2019-05-22 00:09:28

by Chris Packham

[permalink] [raw]
Subject: [PATCH] mtd: cfi_cmdset_0002: dynamically determine the max sectors

Because PPB unlocking unlocks the whole chip cfi_ppb_unlock() needs to
remember the locked status for each sector so it can re-lock the
unaddressed sectors. Dynamically calculate the maximum number of sectors
rather than using a hardcoded value that is too small for larger chips.

Tested with Spansion S29GL01GS11TFI flash device.

Signed-off-by: Chris Packham <[email protected]>
---
drivers/mtd/chips/cfi_cmdset_0002.c | 13 ++++++++-----
1 file changed, 8 insertions(+), 5 deletions(-)

diff --git a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c
index c8fa5906bdf9..a1a7d334aa82 100644
--- a/drivers/mtd/chips/cfi_cmdset_0002.c
+++ b/drivers/mtd/chips/cfi_cmdset_0002.c
@@ -2533,8 +2533,6 @@ struct ppb_lock {
int locked;
};

-#define MAX_SECTORS 512
-
#define DO_XXLOCK_ONEBLOCK_LOCK ((void *)1)
#define DO_XXLOCK_ONEBLOCK_UNLOCK ((void *)2)
#define DO_XXLOCK_ONEBLOCK_GETLOCK ((void *)3)
@@ -2633,6 +2631,7 @@ static int __maybe_unused cfi_ppb_unlock(struct mtd_info *mtd, loff_t ofs,
int i;
int sectors;
int ret;
+ int max_sectors;

/*
* PPB unlocking always unlocks all sectors of the flash chip.
@@ -2640,7 +2639,11 @@ static int __maybe_unused cfi_ppb_unlock(struct mtd_info *mtd, loff_t ofs,
* first check the locking status of all sectors and save
* it for future use.
*/
- sect = kcalloc(MAX_SECTORS, sizeof(struct ppb_lock), GFP_KERNEL);
+ max_sectors = 0;
+ for (i = 0; i < mtd->numeraseregions; i++)
+ max_sectors += regions[i].numblocks;
+
+ sect = kcalloc(max_sectors, sizeof(struct ppb_lock), GFP_KERNEL);
if (!sect)
return -ENOMEM;

@@ -2689,9 +2692,9 @@ static int __maybe_unused cfi_ppb_unlock(struct mtd_info *mtd, loff_t ofs,
}

sectors++;
- if (sectors >= MAX_SECTORS) {
+ if (sectors >= max_sectors) {
printk(KERN_ERR "Only %d sectors for PPB locking supported!\n",
- MAX_SECTORS);
+ max_sectors);
kfree(sect);
return -EINVAL;
}
--
2.21.0


2019-05-22 06:03:37

by Stefan Roese

[permalink] [raw]
Subject: Re: [PATCH] mtd: cfi_cmdset_0002: dynamically determine the max sectors

On 22.05.19 02:06, Chris Packham wrote:
> Because PPB unlocking unlocks the whole chip cfi_ppb_unlock() needs to
> remember the locked status for each sector so it can re-lock the
> unaddressed sectors. Dynamically calculate the maximum number of sectors
> rather than using a hardcoded value that is too small for larger chips.
>
> Tested with Spansion S29GL01GS11TFI flash device.
>
> Signed-off-by: Chris Packham <[email protected]>

This hardcoded limit always annoyed me at that time, so thanks for
"fixing" this:

Reviewed-by: Stefan Roese <[email protected]>

Thanks,
Stefan

2019-06-14 03:25:49

by Chris Packham

[permalink] [raw]
Subject: Re: [PATCH] mtd: cfi_cmdset_0002: dynamically determine the max sectors

Hi All,

I think this may have got lost in the change of maintainer for mtd.

