UBI issues on SAMA5D31 NOR flash

U-Boot Andy Pont 2 participants 10 messages Sep 29, 2014 – Oct 2, 2014

Share & tools

Andy Pont

author

2014-09-29 16:40 UTC #0 lore raw

Hello,

Using the patches from Bo Shen to support the NOR flash on the Atmel SAMA5D31EK hardware platform I am trying to get a UBI root file system within the NOR flash, at offset 0x10000000 in the memory map, to work but get errors from the "ubi part" command using U-Boot 2014.04:

U-Boot> mtdparts

device nor0 <physmap-flash.0>, # parts = 5

 #: name                size            offset          mask_flags
 0: bootstrap           0x00020000      0x00000000      1
 1: u-boot              0x000c0000      0x00020000      1
 2: fdt                 0x00020000      0x000e0000      0
 3: kernel              0x00400000      0x00100000      0
 4: rootfs              0x00a00000      0x00500000      0

active partition: nor0,0 - (bootstrap) 0x00020000 @ 0x00000000

defaults:
mtdids : nor0=physmap-flash.0
mtdparts:
mtdparts=physmap-flash.0:128k(bootstrap)ro,768k(u-boot)ro,128k(fdt),4m(kerne l),10m(rootfs)

U-Boot> ubi part rootfs
UBI: attaching mtd2 to ubi0
UBI: physical eraseblock size: 131072 bytes (128 KiB)
UBI: logical eraseblock size: 130944 bytes
UBI: smallest flash I/O unit: 1
UBI: VID header offset: 64 (aligned 64)
UBI: data offset: 128
UBI error: vtbl_check: volume table check failed: record 0, error 9 UBI error: ubi_init: cannot attach mtd2
UBI error: ubi_init: UBI error: cannot initialize UBI, error -22 UBI init error 22

The UBI settings are:

(2048) Image Size
(130944) Logical Erase Block Size

(1)   Minimum I/O Size
[ ]   mkfs.ubifs Space Fixup                                           
()    mkfs.ubifs Options

*** UBI Image Properties ***
[*] Create UBI Volume
(10MiB) Root File System Volume Size
(128 KiB) Physical Erase Block Size

(1)     Sub-page Size                                                
()      ubinize Configuration File                                   
()      ubinize Options

Any ideas as to what I have missed?

Thanks,

Andy.

Heiko Schocher

2014-09-30 04:39 UTC #1 lore raw

Hello Andy,

Am 29.09.2014 18:40, schrieb Andy Pont:

Quoted 3 lines

Hello,
Using the patches from Bo Shen to support the NOR flash on the Atmel

which patches?

Quoted 22 lines

SAMA5D31EK hardware platform I am trying to get a UBI root file system within the NOR flash, at offset 0x10000000 in the memory map, to work but get errors from the "ubi part" command using U-Boot 2014.04:
U-Boot> mtdparts
device nor0<physmap-flash.0>, # parts = 5
  #: name                size            offset          mask_flags
  0: bootstrap           0x00020000      0x00000000      1
  1: u-boot              0x000c0000      0x00020000      1
  2: fdt                 0x00020000      0x000e0000      0
  3: kernel              0x00400000      0x00100000      0
  4: rootfs              0x00a00000      0x00500000      0
active partition: nor0,0 - (bootstrap) 0x00020000 @ 0x00000000
defaults:
mtdids : nor0=physmap-flash.0
mtdparts:
mtdparts=physmap-flash.0:128k(bootstrap)ro,768k(u-boot)ro,128k(fdt),4m(kerne l),10m(rootfs)

Quoted 8 lines

U-Boot> ubi part rootfs
UBI: attaching mtd2 to ubi0
UBI: physical eraseblock size: 131072 bytes (128 KiB)
UBI: logical eraseblock size: 130944 bytes
UBI: smallest flash I/O unit: 1
UBI: VID header offset: 64 (aligned 64)
UBI: data offset: 128
UBI error: vtbl_check: volume table check failed: record 0, error 9

seems ubi does not find the volume table ... could you activate ubi error messages ubi_err() ?

