NAND Boot Issue

[Kamel BOUHARA]

Hi all,

Im working on my own imx27 designed board and wanted to try barebox on
it.Actually I'm trying it with the phytec board config and it seems to work
fine booting on my nfs kernel and rootfs.
But when Im trying to boot from my 128MiB NAND memory then it doesn't work
and there is no decompression stage from the kernel.
Here is what Im getting :

Board: Phytec phyCard-i.MX27
registered netconsole as cs1
cfi_probe: cfi_flash base: 0xc0000000 size: 0x02000000
NAND device: Manufacturer ID: 0x2c, Chip ID: 0xac (Micron NAND 512MiB 1,8V
8-bit)
Bad block table found at page 262080, version 0x01
Bad block table found at page 262016, version 0x01
Before PHY reset
After PHY reset
ULPI Vendor ID 0xffffff92 Product ID 0xffffff92
No ISP1504 found
***@ehci0: USB EHCI 1.00
ULPI Vendor ID 0xffffff92 Product ID 0xffffff92
No ISP1504 found
***@ehci1: USB EHCI 1.00
Malloc space: 0xa7a00000 -> 0xa7f00000 (size 5 MB)
Stack space : 0xa79f8000 -> 0xa7a00000 (size 32 kB)
running /env/bin/init...

Hit any key to stop autoboot: 1
booting kernel of type uImage from /dev/nand0.kernel.bb
barebox:/ bootu /dev/nand0.kernel.bb
commandline: console=ttymxc0,115200 video=imxfb:Chimei-LW700AT9003
arch_number: 1698

It seems I'm getting some issues with my nand since when I tried to flash it
and the erase task failed :

erasing partition /dev/nand0.kernel.bb

erase: Invalid argument

flashing uImage-apf27 to /dev/nand0.kernel.bb

TFTP from server 192.168.0.2 ('uImage-apf27' -> '/dev/nand0.kernel.bb')
#################################################################
#################################################################
#################################################################
#################################################################
#################################################################
#################################################################
#################################################################
#################################################################
#

How can I fix that ?
Thanks for the help.

[Sascha Hauer]

Post by Kamel BOUHARA
Hi all,
Im working on my own imx27 designed board and wanted to try barebox on
it.Actually I'm trying it with the phytec board config and it seems to work
fine booting on my nfs kernel and rootfs.
But when Im trying to boot from my 128MiB NAND memory then it doesn't work
and there is no decompression stage from the kernel.
Board: Phytec phyCard-i.MX27
registered netconsole as cs1
cfi_probe: cfi_flash base: 0xc0000000 size: 0x02000000
NAND device: Manufacturer ID: 0x2c, Chip ID: 0xac (Micron NAND 512MiB 1,8V
8-bit)
Bad block table found at page 262080, version 0x01
Bad block table found at page 262016, version 0x01
Before PHY reset
After PHY reset
ULPI Vendor ID 0xffffff92 Product ID 0xffffff92
No ISP1504 found
ULPI Vendor ID 0xffffff92 Product ID 0xffffff92
No ISP1504 found
Malloc space: 0xa7a00000 -> 0xa7f00000 (size 5 MB)
Stack space : 0xa79f8000 -> 0xa7a00000 (size 32 kB)
running /env/bin/init...
Hit any key to stop autoboot: 1
booting kernel of type uImage from /dev/nand0.kernel.bb
barebox:/ bootu /dev/nand0.kernel.bb

bootu is for starting raw kernel images. Do you have such an image?
Normally you have an uImage (bootm) or a zImage (bootz)

Post by Kamel BOUHARA
commandline: console=ttymxc0,115200 video=imxfb:Chimei-LW700AT9003
arch_number: 1698
It seems I'm getting some issues with my nand since when I tried to flash it
erasing partition /dev/nand0.kernel.bb
erase: Invalid argument

Works here. I can't see how erase could return -EINVAL. can you please
post the output of 'ls /dev/nand0*'?

Sascha

[Kamel BOUHARA]

Post by Sascha Hauer
bootu is for starting raw kernel images. Do you have such an image?
Normally you have an uImage (bootm) or a zImage (bootz)

Ok so that maybe the solution because I didn't tried with the bootm .... my
bad ! But why did I don't have the "Bad magic number" error like when I
tried the boot command ???