On 22/05/19 12:06 PM, Chris Packham wrote:
> Because PPB unlocking unlocks the whole chip cfi_ppb_unlock() needs to
> remember the locked status for each sector so it can re-lock the
> unaddressed sectors. Dynamically calculate the maximum number of sectors
> rather than using a hardcoded value that is too small for larger chips.
>
> Tested with Spansion S29GL01GS11TFI flash device.
>
> Signed-off-by: Chris Packham <[email protected]>
> ---
> drivers/mtd/chips/cfi_cmdset_0002.c | 13 ++++++++-----
> 1 file changed, 8 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c
> index c8fa5906bdf9..a1a7d334aa82 100644
> --- a/drivers/mtd/chips/cfi_cmdset_0002.c
> +++ b/drivers/mtd/chips/cfi_cmdset_0002.c
> @@ -2533,8 +2533,6 @@ struct ppb_lock {
> int locked;
> };
>
> -#define MAX_SECTORS 512
> -
> #define DO_XXLOCK_ONEBLOCK_LOCK ((void *)1)
> #define DO_XXLOCK_ONEBLOCK_UNLOCK ((void *)2)
> #define DO_XXLOCK_ONEBLOCK_GETLOCK ((void *)3)
> @@ -2633,6 +2631,7 @@ static int __maybe_unused cfi_ppb_unlock(struct mtd_info *mtd, loff_t ofs,
> int i;
> int sectors;
> int ret;
> + int max_sectors;
>
> /*
> * PPB unlocking always unlocks all sectors of the flash chip.
> @@ -2640,7 +2639,11 @@ static int __maybe_unused cfi_ppb_unlock(struct mtd_info *mtd, loff_t ofs,
> * first check the locking status of all sectors and save
> * it for future use.
> */
> - sect = kcalloc(MAX_SECTORS, sizeof(struct ppb_lock), GFP_KERNEL);
> + max_sectors = 0;
> + for (i = 0; i < mtd->numeraseregions; i++)
> + max_sectors += regions[i].numblocks;
> +
> + sect = kcalloc(max_sectors, sizeof(struct ppb_lock), GFP_KERNEL);
> if (!sect)
> return -ENOMEM;
>
> @@ -2689,9 +2692,9 @@ static int __maybe_unused cfi_ppb_unlock(struct mtd_info *mtd, loff_t ofs,
> }
>
> sectors++;
> - if (sectors >= MAX_SECTORS) {
> + if (sectors >= max_sectors) {
> printk(KERN_ERR "Only %d sectors for PPB locking supported!\n",
> - MAX_SECTORS);
> + max_sectors);
> kfree(sect);
> return -EINVAL;
> }
>

2019-06-14 07:20:14

by Joakim Tjernlund

[permalink] [raw]
Subject: Re: [PATCH] mtd: cfi_cmdset_0002: dynamically determine the max sectors

On Fri, 2019-06-14 at 03:23 +0000, Chris Packham wrote:
>
>
> Hi All,
>
> I think this may have got lost in the change of maintainer for mtd.

We need this too, ATM we have a local hack that just changes MAX_SECTORS to 1024

>
> On 22/05/19 12:06 PM, Chris Packham wrote:
> > Because PPB unlocking unlocks the whole chip cfi_ppb_unlock() needs to
> > remember the locked status for each sector so it can re-lock the
> > unaddressed sectors. Dynamically calculate the maximum number of sectors
> > rather than using a hardcoded value that is too small for larger chips.
> >
> > Tested with Spansion S29GL01GS11TFI flash device.
> >
> > Signed-off-by: Chris Packham <[email protected]>
> > ---
> > drivers/mtd/chips/cfi_cmdset_0002.c | 13 ++++++++-----
> > 1 file changed, 8 insertions(+), 5 deletions(-)
> >
> > diff --git a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c
> > index c8fa5906bdf9..a1a7d334aa82 100644
> > --- a/drivers/mtd/chips/cfi_cmdset_0002.c
> > +++ b/drivers/mtd/chips/cfi_cmdset_0002.c
> > @@ -2533,8 +2533,6 @@ struct ppb_lock {
> > int locked;
> > };
> >
> > -#define MAX_SECTORS 512
> > -
> > #define DO_XXLOCK_ONEBLOCK_LOCK ((void *)1)
> > #define DO_XXLOCK_ONEBLOCK_UNLOCK ((void *)2)
> > #define DO_XXLOCK_ONEBLOCK_GETLOCK ((void *)3)
> > @@ -2633,6 +2631,7 @@ static int __maybe_unused cfi_ppb_unlock(struct mtd_info *mtd, loff_t ofs,
> > int i;
> > int sectors;
> > int ret;
> > + int max_sectors;
> >
> > /*
> > * PPB unlocking always unlocks all sectors of the flash chip.
> > @@ -2640,7 +2639,11 @@ static int __maybe_unused cfi_ppb_unlock(struct mtd_info *mtd, loff_t ofs,
> > * first check the locking status of all sectors and save
> > * it for future use.
> > */
> > - sect = kcalloc(MAX_SECTORS, sizeof(struct ppb_lock), GFP_KERNEL);
> > + max_sectors = 0;
> > + for (i = 0; i < mtd->numeraseregions; i++)
> > + max_sectors += regions[i].numblocks;
> > +
> > + sect = kcalloc(max_sectors, sizeof(struct ppb_lock), GFP_KERNEL);
> > if (!sect)
> > return -ENOMEM;
> >
> > @@ -2689,9 +2692,9 @@ static int __maybe_unused cfi_ppb_unlock(struct mtd_info *mtd, loff_t ofs,
> > }
> >
> > sectors++;
> > - if (sectors >= MAX_SECTORS) {
> > + if (sectors >= max_sectors) {
> > printk(KERN_ERR "Only %d sectors for PPB locking supported!\n",
> > - MAX_SECTORS);
> > + max_sectors);
> > kfree(sect);
> > return -EINVAL;
> > }
> >
>
> ______________________________________________________
> Linux MTD discussion mailing list
> https://nam03.safelinks.protection.outlook.com/?url=http%3A%2F%2Flists.infradead.org%2Fmailman%2Flistinfo%2Flinux-mtd%2F&amp;data=02%7C01%7Cjoakim.tjernlund%40infinera.com%7C32742fa3a7134e77f21908d6f078e67b%7C285643de5f5b4b03a1530ae2dc8aaf77%7C1%7C0%7C636960799408384144&amp;sdata=JidMNGuW7GdQO%2FA%2BBs8Q0mnqt%2BWlDUnjsbCRJIDkDvU%3D&amp;reserved=0

2019-06-14 07:39:55

by Vignesh Raghavendra

[permalink] [raw]
Subject: Re: [PATCH] mtd: cfi_cmdset_0002: dynamically determine the max sectors



On 14/06/19 8:53 AM, Chris Packham wrote:
> Hi All,
>
> I think this may have got lost in the change of maintainer for mtd.
>

I do have this in my queue of patches for mtd/next and will forward to
Miquel once he is back.


> On 22/05/19 12:06 PM, Chris Packham wrote:
>> Because PPB unlocking unlocks the whole chip cfi_ppb_unlock() needs to
>> remember the locked status for each sector so it can re-lock the
>> unaddressed sectors. Dynamically calculate the maximum number of sectors
>> rather than using a hardcoded value that is too small for larger chips.
>>
>> Tested with Spansion S29GL01GS11TFI flash device.
>>
>> Signed-off-by: Chris Packham <[email protected]>

Acked-by: Vignesh Raghavendra <[email protected]>

Regards
Vignesh

>> ---
>> drivers/mtd/chips/cfi_cmdset_0002.c | 13 ++++++++-----
>> 1 file changed, 8 insertions(+), 5 deletions(-)
>>
>> diff --git a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c
>> index c8fa5906bdf9..a1a7d334aa82 100644
>> --- a/drivers/mtd/chips/cfi_cmdset_0002.c
>> +++ b/drivers/mtd/chips/cfi_cmdset_0002.c
>> @@ -2533,8 +2533,6 @@ struct ppb_lock {
>> int locked;
>> };
>>
>> -#define MAX_SECTORS 512
>> -
>> #define DO_XXLOCK_ONEBLOCK_LOCK ((void *)1)
>> #define DO_XXLOCK_ONEBLOCK_UNLOCK ((void *)2)
>> #define DO_XXLOCK_ONEBLOCK_GETLOCK ((void *)3)
>> @@ -2633,6 +2631,7 @@ static int __maybe_unused cfi_ppb_unlock(struct mtd_info *mtd, loff_t ofs,
>> int i;
>> int sectors;
>> int ret;
>> + int max_sectors;
>>
>> /*
>> * PPB unlocking always unlocks all sectors of the flash chip.
>> @@ -2640,7 +2639,11 @@ static int __maybe_unused cfi_ppb_unlock(struct mtd_info *mtd, loff_t ofs,
>> * first check the locking status of all sectors and save
>> * it for future use.
>> */
>> - sect = kcalloc(MAX_SECTORS, sizeof(struct ppb_lock), GFP_KERNEL);
>> + max_sectors = 0;
>> + for (i = 0; i < mtd->numeraseregions; i++)
>> + max_sectors += regions[i].numblocks;
>> +
>> + sect = kcalloc(max_sectors, sizeof(struct ppb_lock), GFP_KERNEL);
>> if (!sect)
>> return -ENOMEM;
>>
>> @@ -2689,9 +2692,9 @@ static int __maybe_unused cfi_ppb_unlock(struct mtd_info *mtd, loff_t ofs,
>> }
>>
>> sectors++;
>> - if (sectors >= MAX_SECTORS) {
>> + if (sectors >= max_sectors) {
>> printk(KERN_ERR "Only %d sectors for PPB locking supported!\n",
>> - MAX_SECTORS);
>> + max_sectors);
>> kfree(sect);
>> return -EINVAL;
>> }
>>
>

