public inbox for linux-mtd@lists.infradead.org
 help / color / mirror / Atom feed
* IF eraseall will destryed(phys) the nand flash ?
@ 2002-12-03 21:42 shadow
  2002-12-04 10:01 ` Thomas Gleixner
  0 siblings, 1 reply; 7+ messages in thread
From: shadow @ 2002-12-03 21:42 UTC (permalink / raw)
  To: linux-mtd

Hi all,
I got an Input/Output error when I erase the nand flash. It printed these 
messages.
-------------------------
Erasinand_erase: attempt to erase a bad block at page 0x00000020
ng 8 Kibyte @ 4000 --  0 % complete.
eraseall: /dev/mtd/6: Mnand_erase: attempt to erase a bad block at page 
0x00000030
TD Erase failure: Input/output error
Erasing 8 Kibyte @ 6000 --  1 % complete.
eraseall: /dev/mtd/6: Mnand_erase: attempt to erase a bad block at page 
0x00000040
TD Erase failure: Input/output error
Erasing 8 Kibyte @ 8000 --  1 % complete.
eraseall: /dev/mtd/6: MTD Erase failure: Input/output error
Erased 2048 Kibyte @ 0 -- 100% complete. 
-----------------------------

I found the Input/output error always occurred at the same page even if it 
is a new SMC. But I have a question, the block_size is larger than 
page_size, isn't it ? IF larger why nand driver attempt to erase a bad 
block at page 0x00000020. In my mind, I think one block includes many 
      ^^
pages. Another problem is why the new SMC has some bad blocks and occurred 
at the same page with using eraseall program erase it. Then I put this SMC 
that has bad blocks into Digital Camera that can erase data on the SMC. 
The DC Can't format it, what's up ? eraseall destroyed it ?

IF the SMC has some bad blocks, then how to correct it ?

thanks in advance,

Brian

      

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: IF eraseall will destryed(phys) the nand flash ?
  2002-12-04 10:01 ` Thomas Gleixner
@ 2002-12-03 22:46   ` shadow
  2002-12-04 10:35     ` Thomas Gleixner
  2002-12-04 10:51     ` Henrik Nordstrom
  0 siblings, 2 replies; 7+ messages in thread
From: shadow @ 2002-12-03 22:46 UTC (permalink / raw)
  To: Thomas Gleixner; +Cc: linux-mtd

On Wed, 4 Dec 2002, Thomas Gleixner wrote:

> > Another problem is why the new SMC has some bad blocks and occurred
> > at the same page with using eraseall program erase it. 
> NAND FLASH can have factory marked bad blocks at shipping time. See datasheet.
> Bad block is bad block and does not go away by erasing it. In fact you _MUST 
> NOT_ erase it, as bad block information is lost and 
> 
	I don't understand what is shipping time, your meaming is it was 
destroyed at the shipping time between the factory to retailer ?

Thank your response,

Brian
 

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: IF eraseall will destryed(phys) the nand flash ?
  2002-12-03 21:42 IF eraseall will destryed(phys) the nand flash ? shadow
@ 2002-12-04 10:01 ` Thomas Gleixner
  2002-12-03 22:46   ` shadow
  0 siblings, 1 reply; 7+ messages in thread
From: Thomas Gleixner @ 2002-12-04 10:01 UTC (permalink / raw)
  To: shadow, linux-mtd

On Tuesday 03 December 2002 22:42, shadow wrote:
> Hi all,
> I got an Input/Output error when I erase the nand flash. It printed these
> messages.
> -------------------------
> Erasinand_erase: attempt to erase a bad block at page 0x00000020
> ng 8 Kibyte @ 4000 --  0 % complete.
> eraseall: /dev/mtd/6: Mnand_erase: attempt to erase a bad block at page
> 0x00000030
> TD Erase failure: Input/output error
> Erasing 8 Kibyte @ 6000 --  1 % complete.
> eraseall: /dev/mtd/6: Mnand_erase: attempt to erase a bad block at page
> 0x00000040
> TD Erase failure: Input/output error
> Erasing 8 Kibyte @ 8000 --  1 % complete.
> eraseall: /dev/mtd/6: MTD Erase failure: Input/output error
> Erased 2048 Kibyte @ 0 -- 100% complete.
> -----------------------------
>
> I found the Input/output error always occurred at the same page even if it
> is a new SMC. But I have a question, the block_size is larger than
> page_size, isn't it ? IF larger why nand driver attempt to erase a bad
> block at page 0x00000020. In my mind, I think one block includes many
>       ^^
> pages. 
512Bytes (Pagesize) * 16 (0x10) pages = 8192 Byte = Blocksize

NAND driver refuses to erase a bad blocks.
The attempt comes from eraseall, which does not know that the block is bad. 
And there is no need to know it, as the driver takes care.

So you have 3 bad blocks:
Block Nr. 2 at page 0x20 (32)
Block Nr. 3 at page 0x30 (48)
Block Nr. 4 at page 0x40 (64)

> Another problem is why the new SMC has some bad blocks and occurred
> at the same page with using eraseall program erase it. 
NAND FLASH can have factory marked bad blocks at shipping time. See datasheet.
Bad block is bad block and does not go away by erasing it. In fact you _MUST 
NOT_ erase it, as bad block information is lost and 

> Then I put this SMC  that has bad blocks into Digital Camera that can erase 
> data on the SMC. The DC Can't format it, what's up ? eraseall destroyed it 
Nope eraseall did not destroy anything. It just erased the card.
Your DC can probably use only preformatted SMC. The formatting is SmartMedia 
FAT Filesystem, which is not supported by MTD. By erasing the SMC you deleted 
the SMC FAT preformat.

