MTD utils v1.5.0: MEMERASE64 ioctl failed error

MTD utils v1.5.0: MEMERASE64 ioctl failed error

[Shah, Minal]

Resending this email as it got rejected, being in HTML format.

Hi Mike,
I am using MTD utils version 1.5.0 for DRA7 (Vayu) platform.

I have bad blocks in some of my NAND partitions which I am not able to erase from linux kernel.
Using "-N" option in flash_eraseall ideally should have allowed erasing the bad block but I land up in "MEMERASE64 ioctl failed" error. 
The reason for this error seems that "OFFS64_IOCTLS_SUPPORTED" is not set.
I haven't yet digged in detail to see why this ioctl fails but can you please tell me the possible reasons for the same? 

Please see the logs below for reference.

root@dra7xx-evm:~# flash_eraseall /dev/mtd0
flash_eraseall has been replaced by `flash_erase <mtddev> 0 0`; please use it
flash_erase: Skipping bad block at 00000000
Erasing 128 Kibyte @ 0 -- 100 % complete

root@dra7xx-evm:~# flash_eraseall -N /dev/mtd0
flash_eraseall has been replaced by `flash_erase <mtddev> 0 0`; please use it
Erasing 128 Kibyte @ 0 --  0 % complete [ 2298.842498] nand_erase_nand: attempt to erase a bad block at page 0x00000000
libmtd: error!: MEMERASE64 ioctl failed for eraseblock 0 (mtd0)
        error 5 (Input/output error)
flash_erase: error!: /dev/mtd0: MTD Erase failure
             error 5 (Input/output error)
Erasing 128 Kibyte @ 0 -- 100 % complete

root@dra7xx-evm:~# flash_eraseall -N -u /dev/mtd0
flash_eraseall has been replaced by `flash_erase <mtddev> 0 0`; please use it
Erasing 128 Kibyte @ 0 --  0 % complete libmtd: error!: MEMUNLOCK ioctl failed for eraseblock 0 (mtd0)
        error 95 (Operation not supported)
flash_erase: error!: /dev/mtd0: MTD unlock failure
             error 95 (Operation not supported)
Erasing 128 Kibyte @ 0 -- 100 % complete

root@dra7xx-evm:~# mtd_debug erase /dev/mtd0 0 131072
[ 1932.902496] nand_erase_nand: attempt to erase a bad block at page 0x00000000
MEMERASE: Input/output error


