RE: [Gumstix-users] Flash Error: bad magic number

[Richard T. S. <rs...@pa...>]

I downloaded what I believe is the shipped version (412) of the root fs =
from
http://sourceforge.net/projects/gumstix/.

When all else fails, I use that version.  Unfortunately, WiFi isn't =
enabled
by default (nor is pcmcia) so I have to spend a few minutes typing to =
get
everything working.


-----Original Message-----
From: gum...@li...
[mailto:gum...@li...] On Behalf Of Pooya
Karimian
Sent: Saturday, July 16, 2005 5:29 PM
To: gum...@li...
Subject: [Gumstix-users] Flash Error: bad magic number

Hi,

You can find a more easily readable, detailed version of this email
with error messages on our Wiki page here:
http://autolab.cs.sfu.ca/wiki/GumstixFlashProblem

Our gumstix connex 400 is now showing this message while booting:

%%
U-Boot 1.1.2 (Feb 11 2005 - 12:59:06)


*** Welcome to Gumstix ***

U-Boot code: A3F00000 -> A3F2252C  BSS: -> A3F5738C
RAM Configuration:
Bank #0: a0000000 64 MB
Flash:  4 MB
Hit any key to stop autoboot:  0=20
### JFFS2 loading 'boot/uImage' to 0xa2000000
Scanning JFFS2 FS: . done.
find_inode failed for name=3Dboot
load: Failed to find inode
### JFFS2 LOAD ERROR<0> for boot/uImage!
## Booting image at a2000000 ...
Bad Magic Number
GUM>=20
%%

It happened while I was working with Robostix (the protype version),
and tweener serial board. I tried reflashing gumstix but that failed
too. After loading the root_fs_arm at a2000000, "erase" command shows
some warnings about protected sectors: "Warning: 2 protected sectors
will not be erased!". And the "cp.b" command says: "Can't write to
protected Flash sectors".

The "flinfo" command output shows:

%%
Bank # 1: CFI conformant FLASH (16 x 16)  Size: 4 MB in 32 Sectors
Sector Start Addresses:
00000000 (RO)	00020000 (RO)	00040000 (RO)	00060000 (RO)	00080000

%%

One strange thing here is, that the actual flash chip model is the
same as the one on the older gumstix basix 200, but the "flinfo"
command shows different model numbers. as you can see below:

%%
Bank # 1: INTEL 28F320J3A
Size: 4 MB in 32 Sectors
Sector Start Addresses:
00000000 (RO) 00020000 (RO) 00040000		00060000
00080000	 =20
%%

I can force unprotecting the read-only sectors:

%%
GUM> protect off 1:2-31
GUM> era 1:2-31
GUM> cp.b a2000000 40000 ${filesize}
Copy to Flash... done
%%

But seems that it is not writing into the flash, because the same
problem (bad magic number) occures after rebooting. And the memory
dump of address 40000 confirms that.

These are env variables. I added the noinitrd option because of an old
post I saw in the mailing list but that didn't help either:

%%
bootcmd=3Dfsload a2000000 boot/uImage;bootm a2000000
bootargs=3Dconsole=3DttyS0,115200n8 root=3D1f02 rootfstype=3Djffs2 =
noinitrd
reboot=3Dcold,hard
%%

I was worried to destory the currently working U-BOOT, so I did not
try reflashing the U-BOOT. If you think that will help, is there a way
to get a precompiled safe u-boot.bin instead of compiled it by myself?
Because the one I have in build_arm/u-boot-1.1.2/u-boot.bin is
different from what is in 0 address of the flash (although both seem
to be version 1.1.2).

Can this be a physical problem with the flash chip? If so, why is the
u-boot still working fine?

Any suggestions or comments?

--Pooya Karimian=20
 -Autonomy Lab
 -School of Computing Science
 -Simon Fraser University


-------------------------------------------------------
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
from IBM. Find simple to follow Roadmaps, straightforward articles,
informative Webcasts and more! Get everything you need to get up to
speed, fast. http://ads.osdn.com/?ad_idt77&alloc_id=16492&op=3Dick
_______________________________________________
gumstix-users mailing list
gum...@li...
https://lists.sourceforge.net/lists/listinfo/gumstix-users