility that request's __data_len
>> field may overflow when max_discard_sectors greater than UINT_MAX >> 9
>>
>> If multiple discard requests get merged, merged discard request's
>> size exceeds 4GB, there is possibility that merged discard request's
>> __data_len field may overflow.
>>
>> This patch fixes this issue.
>>
>> Reported-by: Max Filippov <jcmvbkbc@gmail.com>
>> Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
>> Signed-off-by: Vivek Trivedi <t.vivek@samsung.com>
>> Tested-by: Max Filippov <jcmvbkbc@gmail.com>
>> ---
>>  drivers/mmc/card/queue.c |    2 +-
>>  drivers/mmc/core/core.c  |    4 ++--
>>  2 files changed, 3 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/mmc/card/queue.c b/drivers/mmc/card/queue.c
>> index 9447a0e..54726b7 100644
>> --- a/drivers/mmc/card/queue.c
>> +++ b/drivers/mmc/card/queue.c
>> @@ -166,7 +166,7 @@ static void mmc_queue_setup_discard(struct
>> request_queue *q,
>>  		return;
>>
>>  	queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, q);
>> -	q->limits.max_discard_sectors = max_discard;
>> +	blk_queue_max_discard_sectors(q, max_discard);
>>  	if (card->erased_byte == 0 && !mmc_can_discard(card))
>>  		q->limits.discard_zeroes_data = 1;
>>  	q->limits.discard_granularity = card->pref_erase << 9;
>> diff --git a/drivers/mmc/core/core.c b/drivers/mmc/core/core.c
>> index 7b435a3..6ee530c 100644
>> --- a/drivers/mmc/core/core.c
>> +++ b/drivers/mmc/core/core.c
>> @@ -2058,7 +2058,7 @@ static unsigned int mmc_do_calc_max_discard(struct
>> mmc_card *card,
>>  	if (card->erase_shift)
>>  		max_qty = UINT_MAX >> card->erase_shift;
>>  	else if (mmc_card_sd(card))
>> -		max_qty = UINT_MAX;
>> +		max_qty = UINT_MAX >> 9;
>
> No.  This function calculates max discard for the card not the block layer.
>  Apply the block layer limitation at the block layer interface e.g.
> in mmc_queue_setup_discard()
Hi Adrian,
Okay, I will remove change in mmc_do_calc_max_discard and mmc_calc_max_discard.

>> @@ -166,7 +166,7 @@ static void mmc_queue_setup_discard(struct
>> request_queue *q,
>>  		return;
>>
>>  	queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, q);
>> -	q->limits.max_discard_sectors = max_discard;
>> +	blk_queue_max_discard_sectors(q, max_discard);
>>  	if (card->erased_byte == 0 && !mmc_can_discard(card))
>>  		q->limits.discard_zeroes_data = 1;
Above change will handle max discard sectors as we have updated
blk_queue_max_discard_sectors to check upper limit of max discard
sectors as UINT_MAX >> 9.

Thanks.
>
>>  	else
>>  		max_qty = UINT_MAX / card->erase_size;
>>
>> @@ -2100,7 +2100,7 @@ unsigned int mmc_calc_max_discard(struct mmc_card
>> *card)
>>  	unsigned int max_discard, max_trim;
>>
>>  	if (!host->max_discard_to)
>> -		return UINT_MAX;
>> +		return UINT_MAX >> 9;
>
> Ditto.
>
>>
>>  	/*
>>  	 * Without erase_group_def set, MMC erase timeout depends on clock
>>
>
>

Patch

diff --git a/drivers/mmc/card/queue.c b/drivers/mmc/card/queue.c
index 9447a0e..54726b7 100644
--- a/drivers/mmc/card/queue.c
+++ b/drivers/mmc/card/queue.c
@@ -166,7 +166,7 @@  static void mmc_queue_setup_discard(struct request_queue *q,
 		return;
 
 	queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, q);
-	q->limits.max_discard_sectors = max_discard;
+	blk_queue_max_discard_sectors(q, max_discard);
 	if (card->erased_byte == 0 && !mmc_can_discard(card))
 		q->limits.discard_zeroes_data = 1;
 	q->limits.discard_granularity = card->pref_erase << 9;
diff --git a/drivers/mmc/core/core.c b/drivers/mmc/core/core.c
index 7b435a3..6ee530c 100644
--- a/drivers/mmc/core/core.c
+++ b/drivers/mmc/core/core.c
@@ -2058,7 +2058,7 @@  static unsigned int mmc_do_calc_max_discard(struct mmc_card *card,
 	if (card->erase_shift)
 		max_qty = UINT_MAX >> card->erase_shift;
 	else if (mmc_card_sd(card))
-		max_qty = UINT_MAX;
+		max_qty = UINT_MAX >> 9;
 	else
 		max_qty = UINT_MAX / card->erase_size;
 
@@ -2100,7 +2100,7 @@  unsigned int mmc_calc_max_discard(struct mmc_card *card)
 	unsigned int max_discard, max_trim;
 
 	if (!host->max_discard_to)
-		return UINT_MAX;
+		return UINT_MAX >> 9;
 
 	/*
 	 * Without erase_group_def set, MMC erase timeout depends on clock