root@dra7xx-evm:~# nanddump -n -o -p -s 0x0 -l 0x800 --bb=dumpbad /dev/mtd0
Block size 131072, page size 2048, OOB size 64
Dumping data starting at 0x00000000 and ending at 0x00000800...
0x00000000: 52 ff 00 ff 60 ff b3 ff 63 ff a9 ff 55 ff a6 ff
0x00000010: ba ff c3 ff 38 ff b8 ff b0 ff a5 ff 04 ff 3e ff
0x00000020: 51 ff a3 ff 81 ff 53 ff 07 ff 8f ff 75 ff 01 ff
0x00000030: c8 ff da ff 6b ff 6d ff d5 ff 70 ff df ff 27 ff
0x00000040: 70 ff 40 ff da ff d3 ff e9 ff 2f ff 79 ff a3 ff
0x00000050: f3 ff b1 ff 5b ff a3 ff 57 ff 60 ff e1 ff a8 ff
0x00000060: 03 ff 62 ff fb ff 0b ff f2 ff 70 ff 0c ff ba ff
0x00000070: 4b ff 77 ff 28 ff 20 ff e8 ff 07 ff 47 ff 58 ff
0x00000080: 47 ff 21 ff 2c ff 30 ff 51 ff a5 ff d3 ff 44 ff
0x00000090: 57 ff 2f ff e7 ff ae ff 8f ff c8 ff 56 ff 92 ff
0x000000a0: 2a ff 52 ff 9d ff 1c ff c2 ff aa ff d7 ff 0d ff
0x000000b0: 21 ff ff ff 2d ff 09 ff 06 ff 74 ff 62 ff 4e ff
0x000000c0: 96 ff 8e ff 7e ff e7 ff 33 ff 52 ff 2b ff 8a ff
0x000000d0: 81 ff 12 ff 38 ff 10 ff da ff 8f ff a2 ff 04 ff
0x000000e0: e1 ff 40 ff 21 ff a3 ff ea ff f8 ff b1 ff 0b ff
0x000000f0: f7 ff de ff 15 ff fd ff 53 ff 77 ff 4b ff e9 ff
0x00000100: 05 ff ca ff d0 ff 38 ff 1c ff fb ff c3 ff 9d ff
0x00000110: 0d ff fb ff ad ff e7 ff 8a ff 4f ff eb ff 6b ff
0x00000120: 8f ff 0c ff 0f ff 79 ff 04 ff c0 ff 85 ff fb ff
0x00000130: 9e ff 9a ff f9 ff f1 ff 11 ff 44 ff da ff 16 ff
0x00000140: 0e ff aa ff 4e ff 2a ff a5 ff 11 ff c7 ff b2 ff
0x00000150: 0d ff 74 ff 99 ff 97 ff c4 ff 85 ff 03 ff 53 ff
0x00000160: 91 ff 12 ff cd ff 96 ff d2 ff 52 ff 91 ff 70 ff
0x00000170: ec ff 8a ff 62 ff fd ff cf ff 3c ff 13 ff dd ff
0x00000180: e7 ff 61 ff 08 ff 8c ff 73 ff cf ff 3f ff 80 ff
0x00000190: 44 ff d8 ff 17 ff 08 ff 5d ff 1a ff 5b ff ef ff
0x000001a0: 2c ff 28 ff 85 ff fe ff 7a ff 16 ff 6f ff 66 ff
0x000001b0: a1 ff d1 ff 63 ff 70 ff 0d ff 76 ff 4d ff f4 ff
0x000001c0: d8 ff 55 ff 81 ff 4b ff 25 ff c0 ff cb ff 69 ff
0x000001d0: 98 ff e2 ff 71 ff f6 ff fd ff cc ff e5 ff 29 ff
0x000001e0: f5 ff 6a ff 28 ff 6f ff 80 ff 97 ff d6 ff 21 ff
0x000001f0: 68 ff 39 ff 91 ff 75 ff b0 ff df ff 6a ff 88 ff
0x00000200: 34 ff eb ff d3 ff 59 ff ab ff 9e ff c2 ff 43 ff
0x00000210: 80 ff 33 ff 39 ff 7d ff 00 ff 1e ff a7 ff f5 ff
0x00000220: 88 ff cf ff 64 ff 09 ff 66 ff 3a ff 2a ff ce ff
0x00000230: 74 ff bc ff 43 ff 24 ff 9b ff ad ff ac ff cf ff
0x00000240: 98 ff 7f ff 29 ff 43 ff 1d ff eb ff 87 ff 9d ff
0x00000250: 1f ff c0 ff 1b ff 1f ff df ff c2 ff 14 ff 67 ff
0x00000260: 91 ff 78 ff 70 ff f7 ff b3 ff 9b ff c5 ff 27 ff
0x00000270: 57 ff 08 ff 4b ff f2 ff b6 ff f7 ff c1 ff 4e ff
0x00000280: 76 ff ea ff 92 ff 93 ff d6 ff 19 ff 30 ff f5 ff
0x00000290: d9 ff 4b ff 14 ff b8 ff 0d ff 28 ff 20 ff 9e ff