Here is the ls /dev result:

barebox:/ ls /dev/
zero defaultenv mem
nor0 nand0 nand_oob0
ram0 phy0 self_raw
self0 env_raw env0
nor0.barebox nor0.bareboxenv nor0.kernel
nor0.root nand0.barebox nand0.bareboxenv
nand0.kernel nand0.root nand0.root.bb
nand0.kernel.bb nand0.bareboxenv.bb nand0.barebox.bb

Post by Sascha Hauer

Post by Kamel BOUHARA
Hi all,
Im working on my own imx27 designed board and wanted to try barebox on
it.Actually I'm trying it with the phytec board config and it seems to

work

Post by Kamel BOUHARA
fine booting on my nfs kernel and rootfs.
But when Im trying to boot from my 128MiB NAND memory then it doesn't

work

Post by Kamel BOUHARA
and there is no decompression stage from the kernel.
Board: Phytec phyCard-i.MX27
registered netconsole as cs1
cfi_probe: cfi_flash base: 0xc0000000 size: 0x02000000
NAND device: Manufacturer ID: 0x2c, Chip ID: 0xac (Micron NAND 512MiB

1,8V

Post by Kamel BOUHARA
8-bit)
Bad block table found at page 262080, version 0x01
Bad block table found at page 262016, version 0x01
Before PHY reset
After PHY reset
ULPI Vendor ID 0xffffff92 Product ID 0xffffff92
No ISP1504 found
ULPI Vendor ID 0xffffff92 Product ID 0xffffff92
No ISP1504 found
Malloc space: 0xa7a00000 -> 0xa7f00000 (size 5 MB)
Stack space : 0xa79f8000 -> 0xa7a00000 (size 32 kB)
running /env/bin/init...
Hit any key to stop autoboot: 1
booting kernel of type uImage from /dev/nand0.kernel.bb
barebox:/ bootu /dev/nand0.kernel.bb

bootu is for starting raw kernel images. Do you have such an image?
Normally you have an uImage (bootm) or a zImage (bootz)

Post by Kamel BOUHARA
commandline: console=ttymxc0,115200 video=imxfb:Chimei-LW700AT9003
arch_number: 1698
It seems I'm getting some issues with my nand since when I tried to flash

it

Post by Kamel BOUHARA
erasing partition /dev/nand0.kernel.bb
erase: Invalid argument

Works here. I can't see how erase could return -EINVAL. can you please
post the output of 'ls /dev/nand0*'?
Sascha
--
Pengutronix e.K. | |
Industrial Linux Solutions | http://www.pengutronix.de/ |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: [phone] |
Amtsgericht Hildesheim, HRA 2686 | Fax: [phone]-5555 |

[Sascha Hauer]

Post by Kamel BOUHARA

Post by Sascha Hauer
bootu is for starting raw kernel images. Do you have such an image?
Normally you have an uImage (bootm) or a zImage (bootz)

Ok so that maybe the solution because I didn't tried with the bootm .... my
bad ! But why did I don't have the "Bad magic number" error like when I
tried the boot command ???

A raw image does not have any magic number you could check, so the bootu
command relies on the user passing the correct image.

Post by Kamel BOUHARA
barebox:/ ls /dev/
zero defaultenv mem
nor0 nand0 nand_oob0
ram0 phy0 self_raw
self0 env_raw env0
nor0.barebox nor0.bareboxenv nor0.kernel
nor0.root nand0.barebox nand0.bareboxenv
nand0.kernel nand0.root nand0.root.bb
nand0.kernel.bb nand0.bareboxenv.bb nand0.barebox.bb

Sorry, I meant the output of 'ls -l /dev/nand0.*' to check the sizes of
your partitions.

Sascha

[Kamel BOUHARA]

I tried with a zImage, here is what I'm getting :

barebox:/ bootz /dev/nand0.kernel.bb
invalid magic 0x00002018

And here the ls -l /dev/nand0.* output :

barebox:/ ls -l /dev/nand0.*
crw------- 393216 /dev/nand0.barebox.bb
crw------- 131072 /dev/nand0.bareboxenv.bb
crw------- 2883584 /dev/nand0.kernel.bb
crw------- 533200896 /dev/nand0.root.bb
crw------- 533667840 /dev/nand0.root
crw------- 2764800 /dev/nand0.kernel
crw------- 131072 /dev/nand0.bareboxenv
crw------- 307200 /dev/nand0.barebox

Here is how I did my kernel and root partitions, following this steps :

addpart /dev/nand0 300k(barebox)ro,128k(bareboxenv),2700k(kernel),-(root)
nand -a /dev/nand0.*
update -t kernel -d nand
update -t rootfs -d nand

Is there any specific command to erase the nand before flashing it ?

