FW: Handling boot on NAND Bad Blocks

FW: Handling boot on NAND Bad Blocks

[Kjell Ove Røte]

Hi

We have a product on the Freescale imx28 booting from a NAND Flash.
Lately we have seen high numbers of boards failing, and find we have received batches with marked bad blocks in lower boot addresses of Flash.
Typically we see bad blocks in partition where Barebox are located. 

We do also use the barebox command "bcb" to create the Boot Control Block. 
I did just test that using two partitions for the bootstream, and adding these to BCB handles my problem with bad block in one bootstream partition.

I realize we will get into same problem for bareboxenv, kernel and devicetree partitions.
Is there any sort of built-in mechanism or best practice dealing with this? 
Or do I need to make redundant copy of all partition and verifications before load?

We are currently using a rather old Barebox 2013.01.0, which has been working perfectly fine for many years.

Example running an update command on a partition with one bad block below.

barebox@Stentofon Turbine:/ update -t bootstream -d nand -f sti_boot
host 10.5.2.152 is alive

erasing partition /dev/nand0.bootstream.bb

Skipping bad block at 0x00140000

flashing sti_boot to /dev/nand0.bootstream.bb

Regards, 
Kjell Ove

DISCLAIMER:
This e-mail may contain confidential and privileged material for the sole use of the intended recipient. Any review, use, distribution or disclosure by others is strictly prohibited. If you are not the intended recipient (or authorized to receive for the recipient), please contact the sender by reply e-mail and delete all copies of this message.

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: FW: Handling boot on NAND Bad Blocks

[Robert Schwebel]

Hi Kjell,

On Thu, Nov 23, 2017 at 01:45:22PM +0000, Kjell Ove Røte wrote:
> We have a product on the Freescale imx28 booting from a NAND Flash.
> Lately we have seen high numbers of boards failing, and find we have
> received batches with marked bad blocks in lower boot addresses of
> Flash.

Is it possible to have a look at the fails? Log output, maybe dumps...

rsc
-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

RE: FW: Handling boot on NAND Bad Blocks

[Kjell Ove Røte]

Hi Robert

In the original case the board just reboots when the Barebox is to be loaded from the bootstream image. 
In log below you see that the power_prep and boot_prep is run but the Barebox is not started at all. 
So, in this case there is no Barebox issue, but I was looking for if there are some mechanism for handling boot environment, kernel load etc with Bad block management. 
For example something like writing and loading a kernel image, using skip block method, from a partition with bad blocks and image cannot be stored over linear addresses.

Booting an image where Barebox is stored in a NAND Bad Block:

PowerPrep start initialize power...

Configured for 5v only power source. Battery powered operation disabled.

Turbine PowerPrep POWER RESET, EMC hardness
Disable the chip shutting down by a fast falling edge of PSWITCH
Nov 24 201608:57:28
FRAC 0x92925552
Wait for ddr ready 1power 0x00820616
Frac 0x92925552
start change cpu freq
hbus 0x00000003
cpu 0x00010001
start test memory accress
ddr2 0x40000000
finish simple test
0x8050100b
0x8050800f

 

KJELL OVE RØTE
SENIOR SOFTWARE ENGINEER
ZENITEL NORWAY AS

Tel: +47 400 02 573
kjellove.rote@zenitel.com

Sandakerveien 24C, 0473 Oslo, Norway
P.O.Box 1068 Bekkajordet, NO-3194 Horten, Norway
Switchboard: +47 4000 2500
www.zenitel.com

-----Original Message-----
From: Robert Schwebel [mailto:r.schwebel@pengutronix.de] 
Sent: torsdag 23. november 2017 18.12
To: Kjell Ove Røte <kjellove.rote@zenitel.com>
Cc: barebox@lists.infradead.org
Subject: Re: FW: Handling boot on NAND Bad Blocks

Hi Kjell,

On Thu, Nov 23, 2017 at 01:45:22PM +0000, Kjell Ove Røte wrote:
> We have a product on the Freescale imx28 booting from a NAND Flash.
> Lately we have seen high numbers of boards failing, and find we have 
> received batches with marked bad blocks in lower boot addresses of 
> Flash.

