mtd support for large NAND

mtd support for large NAND

[Charles Manning]

Last I looked (2.6.23), mtd_info is limited by a u32 size field to 4Gbytes of 
flash. That's no longer very large and is limiting some applications.

Is anyone working on this?

-- Charles

Re: mtd support for large NAND

[Artem Bityutskiy]

On Thu, 2007-10-11 at 09:11 +1300, Charles Manning wrote:
> Last I looked (2.6.23), mtd_info is limited by a u32 size field to 4Gbytes of 
> flash. That's no longer very large and is limiting some applications.
> 
> Is anyone working on this?

Yeah, the problem is known, but nobody is working on the fix AFAIK.

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)

Re: mtd support for large NAND

[falls huang]

Hello !

  How can I simulate a large nandflash(>4G) in linux PC ? I want to
take a look at it, but I don't have any develop-board that has 4G

On 10/11/07, Artem Bityutskiy <dedekind@infradead.org> wrote:
> On Thu, 2007-10-11 at 09:11 +1300, Charles Manning wrote:
> > Last I looked (2.6.23), mtd_info is limited by a u32 size field to 4Gbytes of
> > flash. That's no longer very large and is limiting some applications.
> >
> > Is anyone working on this?
>
> Yeah, the problem is known, but nobody is working on the fix AFAIK.
>
> --
> Best regards,
> Artem Bityutskiy (Битюцкий Артём)
>
>
> ______________________________________________________
> Linux MTD discussion mailing list
> http://lists.infradead.org/mailman/listinfo/linux-mtd/
>


-- 
Regards
Falls Huang

Re: mtd support for large NAND

[Charles Manning]

On Friday 12 October 2007 21:36:09 falls huang wrote:
> Hello !
>
>   How can I simulate a large nandflash(>4G) in linux PC ? I want to
> take a look at it, but I don't have any develop-board that has 4G

There are a few problems you face here. Firstly, mtd does not support >4G 
which will need fixing.

Secondly, even if your system has > 4G it might not be all simultaneously 
accessible and you will need to use highmem as a way to do temporary mappings 
to access this memory.

I might be tempted to try do this using some sort of userspace driver then 
just backing the driver with regular files (eg. 8x1G files to simulate 8G of 
NAND).


>
> On 10/11/07, Artem Bityutskiy <dedekind@infradead.org> wrote:
> > On Thu, 2007-10-11 at 09:11 +1300, Charles Manning wrote:
> > > Last I looked (2.6.23), mtd_info is limited by a u32 size field to
> > > 4Gbytes of flash. That's no longer very large and is limiting some
> > > applications.
> > >
> > > Is anyone working on this?
> >
> > Yeah, the problem is known, but nobody is working on the fix AFAIK.
> >
> > --
> > Best regards,
> > Artem Bityutskiy (Битюцкий Артём)
> >
> >
> > ______________________________________________________
> > Linux MTD discussion mailing list
> > http://lists.infradead.org/mailman/listinfo/linux-mtd/

Re: mtd support for large NAND

[Artem Bityutskiy]

On Tue, 2007-10-16 at 13:27 +1300, Charles Manning wrote:
> On Friday 12 October 2007 21:36:09 falls huang wrote:
> > Hello !
> >
> >   How can I simulate a large nandflash(>4G) in linux PC ? I want to
> > take a look at it, but I don't have any develop-board that has 4G
> 
> There are a few problems you face here. Firstly, mtd does not support >4G 
> which will need fixing.
> 
> Secondly, even if your system has > 4G it might not be all simultaneously 
> accessible and you will need to use highmem as a way to do temporary mappings 
> to access this memory.
> 
> I might be tempted to try do this using some sort of userspace driver then 
> just backing the driver with regular files (eg. 8x1G files to simulate 8G of 
> NAND).

Or create something like blk2mtd.

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)

Re: mtd support for large NAND

[Jörn Engel]

On Fri, 12 October 2007 16:36:09 +0800, falls huang wrote:
> 
>   How can I simulate a large nandflash(>4G) in linux PC ? I want to
> take a look at it, but I don't have any develop-board that has 4G

Something like this would work:

$ truncate 5GiB foo
$ qemu -s -no-kqemu -hda rootfs -hdb foo -kernel arch/i386/boot/bzImage \
       -append "ro root=/dev/hda block2mtd.block2mtd=/dev/hdb"

Instead of truncate you can also use dd:
$ dd if=/dev/zero of=foo bs=1 seek=5368709119

