nand oob corruption recovery

nand oob corruption recovery

[David Updegraff]

I realize that specs say that if you wipe out mfg. bad block markers 
that there is "no way to recover them".. but...

Are there suggestions for plausibly reliable test sequences that one 
could run a nand chip thru to regenerate valid bad block markers?

Presumeably, it is insufficient to simply try to erase all blocks, 
marking failures as bad...  perhaps verified full page writes ?

Suggestions?

Thanks.

-dbu / Dave Updegraff / dave@cray.com

Re: nand oob corruption recovery

[Thomas Gleixner]

On Thursday 01 July 2004 19:24, David Updegraff wrote:
> I realize that specs say that if you wipe out mfg. bad block markers
> that there is "no way to recover them".. but...
> Are there suggestions for plausibly reliable test sequences that one
> could run a nand chip thru to regenerate valid bad block markers?
>
> Presumeably, it is insufficient to simply try to erase all blocks,
> marking failures as bad...  perhaps verified full page writes ?

Erasing may be not sufficient, as the block could be al ff. Writing all blocks 
to all zeros and verify that it works. Maybe try a 0x55 and 0xAA pattern too.

-- 
Thomas
_____________________________________________________________________
From slash dot org
"When customers are visiting, engineers are not allowed to wear ties. 
That way the customer can tell who is the engineer and who is the 
salesman (and therefore whom to believe.). Ties cut off blood flow 
to the brain, making it easier for the salesmen to do their jobs." 
_____________________________________________________________________
linutronix - competence in embedded & realtime linux
http://www.linutronix.de
mail: tglx@linutronix.de

Re: nand oob corruption recovery

[David Updegraff]

[-- Attachment #1: Type: text/plain, Size: 158 bytes --]

Relevant to this; enclosed is small nandwrite.c patch that fixes 
segfault and at least warns about upcomming OOB destruction in 
large-block devices.

-dbu.

[-- Attachment #2: nandwrite.patch --]
[-- Type: text/x-patch, Size: 974 bytes --]

--- mtd/util/nandwrite.c.orig	2004-07-01 17:09:12.124117544 -0500
+++ mtd/util/nandwrite.c	2004-07-01 17:16:44.497346360 -0500
@@ -42,8 +42,8 @@
 /*
  * Buffer array used for writing data
  */
-unsigned char writebuf[512];
-unsigned char oobbuf[16];
+unsigned char *writebuf;
+unsigned char *oobbuf;
 
 // oob layouts to pass into the kernel as default
 struct nand_oobinfo none_oobinfo = { 
@@ -223,6 +223,19 @@
 		close(fd);
 		exit(1);
 	}
+
+	/* NOW! allocate writebuf and oobbuf.. */
+	writebuf = malloc (meminfo.oobblock);
+	oobbuf = malloc (meminfo.oobsize);
+	if (!writebuf || !oobbuf)
+		perror ("cannot allocate buffers\n");
+	if ((forcejffs2 || forceyaffs || autoplace) && (meminfo.oobsize > 16))
+	{
+		fprintf (stderr, "\nDanger!Danger! Will Robinson, Overriding driver's AUTO_ECC layout for \nlarge-block chips will almost certainly trash OOB!\n Continue? (y/N)\n");
+		if (getchar () != 'y') {
+			exit(1);
+		}
+	}
 	
 	// write without ecc ?
 	if (noecc) {