Quoted 16 lines

UBI error: ubi_init: cannot attach mtd2
UBI error: ubi_init: UBI error: cannot initialize UBI, error -22 UBI init error 22
The UBI settings are:
(2048) Image Size
(130944) Logical Erase Block Size
(1)   Minimum I/O Size
[ ]   mkfs.ubifs Space Fixup
()    mkfs.ubifs Options
*** UBI Image Properties ***
[*] Create UBI Volume
(10MiB) Root File System Volume Size
(128 KiB) Physical Erase Block Size

Quoted 5 lines

(1)     Sub-page Size
()      ubinize Configuration File
()      ubinize Options
Any ideas as to what I have missed?

Did the ubi partition work under linux? Did you tried current ML U-Boot (there MTD/UBI and UBIFS is synced with linux 3.15) ...

bye,
Heiko

Signature · 2 lines

DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany

Andy Pont

author

2014-09-30 09:04 UTC #2 lore raw

Hello Heiko,

Quoted 3 lines

Quoted 1 line
Using the patches from Bo Shen to support the NOR flash on the Atmel
which patches?

I can't find the references in Patchman but they were sent to the mailing list on 18th July 2014 in a series called "ARM: atmel: sama5d3xek: enable NOR flash support". The patches basically just add the board features to enable the IO pins for the NOR flash and to map it into the memory map at 0x10000000.

Quoted 11 lines

Quoted 8 lines
U-Boot> ubi part rootfs
UBI: attaching mtd2 to ubi0
UBI: physical eraseblock size: 131072 bytes (128 KiB)
UBI: logical eraseblock size: 130944 bytes
UBI: smallest flash I/O unit: 1
UBI: VID header offset: 64 (aligned 64)
UBI: data offset: 128
UBI error: vtbl_check: volume table check failed: record 0, error 9
seems ubi does not find the volume table ... could you activate ubi error messages ubi_err() ?

Adding:

#define DEBUG
#define CONFIG_MTD_UBI_DEBUG
#define CONFIG_MTD_UBI_DEBUG_MSG
#define CONFIG_UBIFS_FS_DEBUG

Gives the following output:

U-Boot> ubi part rootfs

---mtdparts_init---
last_ids : nor0=physmap-flash.0
env_ids : nor0=physmap-flash.0
last_parts:
mtdparts=physmap-flash.0:128k(bootstrap)ro,768k(u-boot)ro,128k(fdt),4m(kerne l),10m(rootfs)
env_parts :
mtdparts=physmap-flash.0:128k(bootstrap)ro,768k(u-boot)ro,128k(fdt),4m(kerne l),10m(rootfs)

last_partition : nor0,0
env_partition : nor0,0
--- find_dev_and_part ---
id = rootfs
--- find_dev_and_part ---
id = rootfs
Creating 1 MTD partitions on "nor0":
0x000000500000-0x000000f00000 : "mtd=4"
UBI: attaching mtd2 to ubi0
UBI: physical eraseblock size: 131072 bytes (128 KiB)
UBI: logical eraseblock size: 130944 bytes
UBI: smallest flash I/O unit: 1
UBI: VID header offset: 64 (aligned 64)
UBI: data offset: 128
UBI error: vtbl_check: volume table check failed: record 0, error 9 UBI error: ubi_init: cannot attach mtd2
UBI error: ubi_init: UBI error: cannot initialize UBI, error -22 UBI init error 22
Command failed, result=22

Quoted 2 lines

Did the ubi partition work under linux? Did you tried current ML U-Boot (there MTD/UBI and UBIFS is synced with linux 3.15) ...