0x000002a0: a0 ff 90 ff 95 ff 53 ff 2b ff 5a ff 7a ff 82 ff
0x000002b0: 63 ff c5 ff 74 ff 19 ff bc ff 36 ff 67 ff 32 ff
0x000002c0: 20 ff f9 ff c5 ff f6 ff 12 ff f6 ff eb ff ec ff
0x000002d0: 41 ff ff ff a4 ff 4f ff 27 ff c4 ff ed ff c8 ff
0x000002e0: 55 ff 83 ff 1b ff 80 ff dd ff 96 ff 03 ff 40 ff
0x000002f0: 5b ff 77 ff 59 ff 18 ff ad ff c1 ff 4a ff ce ff
0x00000300: ba ff 10 ff c4 ff cd ff 06 ff b0 ff b9 ff 47 ff
0x00000310: af ff 5d ff 96 ff d7 ff 22 ff 84 ff 9f ff 77 ff
0x00000320: 07 ff ba ff f7 ff e4 ff 50 ff fa ff 25 ff ac ff
0x00000330: 72 ff 7e ff c4 ff 1f ff 3f ff 0e ff ed ff fa ff
0x00000340: 1e ff b2 ff c7 ff 24 ff 62 ff 80 ff 6c ff 11 ff
0x00000350: dd ff 02 ff e8 ff ff ff 86 ff 87 ff 76 ff 8d ff
0x00000360: 42 ff 6e ff 72 ff 92 ff 68 ff 97 ff 3e ff da ff
0x00000370: 15 ff 02 ff fa ff 55 ff 11 ff e7 ff 4f ff 2f ff
0x00000380: 99 ff 16 ff 54 ff fb ff 96 ff c0 ff 0d ff 73 ff
0x00000390: c2 ff f5 ff 73 ff 49 ff 7d ff e9 ff d6 ff bf ff
0x000003a0: 57 ff 48 ff 51 ff c0 ff df ff 90 ff 9a ff f5 ff
0x000003b0: 92 ff 94 ff 4a ff a3 ff 7c ff 99 ff d3 ff 15 ff
0x000003c0: af ff 27 ff 11 ff 45 ff e7 ff 1e ff b8 ff a9 ff
0x000003d0: 13 ff 2b ff f2 ff 90 ff 15 ff c9 ff 4f ff 6c ff
0x000003e0: 11 ff a1 ff 2c ff f1 ff 31 ff c7 ff e6 ff c3 ff
0x000003f0: 5b ff 30 ff 67 ff d7 ff c9 ff 3a ff ed ff 78 ff
0x00000400: 61 ff fe ff bd ff 48 ff 1c ff 75 ff f1 ff 2f ff
0x00000410: a1 ff e4 ff c0 ff b6 ff ad ff 0f ff 22 ff be ff
0x00000420: b0 ff 4f ff af ff e1 ff 16 ff 95 ff a5 ff 71 ff
0x00000430: c5 ff 0c ff 49 ff 8e ff 46 ff 36 ff 06 ff a7 ff
0x00000440: 34 ff c3 ff ef ff 50 ff 39 ff e0 ff 7f ff da ff
0x00000450: c4 ff 3f ff 90 ff 71 ff 4f ff b2 ff 30 ff ff ff
0x00000460: 01 ff df ff e1 ff 17 ff 75 ff 86 ff 89 ff 3a ff
0x00000470: 92 ff d2 ff c9 ff d8 ff 08 ff cf ff 7f ff 3c ff
0x00000480: 93 ff 6e ff 8c ff cc ff 4e ff 0b ff a6 ff 13 ff
0x00000490: 4b ff 36 ff 84 ff 9a ff e8 ff b4 ff 99 ff ea ff
0x000004a0: 94 ff 7a ff 01 ff 09 ff 00 ff 8a ff 43 ff 92 ff
0x000004b0: 5c ff 0c ff 6a ff 64 ff dc ff e9 ff a0 ff 6f ff
0x000004c0: 57 ff 2c ff 3b ff a6 ff 38 ff e1 ff b9 ff 83 ff
0x000004d0: 17 ff 3d ff 1d ff ff ff f2 ff b6 ff e9 ff 86 ff
0x000004e0: 31 ff eb ff 8f ff 31 ff 75 ff d2 ff c4 ff d2 ff
0x000004f0: df ff 2e ff 36 ff bb ff 18 ff d7 ff 2a ff 6f ff
0x00000500: 03 ff 65 ff 15 ff 3b ff 46 ff ce ff be ff 5d ff
0x00000510: 0c ff db ff 5c ff fe ff 92 ff 46 ff 84 ff c3 ff
0x00000520: 31 ff 13 ff f4 ff a6 ff e5 ff b8 ff 78 ff c4 ff
0x00000530: e7 ff af ff 7f ff ff ff 86 ff a9 ff 6e ff 89 ff
0x00000540: 0e ff 84 ff c5 ff 54 ff 52 ff 83 ff b1 ff 5e ff
0x00000550: 5f ff 0e ff 5c ff f1 ff 54 ff e0 ff b4 ff 85 ff
0x00000560: f3 ff a8 ff 2b ff d9 ff 61 ff a4 ff 9d ff 48 ff
