CFI and different Command Definition Tables

Altera_Forum · ‎07-18-2006

I am having problems using Quartus Flash Programmer to erase/program CFI flash made by Atmel.

I have a single 64 megabit chip at49bv6416 on my board hooked up to Nios II using Avalon bus and it uses 16-bit data bus.

I know it has a different Command Definition Table than a similar 64 megabit, also CFI compatible chip made by AMD (am29lv640mh).

For example, "Program Word" command for Atmel chip is following sequence:

555/aa

aaa/55

555/a0

for AMD chip the sequence is:

555/aa

2aa/55

555/a0

how the quartus flash programmer "knows" what sequence to use for my chip ?

Command values does not seem to be coded in CFI table...

Do I configure it in some quartus/project file to match my flash chip?

Altera_Forum · ‎07-18-2006

Have you looked at the flash programmer user's guide (http://www.altera.com/literature/ug/ug_nios2_flash_programmer.pdf).

You may be able to define an override to get your CFI "compliant" flash to function.

- slacker

Altera_Forum · ‎07-19-2006

Yes, I have looked in this file and I learned something new: I can use Flash Programmer in command-line mode to display some more debug messages.

Here is my output from the program for my board:

Using cable "USB-Blaster ", device 1, instance 0x01
Resetting and pausing target processor: OK
Reading System ID at address 0x00900BC8: verified
Found CFI table in 16 bit mode
Raw CFI query table read from device:
   0: 1F 00 D6 00 FF FF FF FF  FF FF FF FF FF FF FF FF  ................
  10: FF FF FF FF FF FF FF FF  FF FF FF FF FF FF FF FF  ................
  20: 51 00 52 00 59 00 02 00  00 00 41 00 00 00 00 00  Q.R.Y.....A.....
  30: 00 00 00 00 00 00 27 00  36 00 09 00 0A 00 04 00  ......&#39;.6.......
  40: 00 00 09 00 10 00 04 00  00 00 03 00 03 00 17 00  ................
CFI query table read from device:
  10: 51 52 59 02 00 41 00 00  00 00 00 27 36 09 0A 04  QRY..A.....&#39;6...
  20: 00 09 10 04 00 03 03 17  01 00 00 00 02 7E 00 00  .............~..
  30: 01 07 00 20 00 FF FF FF  FF FF FF FF FF FF FF FF  ... ............
  40: FF                                                .
CFI extended table read from device:
   0: 50 52 49 31 30 AF 01 00  01 80 03 03 FF FF FF FF  PRI10...........
  10: FF FF FF FF FF FF FF FF  FF FF FF FF FF FF FF FF  ................
Read autoselect code 001F-00D6 (in 16 bit mode)
No CFI override data for 
Device size is 8MByte
Erase regions are:
  offset        0: 127 x 64K
  offset   7F0000:   8 x 8K
Device supports AMD style programming algorithm
Multi-byte programming not supported
Sector erase timeout is 4s
Word program timeout is 256us
              : Checksumming existing contents
00200000      : Checksum failed - needs erase then program
00210000      : Checksum failed - needs erase then program
00220000      : Checksum failed - needs erase then program
00230000      : Checksum failed - needs erase then program
00200000      : Reading existing contents
00210000      : Reading existing contents
00220000      : Reading existing contents
00230000      : Reading existing contents
Checksummed/read 14kB in 0.4s
00200000 ( 0%): Erasing
00210000 (25%): Erasing
00220000 (50%): Erasing
00230000 (75%): Erasing
Erased 256kB in 0.0s
00200000 ( 0%): Programming
Program sequence:
  Write address 0x00000000 <= 0xF0
  Write address 0x0000AAAA <= 0xAA
  Write address 0x00005554 <= 0x55
  Write address 0x0000AAAA <= 0xA0
  Write address 0x00200010 <= 0xD76A
  Read  address 0x00200010 => 0xE4
Program failed at offset 200010
Leaving target processor paused

Doing this, I can see the wrong program sequence is used, but I am not sure how to change it to the correctl one. Yes, I can override CFI table using --override=<filename> parameters, but how do I find numbers for CFI table to override?

Any ideas ?

Altera_Forum · ‎07-19-2006

Get out your datasheet and compare the read CFI table values with what should be there. Use the override table to "replace" these bad values for (in your case) FLASH-001F-00D6.

You might be able to find some examples of doing this, on this forum, or, if not, definitely on Altera's website.

Best of luck and good job using the --debug option. I'm sorry I didn't recommend it to you, in the first place.

Cheers,

- slacker