All of lore.kernel.org
 help / color / mirror / Atom feed
* nandwrite and contiguous space
@ 2014-04-01  9:30 Stas Sergeev
  2014-04-01 18:23 ` Gerhard Sittig
  0 siblings, 1 reply; 6+ messages in thread
From: Stas Sergeev @ 2014-04-01  9:30 UTC (permalink / raw)
  To: linux-mtd

Hello.

I am looking for a NAND flasher that supports searching
for the contiguous region of good blocks (within a specified
range), and flashes an image to it. It seems nandwrite - the
default linux flasher - can only skip bad blocks, but the scenario
I need, doesn't seem to be supported.
Is there any other flasher that does support this?
Or maybe some trick with nandwrite exists?
If not, how difficult is it to implement such a capability into
nandwrite?

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: nandwrite and contiguous space
  2014-04-01  9:30 nandwrite and contiguous space Stas Sergeev
@ 2014-04-01 18:23 ` Gerhard Sittig
  2014-04-01 19:25   ` Stas Sergeev
  2014-04-02  4:19   ` Stas Sergeev
  0 siblings, 2 replies; 6+ messages in thread
From: Gerhard Sittig @ 2014-04-01 18:23 UTC (permalink / raw)
  To: Stas Sergeev; +Cc: linux-mtd

On Tue, 2014-04-01 at 13:30 +0400, Stas Sergeev wrote:
> 
> I am looking for a NAND flasher that supports searching
> for the contiguous region of good blocks (within a specified
> range), and flashes an image to it. It seems nandwrite - the
> default linux flasher - can only skip bad blocks, but the scenario
> I need, doesn't seem to be supported.

Isn't the point that the set of good blocks keeps changing over
time?  What might be good now, may be bad a second later or next
year, doesn't matter as long as it won't stay good forever.

> Is there any other flasher that does support this?
> Or maybe some trick with nandwrite exists?
> If not, how difficult is it to implement such a capability into
> nandwrite?

If you want a sequence of good blocks, this very much sounds like
the LEBs which UBI provides.  Aren't you asking for an UBI
volume?  What am I missing?


virtually yours
Gerhard Sittig
-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr. 5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80  Email: office@denx.de

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: nandwrite and contiguous space
  2014-04-01 18:23 ` Gerhard Sittig
@ 2014-04-01 19:25   ` Stas Sergeev
  2014-04-02  4:19   ` Stas Sergeev
  1 sibling, 0 replies; 6+ messages in thread
From: Stas Sergeev @ 2014-04-01 19:25 UTC (permalink / raw)
  To: Gerhard Sittig; +Cc: linux-mtd

01.04.2014 22:23, Gerhard Sittig пишет:
> On Tue, 2014-04-01 at 13:30 +0400, Stas Sergeev wrote:
>> I am looking for a NAND flasher that supports searching
>> for the contiguous region of good blocks (within a specified
>> range), and flashes an image to it. It seems nandwrite - the
>> default linux flasher - can only skip bad blocks, but the scenario
>> I need, doesn't seem to be supported.
> Isn't the point that the set of good blocks keeps changing over
> time?  What might be good now, may be bad a second later or next
> year, doesn't matter as long as it won't stay good forever.
Hi, I think the point is that if they are no longer ever written,
they won't change its goodness...
Anyway, this is how the ST soc chips work.
They search NAND for a specific signature, and once
found, they map the contiguous block, starting from
that signature, to memory, which allows to run u-boot
in a nor-alike fashion. They do not do BBT lookups when
mapping, AFAIK, and so such a requirement for flasher
does exist.