Post by Sascha Hauer

Post by Kamel BOUHARA

Post by Sascha Hauer
bootu is for starting raw kernel images. Do you have such an image?
Normally you have an uImage (bootm) or a zImage (bootz)

Ok so that maybe the solution because I didn't tried with the bootm ....

my

Post by Kamel BOUHARA
bad ! But why did I don't have the "Bad magic number" error like when I
tried the boot command ???

A raw image does not have any magic number you could check, so the bootu
command relies on the user passing the correct image.

Post by Kamel BOUHARA
barebox:/ ls /dev/
zero defaultenv mem
nor0 nand0 nand_oob0
ram0 phy0 self_raw
self0 env_raw env0
nor0.barebox nor0.bareboxenv nor0.kernel
nor0.root nand0.barebox nand0.bareboxenv
nand0.kernel nand0.root nand0.root.bb
nand0.kernel.bb nand0.bareboxenv.bb nand0.barebox.bb

Sorry, I meant the output of 'ls -l /dev/nand0.*' to check the sizes of
your partitions.
Sascha
--
Pengutronix e.K. | |
Industrial Linux Solutions | http://www.pengutronix.de/ |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: [phone] |
Amtsgericht Hildesheim, HRA 2686 | Fax: [phone]-5555 |

[Sascha Hauer]

Post by Kamel BOUHARA
barebox:/ bootz /dev/nand0.kernel.bb
invalid magic 0x00002018
barebox:/ ls -l /dev/nand0.*
crw------- 393216 /dev/nand0.barebox.bb
crw------- 131072 /dev/nand0.bareboxenv.bb
crw------- 2883584 /dev/nand0.kernel.bb
crw------- 533200896 /dev/nand0.root.bb
crw------- 533667840 /dev/nand0.root
crw------- 2764800 /dev/nand0.kernel
crw------- 131072 /dev/nand0.bareboxenv
crw------- 307200 /dev/nand0.barebox
addpart /dev/nand0 300k(barebox)ro,128k(bareboxenv),2700k(kernel),-(root)

Ok, that puts some light into this. Please make sure your partitions are
aligned to eraseblock boundaries, usually 128k. so:

addpart /dev/nand0 512k(barebox)ro,128k(bareboxenv),3M(kernel),-(root)

barebox currently does not enforce this, it's probably time to change
this. Unaligned partitions are fine, but the partition code should mark
them as readonly.
Currently we have no way of passing the eraseblocksize through to the
partition code, so it's not trivial to enforce alignment.

Sascha

[Kamel BOUHARA]

Post by Sascha Hauer
Unaligned partitions are fine, but the partition code should mark
them as readonly.

That's why I was unable to erase it getting -EINVAL, things are clear now

It works fine, my kernel is booting :) !

Thank you Sascha

Post by Sascha Hauer

Post by Kamel BOUHARA
barebox:/ bootz /dev/nand0.kernel.bb
invalid magic 0x00002018
barebox:/ ls -l /dev/nand0.*
crw------- 393216 /dev/nand0.barebox.bb
crw------- 131072 /dev/nand0.bareboxenv.bb
crw------- 2883584 /dev/nand0.kernel.bb
crw------- 533200896 /dev/nand0.root.bb
crw------- 533667840 /dev/nand0.root
crw------- 2764800 /dev/nand0.kernel
crw------- 131072 /dev/nand0.bareboxenv
crw------- 307200 /dev/nand0.barebox
addpart /dev/nand0 300k(barebox)ro,128k(bareboxenv),2700k(kernel),-(root)

Ok, that puts some light into this. Please make sure your partitions are
addpart /dev/nand0 512k(barebox)ro,128k(bareboxenv),3M(kernel),-(root)
barebox currently does not enforce this, it's probably time to change
this. Unaligned partitions are fine, but the partition code should mark
them as readonly.
Currently we have no way of passing the eraseblocksize through to the
partition code, so it's not trivial to enforce alignment.
Sascha
--
Pengutronix e.K. | |
Industrial Linux Solutions | http://www.pengutronix.de/ |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: [phone] |
Amtsgericht Hildesheim, HRA 2686 | Fax: [phone]-5555 |