RE: [U-Boot-Users] NAND driver issue

December 10, 2005 Aubrey wrote:
On 12/10/05, Dave Ellis DGE@sixnetio.com wrote:
I think the ECC code used to be broken unless
CONFIG_MTD_NAND_ECC_JFFS2
is set in your board config file.
The macro CONFIG_MTD_NAND_ECC_JFFS2 is there black and white in the file "common/cmd_nand.c". Should I add it again to my board config file?
It used to be commented out. I didn't realize it was fixed a long time ago since somehow the fix didn't make it into my local copy. Sorry for the noise.
On the other hand, If I don't want to work on the JFFS2 system, the macro should be disabled. I mean this macro shouldn't affect the direct read/write operation of the nand driver.
It was broken when I worked on it (several years ago). The non-JFFS2 implementation conflicted with the bad block position set by the NAND manufacturers, so I just made the JFFS2 option work. Richard Woodruff fixed the conflict, but I don't think the present code will use the ECC for non-JFFS2 since eccvalid_pos is -1.
I guess nobody cares, otherwise CONFIG_MTD_NAND_ECC_JFFS2 would be defined in the config files, not in cmd_nand.c.
Anyhow, the ST NAND128W3 worked for me (with JFFS2 ECC) when I added the ids to the header files.
Dave Ellis

In message C8F551D39A7C724E987CF8DD11D317830C2EB2@svr3.sixnetio.com you wrote:
The macro CONFIG_MTD_NAND_ECC_JFFS2 is there black and white in the file "common/cmd_nand.c". Should I add it again to my board config file?
It used to be commented out. I didn't realize it was fixed a long time ago since somehow the fix didn't make it into my local copy. Sorry for the noise.
Will everybody who is doing any work on the NAND code please use the "testing-NAND" branch, please?
It makes little sense to work on obsolete code and ignore what is supposed to be the replacement.
Thanks.
Best regards,
Wolfgang Denk
participants (2)
-
Dave Ellis
-
Wolfgang Denk