Unlocked flash is re-locked by unknow

Unlocked flash is re-locked by unknow

[Henrik Bork Steffensen]

Hi,
We have this custom arm based board were we ran Linux 2.4 for 3 years,
now we are moving to 2.6 and has problems with our NOR Flash.

After boot the flash is locked which is normal for this chip.
Then we run flash_unlock and mounts the jffs2 fs.
At this point we can create /remove files.
But after a some time (or some actions) we get write errors.
If we re-run flash_unlock, everything works fine again for
some time (or some actions), then it locks again....

We have tried to figure out which action lock's the device without
success...

We have tried kernel 2.6.16.5 and 2.6.17-rc4 with and without
mtd patch (stable and git).

The flash is: AT49BV320 (4MByte, NOR, almost-CFI-compliant)
It is auto discovered by cfi.
The chip claims to be original Intel flash, but it has another
mapping of its capabilities bits (which is soo nice). We have made
an ugly fix to this issue by changing the bitmap which is read
from the device. Removing this code has no effect on the described
problem, just thought you should now :-)

This is not a HW problem since booting 2.4 makes it work again.

Please provide any hint, a simple link to will be fine.

Thanks in advance.

-- 
med venlig hilsen / best regards / Mit freundlichen Grüßen

Henrik Bork Steffensen

RoseTechnology A/S
http://www.rosetechnology.dk

Re: Unlocked flash is re-locked by unknow

[Nicolas Pitre]

On Tue, 30 May 2006, Henrik Bork Steffensen wrote:

> Hi,
> We have this custom arm based board were we ran Linux 2.4 for 3 years,
> now we are moving to 2.6 and has problems with our NOR Flash.
> 
> After boot the flash is locked which is normal for this chip.
> Then we run flash_unlock and mounts the jffs2 fs.
> At this point we can create /remove files.
> But after a some time (or some actions) we get write errors.
> If we re-run flash_unlock, everything works fine again for
> some time (or some actions), then it locks again....
> 
> We have tried to figure out which action lock's the device without
> success...

A spike on the flash power maybe?

> This is not a HW problem since booting 2.4 makes it work again.

Note that some newer flash parts have a pin that "force-locks" the flash 
when asserted.  Maybe you have one of those and the pin is floating?  
And to sum it all the 2.6 kernel somehow has a different EMI picture 
than the 2.4 and it causes the flash to lock itself that way?  

Or maybe just modifying the kernel version number from 2.6 to 2.4 in the 
source and recompiling will fix it?  ;-)

> Please provide any hint, a simple link to will be fine.

Are you sure your 2.4.x code is not automatically unlocking the flash by 
any chance which might just be masking the problem?


Nicolas

Re: Unlocked flash is re-locked by unknow

[Henrik Bork Steffensen]

Nicolas Pitre wrote:

>A spike on the flash power maybe?
>  
>
Yes, that's it!

The 2.6 seems to generate 20-40% larger spikes on Vcc than
the old 2.4.19-rmk7, which runs 30 Bogomips were the 2.6
runs at 90 Bogomips. (i have no idea of this is just the
bogomips loop which runs faster or the cpu actually runs
faster)

Thanks you very much - we obviously needed fresh eyes on
the subject.


-- 
med venlig hilsen / best regards / Mit freundlichen Grüßen

Henrik Bork Steffensen

RoseTechnology A/S
http://www.rosetechnology.dk