Jörn

-- 
When you close your hand, you own nothing. When you open it up, you
own the whole world.
-- Li Mu Bai in Tiger & Dragon

Re: mtd support for large NAND

[Charles Manning]

On Tuesday 16 October 2007 23:47:27 Jörn Engel wrote:
> On Fri, 12 October 2007 16:36:09 +0800, falls huang wrote:
> >   How can I simulate a large nandflash(>4G) in linux PC ? I want to
> > take a look at it, but I don't have any develop-board that has 4G
>
> Something like this would work:
>
> $ truncate 5GiB foo
> $ qemu -s -no-kqemu -hda rootfs -hdb foo -kernel arch/i386/boot/bzImage \
>        -append "ro root=/dev/hda block2mtd.block2mtd=/dev/hdb"
>
> Instead of truncate you can also use dd:
> $ dd if=/dev/zero of=foo bs=1 seek=5368709119

Plus modifying mtd to handle partitions greater than the u32 size limit.

-- Charles

Re: mtd support for large NAND

[Jörn Engel]

On Wed, 17 October 2007 08:47:46 +1300, Charles Manning wrote:
> On Tuesday 16 October 2007 23:47:27 Jörn Engel wrote:
> > On Fri, 12 October 2007 16:36:09 +0800, falls huang wrote:
> > >   How can I simulate a large nandflash(>4G) in linux PC ? I want to
> > > take a look at it, but I don't have any develop-board that has 4G
> >
> > Something like this would work:
> >
> > $ truncate 5GiB foo
> > $ qemu -s -no-kqemu -hda rootfs -hdb foo -kernel arch/i386/boot/bzImage \
> >        -append "ro root=/dev/hda block2mtd.block2mtd=/dev/hdb"
> >
> > Instead of truncate you can also use dd:
> > $ dd if=/dev/zero of=foo bs=1 seek=5368709119
> 
> Plus modifying mtd to handle partitions greater than the u32 size limit.

I believe Falls Huang wanted to do just that and needed a test setup.
But I may be wrong.

Jörn

-- 
Linux [...] existed just for discussion between people who wanted
to show off how geeky they were.
-- Rob Enderle

Re: mtd support for large NAND

[Charles Manning]

On Wednesday 17 October 2007 08:55:22 Jörn Engel wrote:
> On Wed, 17 October 2007 08:47:46 +1300, Charles Manning wrote:
> > On Tuesday 16 October 2007 23:47:27 Jörn Engel wrote:
> > > On Fri, 12 October 2007 16:36:09 +0800, falls huang wrote:
> > > >   How can I simulate a large nandflash(>4G) in linux PC ? I want to
> > > > take a look at it, but I don't have any develop-board that has 4G
> > >
> > > Something like this would work:
> > >
> > > $ truncate 5GiB foo
> > > $ qemu -s -no-kqemu -hda rootfs -hdb foo -kernel arch/i386/boot/bzImage
> > > \ -append "ro root=/dev/hda block2mtd.block2mtd=/dev/hdb"
> > >
> > > Instead of truncate you can also use dd:
> > > $ dd if=/dev/zero of=foo bs=1 seek=5368709119
> >
> > Plus modifying mtd to handle partitions greater than the u32 size limit.
>
> I believe Falls Huang wanted to do just that and needed a test setup.
> But I may be wrong.

Does block2mtd simulate spare/oob areas too? Depending on what is being tested 
that might or might not be important.


-- Charles

Re: mtd support for large NAND

[Jörn Engel]

[ Someone removed me from Cc:.  Please don't. ]

On Wed, 17 October 2007 09:12:09 +1300, Charles Manning wrote:
> 
> Does block2mtd simulate spare/oob areas too? Depending on what is being tested 
> that might or might not be important.

It doesn't.  If you care enough, send a patch.

Jörn

-- 
The grand essentials of happiness are: something to do, something to
love, and something to hope for.
-- Allan K. Chalmers

Re: mtd support for large NAND

[Charles Manning]

On Wednesday 17 October 2007 09:42:11 Jörn Engel wrote:
> [ Someone removed me from Cc:.  Please don't. ]
>
> On Wed, 17 October 2007 09:12:09 +1300, Charles Manning wrote:
> > Does block2mtd simulate spare/oob areas too? Depending on what is being
> > tested that might or might not be important.
>
> It doesn't.  If you care enough, send a patch.

I don't use block2mtd so I don't really care for my own purposes. I just 
wanted to make this point clear for those that might be expecting an easy 
ride.

-- CHarles