0x00000570: 53 ff 1d ff 47 ff d9 ff c6 ff b5 ff 62 ff d5 ff
0x00000580: 39 ff 27 ff 29 ff 8c ff ab ff db ff ea ff 0a ff
0x00000590: e9 ff 47 ff fb ff 3d ff 27 ff af ff c2 ff 1b ff
0x000005a0: 57 ff ed ff f4 ff b8 ff 91 ff 91 ff 00 ff e4 ff
0x000005b0: ae ff 47 ff bd ff 75 ff fd ff 20 ff 4a ff 36 ff
0x000005c0: 47 ff 73 ff c2 ff f2 ff 4e ff ad ff fc ff 37 ff
0x000005d0: f4 ff f7 ff 74 ff 1b ff a6 ff 36 ff 36 ff fe ff
0x000005e0: 24 ff 2a ff b6 ff b5 ff bc ff b7 ff 9a ff 6a ff
0x000005f0: fe ff 57 ff df ff fb ff 77 ff 29 ff 32 ff bf ff
0x00000600: 9d ff f4 ff b1 ff eb ff a1 ff ae ff 23 ff 95 ff
0x00000610: a5 ff 97 ff b1 ff 4c ff ce ff e7 ff 4a ff f2 ff
0x00000620: 12 ff 00 ff a7 ff ce ff b7 ff 41 ff 38 ff b6 ff
0x00000630: 99 ff 18 ff b1 ff 10 ff 41 ff e3 ff cf ff de ff
0x00000640: d8 ff 81 ff ca ff 79 ff 2f ff ed ff 0f ff d4 ff
0x00000650: 84 ff c0 ff 20 ff 52 ff a7 ff 6a ff 44 ff b9 ff
0x00000660: 6b ff ec ff 87 ff 22 ff 2d ff c0 ff d8 ff c6 ff
0x00000670: d8 ff 8a ff d7 ff 19 ff 6d ff a6 ff f8 ff 45 ff
0x00000680: 27 ff c2 ff bf ff 56 ff af ff ce ff 2b ff 33 ff
0x00000690: 8e ff 4b ff 86 ff 35 ff b6 ff ca ff ef ff 21 ff
0x000006a0: b6 ff 76 ff 43 ff e4 ff 36 ff 1c ff aa ff 0e ff
0x000006b0: a6 ff 81 ff 28 ff 13 ff 28 ff 20 ff 59 ff 4f ff
0x000006c0: e2 ff 18 ff a6 ff 91 ff e6 ff d1 ff c4 ff 74 ff
0x000006d0: 1c ff 4a ff a9 ff d2 ff 15 ff 98 ff f3 ff cb ff
0x000006e0: 0f ff 37 ff af ff 45 ff 53 ff 5a ff 54 ff f9 ff
0x000006f0: db ff 7c ff 0c ff 03 ff 9c ff 65 ff 53 ff 7e ff
0x00000700: 7d ff f9 ff 0f ff 63 ff ca ff d3 ff d7 ff e6 ff
0x00000710: 1e ff 81 ff b9 ff 33 ff 19 ff ac ff fe ff 28 ff
0x00000720: e3 ff ae ff 6e ff 36 ff 08 ff c2 ff 2f ff e3 ff
0x00000730: 3e ff 3c ff e7 ff da ff a1 ff 3a ff 58 ff 1f ff
0x00000740: 33 ff 67 ff 82 ff fd ff 3a ff 5a ff e3 ff 58 ff
0x00000750: db ff 9c ff 8b ff f4 ff 49 ff 8a ff 1d ff 2c ff
0x00000760: 38 ff 8b ff 63 ff 40 ff 4d ff 92 ff 23 ff 8b ff
0x00000770: ce ff 0a ff 65 ff 70 ff 44 ff bd ff 8f ff 77 ff
0x00000780: 24 ff 11 ff 74 ff 5e ff 6b ff 58 ff b7 ff 46 ff
0x00000790: f4 ff 42 ff 3b ff 3d ff cc ff 58 ff 6a ff 04 ff
0x000007a0: e3 ff cd ff 44 ff 30 ff 5f ff 68 ff bb ff 2e ff
0x000007b0: 72 ff 20 ff 9e ff b7 ff dd ff 2d ff 2e ff 01 ff
0x000007c0: 3e ff a3 ff 5f ff aa ff fb ff 16 ff f0 ff ef ff
0x000007d0: 59 ff 2b ff 2d ff 25 ff 83 ff 97 ff 2a ff 66 ff
0x000007e0: 64 ff 6e ff 96 ff c3 ff d6 ff 51 ff f1 ff 49 ff
0x000007f0: 71 ff 8f ff 00 ff 4e ff bc ff 2e ff 4f ff fb ff
  OOB Data: d1 ff af ff a5 ff cc ff c5 ff 95 ff bc ff 1e ff
  OOB Data: c1 ff e9 ff 44 ff 44 ff 80 ff 6e ff ab ff e4 ff
  OOB Data: dc ff 41 ff a7 ff b3 ff 93 ff 99 ff fc ff 04 ff
  OOB Data: 28 ff fc ff 53 ff e5 ff 2a ff a2 ff e0 ff fc ff



