MTD: OneNAND: Return only negative error codes

PATCH In review Adrian Hunter 3 participants 3 messages Oct 12, 2007 – Oct 14, 2007 Memory technology devices (MTD)

lore

Share & tools

lore

Adrian Hunter

author

2007-10-12 07:27 UTC #0 lore raw

PATCH 1/1 MTD: OneNAND: Return only negative error codes

The OneNAND driver was confusing JFFS2 by returning positive error codes.

Signed-off-by Adrian Hunter <[email protected]>

---
drivers/mtd/onenand/onenand_base.c | 4 ++--

1 files changed, 2 insertions(+), 2 deletions(-)

drivers/mtd/onenand/onenand_base.c +0 -0

index f633941..a8c426b 100644
@@ -327,7 +327,7 @@ static int onenand_wait(struct mtd_info *mtd, int state)

		printk(KERN_ERR "onenand_wait: controller error = 0x%04x\n", ctrl);
		if (ctrl & ONENAND_CTRL_LOCK)
			printk(KERN_ERR "onenand_wait: it's locked error.\n");
-		return ctrl;
+		return -EIO;
	}

	if (interrupt & ONENAND_INT_READ) {
@@ -336,7 +336,7 @@ static int onenand_wait(struct mtd_info *mtd, int state)
			if (ecc & ONENAND_ECC_2BIT_ALL) {
				printk(KERN_ERR "onenand_wait: ECC error = 0x%04x\n", ecc);
				mtd->ecc_stats.failed++;
-				return ecc;
+				return -EBADMSG;
			} else if (ecc & ONENAND_ECC_1BIT_ALL) {
				printk(KERN_INFO "onenand_wait: correctable ECC error = 0x%04x\n", ecc);
				mtd->ecc_stats.corrected++;

Signature · 1 line

1.4.4.2

Kyungmin Park

Acked-by on v1 1/1

2007-10-12 09:30 UTC #1 lore raw

Quoted 2 lines

The OneNAND driver was confusing JFFS2 by returning positive error codes.

It's only possible the write case error and it returns the positive 'ctrl' error. Other case, "2-bit ecc error' is handled with 'mtd->ecc_stat' correctly.

Well, no problem to change to the negative value.

Acked-by Kyungmin Park <[email protected]>

Quoted 29 lines

---
 drivers/mtd/onenand/onenand_base.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

drivers/mtd/onenand/onenand_base.c +2 -3

index f633941..a8c426b 100644
@@ -327,7 +327,7 @@ static int onenand_wait(struct mtd_info *mtd, int state)
 		printk(KERN_ERR "onenand_wait: controller error = 0x%04x\n", ctrl);
 		if (ctrl & ONENAND_CTRL_LOCK)
 			printk(KERN_ERR "onenand_wait: it's locked error.\n");
-		return ctrl;
+		return -EIO;
 	}
 
 	if (interrupt & ONENAND_INT_READ) {

@@ -336,7 +336,7 @@ static int onenand_wait(struct mtd_info *mtd, int state)
 			if (ecc & ONENAND_ECC_2BIT_ALL) {
 				printk(KERN_ERR "onenand_wait: ECC error = 0x%04x\n", ecc);
 				mtd->ecc_stats.failed++;
-				return ecc;
+				return -EBADMSG;
 			} else if (ecc & ONENAND_ECC_1BIT_ALL) {
 				printk(KERN_INFO "onenand_wait: correctable ECC error = 0x%04x\n", ecc);
 				mtd->ecc_stats.corrected++;
--

1.4.4.2

Artem Bityutskiy

2007-10-14 08:17 UTC #2 lore raw

Quoted 9 lines · Kyungmin Park

Quoted 2 lines
The OneNAND driver was confusing JFFS2 by returning positive error codes.
It's only possible the write case error and it returns the positive 'ctrl' error. Other case, "2-bit ecc error' is handled with 'mtd->ecc_stat' correctly.
Well, no problem to change to the negative value.
Acked-by Kyungmin Park <[email protected]>

Yeah, it's rare but we caught this in practice, and JFFS2 just calls BUG() if OneNAND returns positive number while JFFS2 is scanning the flash medium.

Signature · 2 lines

Best regards,
Artem Bityutskiy (Битюцкий Артём)