>> Is there any other flasher that does support this?
>> Or maybe some trick with nandwrite exists?
>> If not, how difficult is it to implement such a capability into
>> nandwrite?
> If you want a sequence of good blocks, this very much sounds like
> the LEBs which UBI provides.  Aren't you asking for an UBI
> volume?  What am I missing?
Unfortunately, I don't think the aforementioned ST chips
can boot from UBI. Once booted, I want UBI for sure.
But the only alternative to the aforementioned requirement
would be to assume that 2 first blocks are good, and fit
u-boot there, and it they are bad - throw away such a nand.
I am considering this too, but I wonder how many nands
will be wasted this way... Likely none, but who knows. :)

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: nandwrite and contiguous space
  2014-04-01 18:23 ` Gerhard Sittig
  2014-04-01 19:25   ` Stas Sergeev
@ 2014-04-02  4:19   ` Stas Sergeev
  2014-04-04 22:00     ` Richard Weinberger
  1 sibling, 1 reply; 6+ messages in thread
From: Stas Sergeev @ 2014-04-02  4:19 UTC (permalink / raw)
  To: Gerhard Sittig; +Cc: linux-mtd

01.04.2014 22:23, Gerhard Sittig пишет:
> If you want a sequence of good blocks, this very much sounds like
> the LEBs which UBI provides.  Aren't you asking for an UBI
> volume?  What am I missing?
Is is possible to configure ubi leb size to be multiple
physically contiguous eraseblocks? With some googling
it seems not, or what am I missing?

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: nandwrite and contiguous space
  2014-04-02  4:19   ` Stas Sergeev
@ 2014-04-04 22:00     ` Richard Weinberger
  2014-04-12  8:30       ` Stas Sergeev
  0 siblings, 1 reply; 6+ messages in thread
From: Richard Weinberger @ 2014-04-04 22:00 UTC (permalink / raw)
  To: Stas Sergeev; +Cc: Gerhard Sittig, linux-mtd

On Wed, Apr 2, 2014 at 6:19 AM, Stas Sergeev <stsp@list.ru> wrote:
> 01.04.2014 22:23, Gerhard Sittig пишет:
>>
>> If you want a sequence of good blocks, this very much sounds like
>>
>> the LEBs which UBI provides.  Aren't you asking for an UBI
>> volume?  What am I missing?
>
> Is is possible to configure ubi leb size to be multiple
> physically contiguous eraseblocks? With some googling
> it seems not, or what am I missing?

No, this is not possible.

-- 
Thanks,
//richard

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: nandwrite and contiguous space
  2014-04-04 22:00     ` Richard Weinberger
@ 2014-04-12  8:30       ` Stas Sergeev
  0 siblings, 0 replies; 6+ messages in thread
From: Stas Sergeev @ 2014-04-12  8:30 UTC (permalink / raw)
  To: Richard Weinberger; +Cc: Gerhard Sittig, linux-mtd

[-- Attachment #1: Type: text/plain, Size: 796 bytes --]

05.04.2014 02:00, Richard Weinberger пишет:
> On Wed, Apr 2, 2014 at 6:19 AM, Stas Sergeev <stsp@list.ru> wrote:
>> 01.04.2014 22:23, Gerhard Sittig пишет:
>>> If you want a sequence of good blocks, this very much sounds like
>>>
>>> the LEBs which UBI provides.  Aren't you asking for an UBI
>>> volume?  What am I missing?
>> Is is possible to configure ubi leb size to be multiple
>> physically contiguous eraseblocks? With some googling
>> it seems not, or what am I missing?
> No, this is not possible.
OK so I patched nandwrite.
The patch is attached only for the preliminary review or
if someone wants to help me with testing this.
The problem is that I don't know how to even test it.
What tool should I use to mark blocks as bad, test
the patch, and then unmark the blocks back to good?

[-- Attachment #2: nandpatch.tar.gz --]
[-- Type: application/gzip, Size: 4735 bytes --]

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2014-04-12  8:30 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-04-01  9:30 nandwrite and contiguous space Stas Sergeev
2014-04-01 18:23 ` Gerhard Sittig
2014-04-01 19:25   ` Stas Sergeev
2014-04-02  4:19   ` Stas Sergeev
2014-04-04 22:00     ` Richard Weinberger
2014-04-12  8:30       ` Stas Sergeev

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.