I haven't booted to Linux yet. If I try to boot with the kernel that is in the NOR flash then I get the following output from the console messages:

0.nor_flash: Found 1 x16 devices at 0x0 in 16-bit bank. Manufacturer ID 0x000089 Chip ID 0x00899a
Intel/Sharp Extended Query Table at 0x010A
Intel/Sharp Extended Query Table at 0x010A
Intel/Sharp Extended Query Table at 0x010A
Intel/Sharp Extended Query Table at 0x010A
Intel/Sharp Extended Query Table at 0x010A
Using buffer write method
Using auto-unlock on power-up/resume
cfi_cmdset_0001: Erase suspend on write enabled
UBI error: ubi_init: UBI error: cannot initialize UBI, error -19

Is there something that I need to do in order to tell U-Boot and the kernel that the base address of the NOR flash device is 0x10000000 rather than 0x0?

Thanks,

Andy.

Heiko Schocher

2014-09-30 09:36 UTC #3 lore raw

Hello Andy,

Am 30.09.2014 11:04, schrieb Andy Pont:

Quoted 56 lines

Hello Heiko,
Quoted 3 lines
Quoted 1 line
Using the patches from Bo Shen to support the NOR flash on the Atmel
which patches?
I can't find the references in Patchman but they were sent to the mailing list on 18th July 2014 in a series called "ARM: atmel: sama5d3xek: enable NOR flash support". The patches basically just add the board features to enable the IO pins for the NOR flash and to map it into the memory map at 0x10000000.
Quoted 11 lines
Quoted 8 lines
U-Boot> ubi part rootfs
UBI: attaching mtd2 to ubi0
UBI: physical eraseblock size: 131072 bytes (128 KiB)
UBI: logical eraseblock size: 130944 bytes
UBI: smallest flash I/O unit: 1
UBI: VID header offset: 64 (aligned 64)
UBI: data offset: 128
UBI error: vtbl_check: volume table check failed: record 0, error 9
seems ubi does not find the volume table ... could you activate ubi error messages ubi_err() ?
Adding:
#define DEBUG
#define CONFIG_MTD_UBI_DEBUG
#define CONFIG_MTD_UBI_DEBUG_MSG
#define CONFIG_UBIFS_FS_DEBUG
Gives the following output:
U-Boot> ubi part rootfs
---mtdparts_init---
last_ids : nor0=physmap-flash.0
env_ids : nor0=physmap-flash.0
last_parts:
mtdparts=physmap-flash.0:128k(bootstrap)ro,768k(u-boot)ro,128k(fdt),4m(kerne l),10m(rootfs)
env_parts :
mtdparts=physmap-flash.0:128k(bootstrap)ro,768k(u-boot)ro,128k(fdt),4m(kerne l),10m(rootfs)
last_partition : nor0,0
env_partition : nor0,0
--- find_dev_and_part ---
id = rootfs
--- find_dev_and_part ---
id = rootfs
Creating 1 MTD partitions on "nor0":
0x000000500000-0x000000f00000 : "mtd=4"
UBI: attaching mtd2 to ubi0
UBI: physical eraseblock size: 131072 bytes (128 KiB)
UBI: logical eraseblock size: 130944 bytes

Quoted 8 lines

UBI: smallest flash I/O unit: 1
UBI: VID header offset: 64 (aligned 64)
UBI: data offset: 128
UBI error: vtbl_check: volume table check failed: record 0, error 9 UBI error: ubi_init: cannot attach mtd2
UBI error: ubi_init: UBI error: cannot initialize UBI, error -22 UBI init error 22
Command failed, result=22

Hmm.. I thought to see something like this:

"too large reserved_pebs %d, good PEBs %d"

I am not familiar with this HW ... I think, you have to debug into it ... Maybe Bo has an idea?