Regards,
Minal
+91-80-2519-3129

Re: MTD utils v1.5.0: MEMERASE64 ioctl failed error

[Brian Norris]

Hi Minal,

On Wed, Oct 09, 2013 at 05:05:11AM +0000, Shah, Minal wrote:
> I am using MTD utils version 1.5.0 for DRA7 (Vayu) platform.
> 
> I have bad blocks in some of my NAND partitions which I am not able to erase from linux kernel.

MTD does not allow erasing bad blocks. This is prohibited by NAND
datasheets. There are occasions where, for debugging purposes, one might
need to erase a "bad" block that is not actually bad, but MTD does not
provide an interface for doing this. An interface for doing so was
discussed a while back, but nothing was merged.

> Using "-N" option in flash_eraseall ideally should have allowed erasing the bad block but I land up in "MEMERASE64 ioctl failed" error. 

"-N" just means that the user-space tool (flash_erase) won't try to skip
bad blocks. But that doesn't circumvent the in-kernel protection that
prevents the erase from succeeding. So the "Input/output" error you see
is completely expected and currently unavoidable, unless you modify your
kernel to remove this protection.

[snip]

So why do you want to erase bad blocks?

Brian

Re: MTD utils v1.5.0: MEMERASE64 ioctl failed error

[Ezequiel Garcia]

On Fri, Oct 11, 2013 at 03:24:59PM -0700, Brian Norris wrote:
> Hi Minal,
> 
> On Wed, Oct 09, 2013 at 05:05:11AM +0000, Shah, Minal wrote:
> > I am using MTD utils version 1.5.0 for DRA7 (Vayu) platform.
> > 
> > I have bad blocks in some of my NAND partitions which I am not able to erase from linux kernel.
> 
> MTD does not allow erasing bad blocks. This is prohibited by NAND
> datasheets. There are occasions where, for debugging purposes, one might
> need to erase a "bad" block that is not actually bad, but MTD does not
> provide an interface for doing this. An interface for doing so was
> discussed a while back, but nothing was merged.
> 

Ah, interesting. As you know I'm doing some very intrusive debugging
while developing the Armada driver and I usually need to circumvent
(manually) the kernel bad block erase check. Sometimes my blocks are not
really bad, yet they are detected as bad.

I'd be interesting in having some way to remove the check easily, but
since this is of no benefit, unless you're developing a crazy driver
I'm not sure how to do this.

What do you think?

