[PATCH 00/12] Marvell NAND controller rework with ->exec_op()

On Wed, 3 Jan 2018 21:10:28 +0100
Boris Brezillon [off-list ref] wrote:

On Wed, 03 Jan 2018 20:58:29 +0100
Robert Jarzmik [off-list ref] wrote:

quoted

Miquel RAYNAL [off-list ref] writes:
  

quoted

On Tue, 02 Jan 2018 20:21:09 +0100
Robert Jarzmik [off-list ref] wrote:
   

quoted

Miquel RAYNAL [off-list ref] writes:
    

quoted

I think the ECC issue you faced was related to pages being written
*and* empty. If this guess is right, the board should boot fine with
these changes.

Otherwise, please add the DEBUG define as before in both the core
and the driver and do not hesitate to add another dump_stack()
where it crashes (if applicable).      

The problem looks still the same :
[    3.560163] Bad block table not found for chip 0    

Mmmmh ok.

Can you please add this patch:
http://code.bulix.org/61at9p-254626    

Well, it looks a bit better, see attached log in [1].
Now the BBT is detected ...
[    3.310841] Bad block table found at page 131008, version 0x01
...
[    3.354944] Bad block table found at page 130944, version 0x01

But all blocks are considered bad ... as if the bit logic was inverted for the
meaning of "bad" or "good" block, see :
[    3.379825] nand_read_bbt: bad block at 0x000000000000  

Hm, that's weird. Can you try with the old driver (pxa3xx)?

Alternatively, you can type 'nand bad' from uboot to check if it
detects the same bad blocks.