Quoted 6 lines

Quoted 2 lines
Did the ubi partition work under linux? Did you tried current ML U-Boot (there MTD/UBI and UBIFS is synced with linux 3.15) ...
I haven't booted to Linux yet. If I try to boot with the kernel that is in the NOR flash then I get the following output from the console messages:

Quoted 11 lines

0.nor_flash: Found 1 x16 devices at 0x0 in 16-bit bank. Manufacturer ID 0x000089 Chip ID 0x00899a
Intel/Sharp Extended Query Table at 0x010A
Intel/Sharp Extended Query Table at 0x010A
Intel/Sharp Extended Query Table at 0x010A
Intel/Sharp Extended Query Table at 0x010A
Intel/Sharp Extended Query Table at 0x010A
Using buffer write method
Using auto-unlock on power-up/resume
cfi_cmdset_0001: Erase suspend on write enabled
UBI error: ubi_init: UBI error: cannot initialize UBI, error -19

Hmm.. linux returns another error ...

Quoted 2 lines

Is there something that I need to do in order to tell U-Boot and the kernel that the base address of the NOR flash device is 0x10000000 rather than 0x0?

I don t know ... but it seems, the chip is detected, so the base addr should be correct.

bye,
Heiko

Signature · 2 lines

DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany

Andy Pont

author

2014-09-30 11:19 UTC #4 lore raw

Hello Heiko,

Quoted 5 lines

Hmm.. I thought to see something like this:
"too large reserved_pebs %d, good PEBs %d"
I am not familiar with this HW ... I think, you have to debug into it

Changing the dbg_err() to ubi_err() and also adding a bit more output then I get:

UBI error: vtbl_check: reserved_pebs 81, ubi->good_peb_count 80 UBI error: vtbl_check: too large reserved_pebs, good PEBs 80

Use of UBI is new to me so where are the PEBs configured?

Thanks,

Andy.

Heiko Schocher

2014-10-01 11:35 UTC #5 lore raw

Am 30.09.2014 13:19, schrieb Andy Pont:

Quoted 8 lines

Hello Heiko,
Quoted 5 lines
Hmm.. I thought to see something like this:
"too large reserved_pebs %d, good PEBs %d"
I am not familiar with this HW ... I think, you have to debug into it

Quoted 7 lines

Changing the dbg_err() to ubi_err() and also adding a bit more output then I get:
UBI error: vtbl_check: reserved_pebs 81, ubi->good_peb_count 80 UBI error: vtbl_check: too large reserved_pebs, good PEBs 80
Use of UBI is new to me so where are the PEBs configured?

Good question ... looking into vtbl_check(), the "reserved_pebs" value is read from the record in the volume table:

                 reserved_pebs = be32_to_cpu(vtbl[i].reserved_pebs);

"@reserved_pebs: how many physical eraseblocks are reserved for this volume"

The "ubi->good_peb_count 80" seems correct to me.

maybe ubi image creation was bad?
I can only speculate ... sorry.

bye,
Heiko

Signature · 2 lines

DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany

Andy Pont

author

2014-10-02 08:16 UTC #6 lore raw

Hello Heiko,

Quoted 5 lines

Quoted 4 lines
UBI error: vtbl_check: reserved_pebs 81, ubi->good_peb_count 80 UBI error: vtbl_check: too large reserved_pebs, good PEBs 80
Use of UBI is new to me so where are the PEBs configured?

Quoted 7 lines

Good question ... looking into vtbl_check(), the "reserved_pebs" value is read from the record in the volume table:
                 reserved_pebs = be32_to_cpu(vtbl[i].reserved_pebs);
"@reserved_pebs: how many physical eraseblocks are reserved for this volume"

I have changed the size of the partition as defined in U-Boot and also reduced the maximum image size when creating the UBI image and have found a combination that moves me on to the next error:

U-Boot> ubi part rootfs
UBI: attaching mtd2 to ubi0
UBI: physical eraseblock size: 131072 bytes (128 KiB)
UBI: logical eraseblock size: 130944 bytes
UBI: smallest flash I/O unit: 1
UBI: VID header offset: 64 (aligned 64)
UBI: data offset: 128
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
UBI error: ubi_io_write: error -5 while writing 64 bytes to PEB 0:0, written 0 bytes
UBI error: erase_worker: failed to erase PEB 0, error -5
UBI error: erase_worker: bad physical eraseblock 0 detected
UBI warning: ubi_ro_mode: switch to read-only mode
UBI error: do_work: work failed with error code -5
UBI error: autoresize: cannot auto-resize volume 0
UBI error: ubi_init: cannot attach mtd2
UBI error: ubi_init: UBI error: cannot initialize UBI, error -30 UBI init error 30

Hopefully this may be the last hurdle to overcome but somehow I think maybe not!