Maybe some hidden /sys/kernel/debug/mtd file to turn that option, just
for the use of a developer?
-- 
Ezequiel García, Free Electrons
Embedded Linux, Kernel and Android Engineering
http://free-electrons.com

RE: MTD utils v1.5.0: MEMERASE64 ioctl failed error

[Shah, Minal]

Hi Brian,
Thanks a lot for the explanation; I understood the reason for failure.

> So why do you want to erase bad blocks?
The reason why I want to erase bad blocks is because they actually don't seem to be bad. 
Block0 itself is shown as bad.
I have few partitions (like MLO and u-boot env) of size equal to 1 erase block. The block is marked as bad and hence I cannot write anything to this partition. 
One of the solutions is to increase that partition size to 2 blocks so even if 1 block is bad then it will atleast be able to write to the other block of the same partition. But I don't think I should do this while the block is not really bad (what if all the blocks for some reason are marked as bad while they are actually not).

This question came in my mind from the fact that "nand scrub" command is allowed from u-boot which allows to also completely erase all the bad blocks; so why not from the kernel level.

Regards,
Minal

-----Original Message-----
From: Brian Norris [mailto:computersforpeace@gmail.com] 
Sent: Saturday, October 12, 2013 3:55 AM
To: Shah, Minal
Cc: Mike Frysinger; Mukherjee, Somnath; linux-mtd@lists.infradead.org; Gupta, Pekon; Parikh, Urmil
Subject: Re: MTD utils v1.5.0: MEMERASE64 ioctl failed error

Hi Minal,

On Wed, Oct 09, 2013 at 05:05:11AM +0000, Shah, Minal wrote:
> I am using MTD utils version 1.5.0 for DRA7 (Vayu) platform.
> 
> I have bad blocks in some of my NAND partitions which I am not able to erase from linux kernel.

MTD does not allow erasing bad blocks. This is prohibited by NAND
datasheets. There are occasions where, for debugging purposes, one might
need to erase a "bad" block that is not actually bad, but MTD does not
provide an interface for doing this. An interface for doing so was
discussed a while back, but nothing was merged.

> Using "-N" option in flash_eraseall ideally should have allowed erasing the bad block but I land up in "MEMERASE64 ioctl failed" error. 

"-N" just means that the user-space tool (flash_erase) won't try to skip
bad blocks. But that doesn't circumvent the in-kernel protection that
prevents the erase from succeeding. So the "Input/output" error you see
is completely expected and currently unavoidable, unless you modify your
kernel to remove this protection.

[snip]

So why do you want to erase bad blocks?

Brian

Re: MTD utils v1.5.0: MEMERASE64 ioctl failed error

[Mike Dunn]

On 10/15/2013 01:32 AM, Shah, Minal wrote:
> Hi Brian,
> Thanks a lot for the explanation; I understood the reason for failure.
> 
>> So why do you want to erase bad blocks?
> The reason why I want to erase bad blocks is because they actually don't seem to be bad. 
> Block0 itself is shown as bad.
> I have few partitions (like MLO and u-boot env) of size equal to 1 erase block. The block is marked as bad and hence I cannot write anything to this partition. 
> One of the solutions is to increase that partition size to 2 blocks so even if 1 block is bad then it will atleast be able to write to the other block of the same partition. But I don't think I should do this while the block is not really bad (what if all the blocks for some reason are marked as bad while they are actually not).
> 
> This question came in my mind from the fact that "nand scrub" command is allowed from u-boot which allows to also completely erase all the bad blocks; so why not from the kernel level.
> 
> Regards,
> Minal
> 
> -----Original Message-----
> From: Brian Norris [mailto:computersforpeace@gmail.com] 
> Sent: Saturday, October 12, 2013 3:55 AM
> To: Shah, Minal
> Cc: Mike Frysinger; Mukherjee, Somnath; linux-mtd@lists.infradead.org; Gupta, Pekon; Parikh, Urmil
> Subject: Re: MTD utils v1.5.0: MEMERASE64 ioctl failed error
> 
> Hi Minal,
> 
> On Wed, Oct 09, 2013 at 05:05:11AM +0000, Shah, Minal wrote:
>> I am using MTD utils version 1.5.0 for DRA7 (Vayu) platform.
>>
>> I have bad blocks in some of my NAND partitions which I am not able to erase from linux kernel.
> 
> MTD does not allow erasing bad blocks. This is prohibited by NAND
> datasheets. There are occasions where, for debugging purposes, one might
> need to erase a "bad" block that is not actually bad, but MTD does not
> provide an interface for doing this. An interface for doing so was
> discussed a while back, but nothing was merged.