> IF the SMC has some bad blocks, then how to correct it ?
Bad block is a bad block and will be a bad block all the time. NAND aware 
filesystems handle bad blocks correctly. Bad blocks just reduce the available 
storage size of your chip. Your 2MB partition has now 2MB - 3*8kB storage 
size. Don't worry about. Mount JFFS2 and see what happens.

-- 
Thomas
____________________________________________________
linutronix - competence in embedded & realtime linux
http://www.linutronix.de
mail: tglx@linutronix.de

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: IF eraseall will destryed(phys) the nand flash ?
  2002-12-03 22:46   ` shadow
@ 2002-12-04 10:35     ` Thomas Gleixner
  2002-12-04 13:00       ` shadow
  2002-12-04 10:51     ` Henrik Nordstrom
  1 sibling, 1 reply; 7+ messages in thread
From: Thomas Gleixner @ 2002-12-04 10:35 UTC (permalink / raw)
  To: shadow; +Cc: linux-mtd

On Tuesday 03 December 2002 23:46, shadow wrote:
> On Wed, 4 Dec 2002, Thomas Gleixner wrote:
> > > Another problem is why the new SMC has some bad blocks and occurred
> > > at the same page with using eraseall program erase it.
> >
> > NAND FLASH can have factory marked bad blocks at shipping time. See
> > datasheet. Bad block is bad block and does not go away by erasing it. In
> > fact you _MUST NOT_ erase it, as bad block information is lost and
>
> 	I don't understand what is shipping time, your meaming is it was
> destroyed at the shipping time between the factory to retailer ?
Nope, the blocks were bad, when the chip was produced or do you believe that 
the postman is a heavy source of gamma rays ? :)

-- 
Thomas
____________________________________________________
linutronix - competence in embedded & realtime linux
http://www.linutronix.de
mail: tglx@linutronix.de

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: IF eraseall will destryed(phys) the nand flash ?
  2002-12-03 22:46   ` shadow
  2002-12-04 10:35     ` Thomas Gleixner
@ 2002-12-04 10:51     ` Henrik Nordstrom
  2002-12-04 13:03       ` shadow
  1 sibling, 1 reply; 7+ messages in thread
From: Henrik Nordstrom @ 2002-12-04 10:51 UTC (permalink / raw)
  To: shadow; +Cc: linux-mtd

tis 2002-12-03 klockan 23.46 skrev shadow:

> I don't understand what is shipping time, your meaming is it was 
> destroyed at the shipping time between the factory to retailer ?

The errors was detected at the factory and marked as bad blocks before
the flash was shipped to the retailer.

Just as there is small glitches in almost all TFT screens you can find
there is quite commonly a small amount of errors in the manufacturement
of flash memories.. 

Regards
Henrik

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: IF eraseall will destryed(phys) the nand flash ?
  2002-12-04 10:35     ` Thomas Gleixner
@ 2002-12-04 13:00       ` shadow
  0 siblings, 0 replies; 7+ messages in thread
From: shadow @ 2002-12-04 13:00 UTC (permalink / raw)
  To: Thomas Gleixner; +Cc: linux-mtd

On Wed, 4 Dec 2002, Thomas Gleixner wrote:

> On Tuesday 03 December 2002 23:46, shadow wrote:
> > On Wed, 4 Dec 2002, Thomas Gleixner wrote:
> > > > Another problem is why the new SMC has some bad blocks and occurred
> > > > at the same page with using eraseall program erase it.
> > >
> > > NAND FLASH can have factory marked bad blocks at shipping time. See
> > > datasheet. Bad block is bad block and does not go away by erasing it. In
> > > fact you _MUST NOT_ erase it, as bad block information is lost and
> >
> > 	I don't understand what is shipping time, your meaming is it was
> > destroyed at the shipping time between the factory to retailer ?
> Nope, the blocks were bad, when the chip was produced or do you believe that 
> the postman is a heavy source of gamma rays ? :)
> 

Thank your response,

Brian

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: IF eraseall will destryed(phys) the nand flash ?
  2002-12-04 10:51     ` Henrik Nordstrom
@ 2002-12-04 13:03       ` shadow
  0 siblings, 0 replies; 7+ messages in thread
From: shadow @ 2002-12-04 13:03 UTC (permalink / raw)
  To: Henrik Nordstrom; +Cc: linux-mtd

On 4 Dec 2002, Henrik Nordstrom wrote:

> tis 2002-12-03 klockan 23.46 skrev shadow:
> 
> > I don't understand what is shipping time, your meaming is it was 
> > destroyed at the shipping time between the factory to retailer ?
> 
> The errors was detected at the factory and marked as bad blocks before
> the flash was shipped to the retailer.
> 
> Just as there is small glitches in almost all TFT screens you can find
> there is quite commonly a small amount of errors in the manufacturement
> of flash memories.. 
> 
> Regards
> Henrik
> 
Thanks in advance,

Brian

^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2002-12-05  0:19 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2002-12-03 21:42 IF eraseall will destryed(phys) the nand flash ? shadow
2002-12-04 10:01 ` Thomas Gleixner
2002-12-03 22:46   ` shadow
2002-12-04 10:35     ` Thomas Gleixner
2002-12-04 13:00       ` shadow
2002-12-04 10:51     ` Henrik Nordstrom
2002-12-04 13:03       ` shadow

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox