2017-06-01 21:30:13

by Chris Packham

[permalink] [raw]
Subject: Re: [PATCH 3/4] mtd: mchp23k256: add partitioning support

On 02/06/17 06:43, Brian Norris wrote:
> On Wed, May 17, 2017 at 05:29:11PM +0200, Boris Brezillon wrote:
>> Hi Chris,
>>
>> On Wed, 17 May 2017 17:39:07 +1200
>> Chris Packham <[email protected]> wrote:
>>
>>> Setting the of_node for the mtd device allows the generic mtd code to
>>> setup the partitions. Additionally we must specify a non-zero erasesize
>>> for the partitions to be writeable.
>>>
>>> Signed-off-by: Chris Packham <[email protected]>
>>> ---
>>> drivers/mtd/devices/mchp23k256.c | 5 +++++
>>> 1 file changed, 5 insertions(+)
>>>
>>> diff --git a/drivers/mtd/devices/mchp23k256.c b/drivers/mtd/devices/mchp23k256.c
>>> index 2542f5b8b63f..02c6b9dcbd3e 100644
>>> --- a/drivers/mtd/devices/mchp23k256.c
>>> +++ b/drivers/mtd/devices/mchp23k256.c
>>> @@ -143,6 +143,7 @@ static int mchp23k256_probe(struct spi_device *spi)
>>>
>>> data = dev_get_platdata(&spi->dev);
>>>
>>> + mtd_set_of_node(&flash->mtd, spi->dev.of_node);
>>> flash->mtd.dev.parent = &spi->dev;
>>> flash->mtd.type = MTD_RAM;
>>> flash->mtd.flags = MTD_CAP_RAM;
>>> @@ -151,6 +152,10 @@ static int mchp23k256_probe(struct spi_device *spi)
>>> flash->mtd._read = mchp23k256_read;
>>> flash->mtd._write = mchp23k256_write;
>>>
>>> + flash->mtd.erasesize = PAGE_SIZE;
>>> + while (flash->mtd.size & (flash->mtd.erasesize - 1))
>>> + flash->mtd.erasesize >>= 1;
>>> +
>>
>> Can we fix allocate_partition() to properly handle the
>> master->erasesize == 0 case instead of doing that?
>
> Is everything actually ready for the eraseblock size to be 0?

That was my initial motivation for faking it.

> That would
> seem surprising to many applications, I would think. Can you, for
> instance, even use UBI on such a device?

I've tried ext2 and I believe Andrew has tried minix fs. We're talking
SRAM so UBI/UBIFS doesn't really provide much benefit for this use-case.

> BTW, I feel like this check is a little more natural to do with
> 'mtd->flags & MTD_NO_ERASE', rather than checking the (apparently
> meaningless) erasesize.
>
> (I realize there's a later version of these patches, but I figured I'd
> put my comments where the suggestion was brought up.)
>
> Brian
>



2017-06-01 22:23:25

by Brian Norris

[permalink] [raw]
Subject: Re: [PATCH 3/4] mtd: mchp23k256: add partitioning support

On Thu, Jun 01, 2017 at 09:30:07PM +0000, Chris Packham wrote:
> On 02/06/17 06:43, Brian Norris wrote:
> > On Wed, May 17, 2017 at 05:29:11PM +0200, Boris Brezillon wrote:
> >> Can we fix allocate_partition() to properly handle the
> >> master->erasesize == 0 case instead of doing that?
> >
> > Is everything actually ready for the eraseblock size to be 0?
>
> That was my initial motivation for faking it.

Understood. I think it's probably better to avoid hacking drivers like
you were about to, but I was also curious if anyone had thought through
the implications of *not* forcing a non-zero size.

> > That would
> > seem surprising to many applications, I would think. Can you, for
> > instance, even use UBI on such a device?
>
> I've tried ext2 and I believe Andrew has tried minix fs. We're talking
> SRAM so UBI/UBIFS doesn't really provide much benefit for this use-case.

Right. But that's not necessarily true for all NO_ERASE devices, so we'd
probably want to think about that before allowing it.

Brian