Yes, this is a shortcoming in mtd, I think.  Blocks can be marked as bad, but
never "un-marked".  This is the reason I needed a module parameter
"ignore_badblocks" in the docg4 driver.

But any solution is complicated by the diversity of methods for marking bad
blocks (fixed or relocatable flash table, oob marker), and various reasons why
it is so marked (factory, user).  This is why u-boot says that 'nand scrub' is
potentially "unsafe".  You have to know the internals of your device.

Mike

Re: MTD utils v1.5.0: MEMERASE64 ioctl failed error

[Mike Frysinger]

[-- Attachment #1: Type: Text/Plain, Size: 1220 bytes --]

On Tuesday 15 October 2013 04:32:23 Shah, Minal wrote:
> Hi Brian,
> Thanks a lot for the explanation; I understood the reason for failure.
> 
> > So why do you want to erase bad blocks?
> 
> The reason why I want to erase bad blocks is because they actually don't
> seem to be bad. Block0 itself is shown as bad.
> I have few partitions (like MLO and u-boot env) of size equal to 1 erase
> block. The block is marked as bad and hence I cannot write anything to
> this partition. One of the solutions is to increase that partition size to
> 2 blocks so even if 1 block is bad then it will atleast be able to write
> to the other block of the same partition. But I don't think I should do
> this while the block is not really bad (what if all the blocks for some
> reason are marked as bad while they are actually not).
> 
> This question came in my mind from the fact that "nand scrub" command is
> allowed from u-boot which allows to also completely erase all the bad
> blocks; so why not from the kernel level.

as mentioned, there was discussion some time ago about adding this 
functionality.  but the work never finished.  if you're interested, i'd suggest 
picking up the work there.
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: MTD utils v1.5.0: MEMERASE64 ioctl failed error

[Brian Norris]

On Tue, Oct 15, 2013 at 12:26:48PM -0400, Mike Frysinger wrote:
> On Tuesday 15 October 2013 04:32:23 Shah, Minal wrote:
> > Hi Brian,
> > Thanks a lot for the explanation; I understood the reason for failure.
> > 
> > > So why do you want to erase bad blocks?
> > 
> > The reason why I want to erase bad blocks is because they actually don't
> > seem to be bad. Block0 itself is shown as bad.
> > I have few partitions (like MLO and u-boot env) of size equal to 1 erase
> > block. The block is marked as bad and hence I cannot write anything to
> > this partition. One of the solutions is to increase that partition size to
> > 2 blocks so even if 1 block is bad then it will atleast be able to write
> > to the other block of the same partition. But I don't think I should do
> > this while the block is not really bad (what if all the blocks for some
> > reason are marked as bad while they are actually not).
> > 
> > This question came in my mind from the fact that "nand scrub" command is
> > allowed from u-boot which allows to also completely erase all the bad
> > blocks; so why not from the kernel level.
> 
> as mentioned, there was discussion some time ago about adding this 
> functionality.  but the work never finished.  if you're interested, i'd suggest 
> picking up the work there.

Really old work:

  http://lists.infradead.org/pipermail/linux-mtd/2010-October/032577.html

More recent discussion:

  http://thread.gmane.org/gmane.linux.drivers.mtd/41624/

There were several proposals thrown around, and we ended up in a cycle
of over-engineering suggestions. I don't really see why we can't just
implement an ioctl (MEMNUKE? MEMFORCEERASE? MEMSCRUB?) that skips the
"don't erase bad blocks" check in MEMERASE64. There may be other
reasonable alternatives, but I would warn against anything too complex.

Brian