Regards,

Andy.

Heiko Schocher

2014-10-02 09:10 UTC #7 lore raw

Hello Andy,

Am 02.10.2014 10:16, schrieb Andy Pont:

Quoted 37 lines

Hello Heiko,
Quoted 5 lines
Quoted 4 lines
UBI error: vtbl_check: reserved_pebs 81, ubi->good_peb_count 80 UBI error: vtbl_check: too large reserved_pebs, good PEBs 80
Use of UBI is new to me so where are the PEBs configured?
Quoted 7 lines
Good question ... looking into vtbl_check(), the "reserved_pebs" value is read from the record in the volume table:
                  reserved_pebs = be32_to_cpu(vtbl[i].reserved_pebs);
"@reserved_pebs: how many physical eraseblocks are reserved for this volume"
I have changed the size of the partition as defined in U-Boot and also reduced the maximum image size when creating the UBI image and have found a combination that moves me on to the next error:
U-Boot> ubi part rootfs
UBI: attaching mtd2 to ubi0
UBI: physical eraseblock size: 131072 bytes (128 KiB)
UBI: logical eraseblock size: 130944 bytes
UBI: smallest flash I/O unit: 1
UBI: VID header offset: 64 (aligned 64)
UBI: data offset: 128
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!

Quoted 8 lines

- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
- Warning: 1 protected sectors will not be erased!
UBI error: ubi_io_write: error -5 while writing 64 bytes to PEB 0:0, written 0 bytes

What says flinfo ? It seems you have protected sectors on your nor flash ... you must unprotect them before using it ...

Quoted 11 lines

UBI error: erase_worker: failed to erase PEB 0, error -5 UBI error: erase_worker: bad physical eraseblock 0 detected UBI warning: ubi_ro_mode: switch to read-only mode
UBI error: do_work: work failed with error code -5
UBI error: autoresize: cannot auto-resize volume 0
UBI error: ubi_init: cannot attach mtd2
UBI error: ubi_init: UBI error: cannot initialize UBI, error -30 UBI init error 30
Hopefully this may be the last hurdle to overcome but somehow I think maybe not!

Cheer up!

bye,
Heiko

Signature · 2 lines

DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany

Andy Pont

author

2014-10-02 09:16 UTC #8 lore raw

Hello Heiko,

Quoted 2 lines

What says flinfo ? It seems you have protected sectors on your nor flash ... you must unprotect them before using it ...

flinfo says that all sectors are read only as the flash device supports block locking and powers up with all sectors in their locked state.

The board configuration file includes CONFIG_SYS_FLASH_PROTECTION in order to allow the "protect off" command to work prior to manual updates of the flash content.

Quoted 1 line

Cheer up!

I'll try! :-)

Regards,

Andy.

Andy Pont

author

2014-10-02 09:44 UTC #9 lore raw

Hello Heiko,

Quoted 9 lines

Quoted 2 lines
What says flinfo ? It seems you have protected sectors on your nor flash ... you must unprotect them before using it ...
flinfo says that all sectors are read only as the flash device supports block locking and powers up with all sectors in their locked state.
The board configuration file includes CONFIG_SYS_FLASH_PROTECTION in order to allow the "protect off" command to work prior to manual updates of the flash content.

Running "protect off" on the UBI area before "ubi part rootfs" makes the command work without any issues. Now just have to figure out why the Linux kernel isn't detecting the partition and mounting it...

Regards,

Andy.