2007-04-03 05:21:47     SPI eeprom write read

Document created by Aaronwu Employee on Sep 17, 2013
Version 1Show Document
  • View in full screen mode

2007-04-03 05:21:47     SPI eeprom write read

Richard Aldrich (UNITED KINGDOM)

Message: 23979    I'm using the latest SVN versions of uboot, toolchain and uclinux (updated and compiled yesterday).


I followed the document http://docs.blackfin.uclinux.org/doku.php?id=accessing_serial_nor_flash_with_u-boot

to attempt to write to SPI flash (ST M25P64). I call the eeprom write command, and it appears to works, then I read it back to another memory location.


However, when I do the memory compare, it says that 0 bytes are the same (output is something like:

cmp.b 0x1000000 0x2000000 $(filesize)

byte at 0x01000000 (0x40) != byte at 0x02000000 (0x00)

Total of 0 bytes were the same)


where 0x2000000 is the location I have read the eeprom to.


Does anyone have any ideas on what I should investigate next?




2007-04-04 00:37:59     Re: SPI eeprom write read

Mike Frysinger (UNITED STATES)

Message: 24009    this is a custom board right ?  i would first verify memory is sane by using the u-boot "mtest" command ... then try writing out all like 0x00 and 0xFF to the flash and then read it back out and use md.b to verify the location you read back out matches




2007-04-04 09:01:50     Re: SPI eeprom write read

Richard Aldrich (UNITED KINGDOM)

Message: 24039    Ok, I am now able to write to and read from the SPI memory (its on a BF 537-stamp board). My problem came down to the wire length / type that had been used to connect the memory to a header.


I am able to load the uImage and read it back, and boot it and this process works well. All the bytes are identical.


However, I now want to investigate booting the u-boot from SPI. What I understand is that I compile u-boot with the



in the file found at: u-boot/include/configs/bf537-stamp.h


I do this, and then I have written the u-boot.ldr image to the eeprom, at offset 0x0.


I then switch SW16 to 3, and power up. ...nothing happens, I don't get anything back in minicom.


Am I missing a stage along the way? Btw, the board is Rev 0.2




2007-04-04 14:52:21     Re: SPI eeprom write read

Mike Frysinger (UNITED STATES)

Message: 24062    nope, the steps you've listed are correct ...


in order to take one step out of the equation though, you should try using one of our tested releases:



we have a SPI LDR for u-boot there for the BF537-STAMP ... that should eliminate question of whether it's a software (compiling u-boot) or hardware (using correct settings)




2008-04-01 10:20:03     Re: SPI eeprom write read

Hadi Boulfelfel (FRANCE)

Message: 53456    I think it's :  #define BFIN_BOOT_MODE        BF537_BOOT_SPI_MASTER     not     BF537_SPI_MASTER_BOOT




2008-04-11 10:26:53     Re: SPI eeprom write read

Pafone Lee (CHINA)

Message: 54093    There is also a difference between the docs and the u-boot source code

in http://docs.blackfin.uclinux.org/doku.php?id=bootloaders:u-boot:release-notes:2007r1

<pre class="code">// ADSP-BF533

#define BF533_BYPASS_BOOT 0x0001

#define BF533_PARA_BOOT 0x0002

#define BF533_SPI_BOOT 0x0004 but in u-boot2008-R1-1.1.6  source code /include/configs/bf533-ezkit.h


I don't know whether it is a mistake.




2008-04-11 11:58:54     Re: SPI eeprom write read

Mike Frysinger (UNITED STATES)

Message: 54096    there is no difference nor error with the documentation


you're comparing old release notes against a newer release.  that'll never work.