[U-Boot-Users] cfi driver - wrong sector protection detection with AMD flashes

[Martin Krause]

Hi all,

I tried the cfi flash driver on our custom MPC5200 board (TQM5200).
It works fine with different flash chips soldered. The only problem
is with an AMD (Spansion) AM29L160M Flash. There u-boot comes up with
_all_ sectors protected:

=> flinfo

Bank # 1: CFI conformant FLASH (32 x 16)  Size: 4 MB in 35 Sectors
 Erase timeout 16384 ms, write timeout 0 ms, buffer write timeout 1 ms, buffer s
ize 1
  Sector Start Addresses:
 FC000000  RO  FC008000  RO  FC00C000  RO  FC010000  RO  FC020000  RO
 FC040000  RO  FC060000 ERO  FC080000 ERO  FC0A0000 ERO  FC0C0000 ERO
 FC0E0000 ERO  FC100000 ERO  FC120000 ERO  FC140000 ERO  FC160000 ERO
 FC180000 ERO  FC1A0000 ERO  FC1C0000 ERO  FC1E0000 ERO  FC200000 ERO
 FC220000 ERO  FC240000 ERO  FC260000 ERO  FC280000 ERO  FC2A0000 ERO
 FC2C0000 ERO  FC2E0000 ERO  FC300000 ERO  FC320000 ERO  FC340000 ERO
 FC360000 ERO  FC380000 ERO  FC3A0000 ERO  FC3C0000 ERO  FC3E0000 ERO
=>

I tracked down the problem to the function flash_get_size() in cfi_flash.c.
There a special mechanism is used to query the flash for the protection
status of each sector. As far as I know, the used mechanism is not part of
the CFI Spec. And worse the mechanism is different for Intel and AMD
(Spansion).

For Intel flashes the sector protection status could be read when the flash
is in the CFI query state (at offset 02h on each sector address). For AMD
flashes this works only in the "autoselect" mode and in the actual driver
the "autoselect" mode is not used. And I could not say, if autoselect and CFI
Query mode would work at the same time.

I wonder if anyone else encountered similar problems with AMD flahes and
sector protection detection?

As solution for flashes with AMD command set I disabled the sector
protection detection in the cfi driver (see patch). Comments are welcome.

Is there anybody out there who uses the cfi driver and _really_ protected
flash sectors (not the software based protection used by u-boot)? If not,
we could perhaps disable the protection detection mechanism completely. Also
because the used mechanism is not part of the cfi specification.

Regards,
Martin

------------------------------------------------------------------
Martin Krause
TQ-Systems GmbH
Gut Delling, Mühlstr. 2, 82229 Seefeld, Germany
Fax: +49 8153 4223
Phone: +49 8153 9308 157
Personal Fax: +49 8153 9308 7157
Email: martin.krause at tqs.de
http://www.tq-group.com
------------------------------------------------------------------
-------------- next part --------------
A non-text attachment was scrubbed...
Name: patch
Type: application/octet-stream
Size: 877 bytes
Desc: patch
Url : http://lists.denx.de/pipermail/u-boot/attachments/20041026/3a9bb726/attachment.obj