--
Regards
Vignesh

2019-06-28 14:33:49

by Vignesh Raghavendra

[permalink] [raw]
Subject: Re: [PATCH] mtd: cfi_cmdset_0002: dynamically determine the max sectors

Hi,

On 22-May-19 5:36 AM, Chris Packham wrote:
> Because PPB unlocking unlocks the whole chip cfi_ppb_unlock() needs to
> remember the locked status for each sector so it can re-lock the
> unaddressed sectors. Dynamically calculate the maximum number of sectors
> rather than using a hardcoded value that is too small for larger chips.
>
> Tested with Spansion S29GL01GS11TFI flash device.
>
> Signed-off-by: Chris Packham <[email protected]>
> ---

Applied to https://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux.git
branch mtd/next.

Regards
Vignesh

> drivers/mtd/chips/cfi_cmdset_0002.c | 13 ++++++++-----
> 1 file changed, 8 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/mtd/chips/cfi_cmdset_0002.c b/drivers/mtd/chips/cfi_cmdset_0002.c
> index c8fa5906bdf9..a1a7d334aa82 100644
> --- a/drivers/mtd/chips/cfi_cmdset_0002.c
> +++ b/drivers/mtd/chips/cfi_cmdset_0002.c
> @@ -2533,8 +2533,6 @@ struct ppb_lock {
> int locked;
> };
>
> -#define MAX_SECTORS 512
> -
> #define DO_XXLOCK_ONEBLOCK_LOCK ((void *)1)
> #define DO_XXLOCK_ONEBLOCK_UNLOCK ((void *)2)
> #define DO_XXLOCK_ONEBLOCK_GETLOCK ((void *)3)
> @@ -2633,6 +2631,7 @@ static int __maybe_unused cfi_ppb_unlock(struct mtd_info *mtd, loff_t ofs,
> int i;
> int sectors;
> int ret;
> + int max_sectors;
>
> /*
> * PPB unlocking always unlocks all sectors of the flash chip.
> @@ -2640,7 +2639,11 @@ static int __maybe_unused cfi_ppb_unlock(struct mtd_info *mtd, loff_t ofs,
> * first check the locking status of all sectors and save
> * it for future use.
> */
> - sect = kcalloc(MAX_SECTORS, sizeof(struct ppb_lock), GFP_KERNEL);
> + max_sectors = 0;
> + for (i = 0; i < mtd->numeraseregions; i++)
> + max_sectors += regions[i].numblocks;
> +
> + sect = kcalloc(max_sectors, sizeof(struct ppb_lock), GFP_KERNEL);
> if (!sect)
> return -ENOMEM;
>
> @@ -2689,9 +2692,9 @@ static int __maybe_unused cfi_ppb_unlock(struct mtd_info *mtd, loff_t ofs,
> }
>
> sectors++;
> - if (sectors >= MAX_SECTORS) {
> + if (sectors >= max_sectors) {
> printk(KERN_ERR "Only %d sectors for PPB locking supported!\n",
> - MAX_SECTORS);
> + max_sectors);
> kfree(sect);
> return -EINVAL;
> }
>