Is it possible to have a look at the fails? Log output, maybe dumps...

rsc
-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

DISCLAIMER:
This e-mail may contain confidential and privileged material for the sole use of the intended recipient. Any review, use, distribution or disclosure by others is strictly prohibited. If you are not the intended recipient (or authorized to receive for the recipient), please contact the sender by reply e-mail and delete all copies of this message.
_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: FW: Handling boot on NAND Bad Blocks

[Robert Schwebel]

On Fri, Nov 24, 2017 at 08:49:22AM +0000, Kjell Ove Røte wrote:
> In the original case the board just reboots when the Barebox is to be loaded from the bootstream image. 
> In log below you see that the power_prep and boot_prep is run but the Barebox is not started at all. 
> So, in this case there is no Barebox issue, but I was looking for if there are some mechanism for handling boot environment, kernel load etc with Bad block management. 
> For example something like writing and loading a kernel image, using skip block method, from a partition with bad blocks and image cannot be stored over linear addresses.
> 
> Booting an image where Barebox is stored in a NAND Bad Block:
> 
> PowerPrep start initialize power...
> 
> Configured for 5v only power source. Battery powered operation disabled.
> 
> Turbine PowerPrep POWER RESET, EMC hardness
> Disable the chip shutting down by a fast falling edge of PSWITCH
> Nov 24 201608:57:28
> FRAC 0x92925552
> Wait for ddr ready 1power 0x00820616
> Frac 0x92925552
> start change cpu freq
> hbus 0x00000003
> cpu 0x00010001
> start test memory accress
> ddr2 0x40000000
> finish simple test
> 0x8050100b
> 0x8050800f

Hmm, that doesn't look familiar to me (we had other NAND issues
recently).

rsc
-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

RE: FW: Handling boot on NAND Bad Blocks

[Kjell Ove Røte]

Hi Robert

Did lookup the Boot ROM error codes:

0x8050100B Load command payload CRC failed.
0x8050800F ECC failed - data is not valid.

So, I assume in our case that it is stored a CRC of bootstream in the BCB image (with command bcb nand0.bootstream)
The MXS Bootrom code CRC check fails since image is written over a bad block. We do write bootstream image with Barebox doing a copy to a badblock aware device (nand0.bootstream.bb), so I guess the problem is that the ROM boot code is not skipping bad block on read. 
Partitioning with two bootstreams, and using command "bcb nand0.bootstream nand0.bootstream2", we do handle bad block on one of the bootstream partitions, but if one bad block in each we are stuck.

On my other questions on the other boot images like kernel and environment I do now see how it works with the bad block devices, and handling of skipping bad blocks. Adjusting some on the partition sizes allocating enough free erase blocks, we get good robustness on bad blocks in the rest of the NAND.


Best re


 

KJELL OVE RØTE
SENIOR SOFTWARE ENGINEER
ZENITEL NORWAY AS

Tel: +47 400 02 573
kjellove.rote@zenitel.com

Sandakerveien 24C, 0473 Oslo, Norway
P.O.Box 1068 Bekkajordet, NO-3194 Horten, Norway
Switchboard: +47 4000 2500
www.zenitel.com

-----Original Message-----
From: Robert Schwebel [mailto:robert@schwebel.de] 
Sent: torsdag 30. november 2017 22.44
To: Kjell Ove Røte <kjellove.rote@zenitel.com>
Cc: barebox@lists.infradead.org
Subject: Re: FW: Handling boot on NAND Bad Blocks

On Fri, Nov 24, 2017 at 08:49:22AM +0000, Kjell Ove Røte wrote:
> In the original case the board just reboots when the Barebox is to be loaded from the bootstream image. 
> In log below you see that the power_prep and boot_prep is run but the Barebox is not started at all. 
> So, in this case there is no Barebox issue, but I was looking for if there are some mechanism for handling boot environment, kernel load etc with Bad block management. 
> For example something like writing and loading a kernel image, using skip block method, from a partition with bad blocks and image cannot be stored over linear addresses.
> 
> Booting an image where Barebox is stored in a NAND Bad Block:
> 
> PowerPrep start initialize power...
> 
> Configured for 5v only power source. Battery powered operation disabled.
> 
> Turbine PowerPrep POWER RESET, EMC hardness Disable the chip shutting 
> down by a fast falling edge of PSWITCH Nov 24 201608:57:28 FRAC 
> 0x92925552 Wait for ddr ready 1power 0x00820616 Frac 0x92925552 start 
> change cpu freq hbus 0x00000003 cpu 0x00010001 start test memory 
> accress
> ddr2 0x40000000
> finish simple test
> 0x8050100b
> 0x8050800f

Hmm, that doesn't look familiar to me (we had other NAND issues recently).

rsc
-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

DISCLAIMER:
This e-mail may contain confidential and privileged material for the sole use of the intended recipient. Any review, use, distribution or disclosure by others is strictly prohibited. If you are not the intended recipient (or authorized to receive for the recipient), please contact the sender by reply e-mail and delete all copies of this message.
_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox