ARM: omap2: mtd split nand_scan in ident and tail

PATCH Superseded Jan Weitzel 3 participants 7 messages Apr 14, 2011 – Apr 20, 2011

Subsystems (2)

Memory technology devices (MTD) NAND flash

lore

Share & tools

lore

Jan Weitzel

author

2011-04-14 09:15 UTC #0 lore raw

PATCH 1/1 ARM: omap2: mtd split nand_scan in ident and tail

nand_scan calls nand_scan_ident and nand_scan_tail, setting values like oobvail according to ecc.layout. If we change the layout afterwards values are wrong.

Signed-off-by Jan Weitzel <[email protected]>

---
 drivers/mtd/nand/omap2.c |   10 ++++++++--
 1 files changed, 8 insertions(+), 2 deletions(-)

drivers/mtd/nand/omap2.c +8 -2

index da9a351..288423f 100644
@@ -1073,9 +1073,9 @@ static int __devinit omap_nand_probe(struct platform_device *pdev)
 	/* DIP switches on some boards change between 8 and 16 bit
 	 * bus widths for flash.  Try the other width if the first try fails.
 	 */
-	if (nand_scan(&info->mtd, 1)) {
+	if (nand_scan_ident(&info->mtd, 1, NULL)) {
 		info->nand.options ^= NAND_BUSWIDTH_16;
-		if (nand_scan(&info->mtd, 1)) {
+		if (nand_scan_ident(&info->mtd, 1, NULL)) {
 			err = -ENXIO;
 			goto out_release_mem_region;
 		}

@@ -1101,6 +1101,12 @@ static int __devinit omap_nand_probe(struct platform_device *pdev)
 		info->nand.ecc.layout = &omap_oobinfo;
 	}
 
+	/* second phase scan */
+	if (nand_scan_tail(&info->mtd)) {
+		err = -ENXIO;
+		goto out_release_mem_region;
+	}
+
 #ifdef CONFIG_MTD_PARTITIONS
 	err = parse_mtd_partitions(&info->mtd, part_probes, &info->parts, 0);
 	if (err > 0)

Signature · 1 line

1.7.0.4

Jan Weitzel

author

2011-04-15 13:35 UTC #1 lore raw

Am Donnerstag, den 14.04.2011, 11:15 +0200 schrieb Jan Weitzel:

Quoted 27 lines

nand_scan calls nand_scan_ident and nand_scan_tail, setting values like oobvail according to ecc.layout. If we change the layout afterwards values are wrong.

Signed-off-by Jan Weitzel <[email protected]>

---
 drivers/mtd/nand/omap2.c |   10 ++++++++--
 1 files changed, 8 insertions(+), 2 deletions(-)

drivers/mtd/nand/omap2.c +2 -2

index da9a351..288423f 100644
@@ -1073,9 +1073,9 @@ static int __devinit omap_nand_probe(struct platform_device *pdev)
 	/* DIP switches on some boards change between 8 and 16 bit
 	 * bus widths for flash.  Try the other width if the first try fails.
 	 */
-	if (nand_scan(&info->mtd, 1)) {
+	if (nand_scan_ident(&info->mtd, 1, NULL)) {
 		info->nand.options ^= NAND_BUSWIDTH_16;
-		if (nand_scan(&info->mtd, 1)) {
+		if (nand_scan_ident(&info->mtd, 1, NULL)) {
 			err = -ENXIO;
 			goto out_release_mem_region;
 		}

@@ -1101,6 +1101,12 @@ static int __devinit omap_nand_probe(struct platform_device *pdev)
 		info->nand.ecc.layout = &omap_oobinfo;
 	}

drivers/mtd/nand/omap2.c static int __devinit omap_nand_probe(struct platform_device *pdev) · line 1112

Quoted 10 lines

drivers/mtd/nand/omap2.c +6 -0

 
+	/* second phase scan */
+	if (nand_scan_tail(&info->mtd)) {
+		err = -ENXIO;
+		goto out_release_mem_region;
+	}
+
 #ifdef CONFIG_MTD_PARTITIONS
 	err = parse_mtd_partitions(&info->mtd, part_probes, &info->parts, 0);
 	if (err > 0)

So no comments? Is any rework needed? Without this patch I am not able to mount partions with OMAP_ECC_HAMMING_CODE_HW_ROMCODE. Regards, Jan

Artem Bityutskiy

2011-04-15 14:06 UTC #2 lore raw

Quoted 29 lines · Jan Weitzel

Am Donnerstag, den 14.04.2011, 11:15 +0200 schrieb Jan Weitzel:

Quoted 28 lines

nand_scan calls nand_scan_ident and nand_scan_tail, setting values like oobvail according to ecc.layout. If we change the layout afterwards values are wrong.

Signed-off-by Jan Weitzel <[email protected]>

---
 drivers/mtd/nand/omap2.c |   10 ++++++++--
 1 files changed, 8 insertions(+), 2 deletions(-)

drivers/mtd/nand/omap2.c +2 -2

index da9a351..288423f 100644
@@ -1073,9 +1073,9 @@ static int __devinit omap_nand_probe(struct platform_device *pdev)
 	/* DIP switches on some boards change between 8 and 16 bit
 	 * bus widths for flash.  Try the other width if the first try fails.
 	 */
-	if (nand_scan(&info->mtd, 1)) {
+	if (nand_scan_ident(&info->mtd, 1, NULL)) {
 		info->nand.options ^= NAND_BUSWIDTH_16;
-		if (nand_scan(&info->mtd, 1)) {
+		if (nand_scan_ident(&info->mtd, 1, NULL)) {
 			err = -ENXIO;
 			goto out_release_mem_region;
 		}

@@ -1101,6 +1101,12 @@ static int __devinit omap_nand_probe(struct platform_device *pdev)
 		info->nand.ecc.layout = &omap_oobinfo;
 	}

Quoted 12 lines

drivers/mtd/nand/omap2.c static int __devinit omap_nand_probe(struct platform_device *pdev) · line 1112
Quoted 9 lines
drivers/mtd/nand/omap2.c +6 -0
+	/* second phase scan */
+	if (nand_scan_tail(&info->mtd)) {
+		err = -ENXIO;
+		goto out_release_mem_region;
+	}
+
 #ifdef CONFIG_MTD_PARTITIONS
 	err = parse_mtd_partitions(&info->mtd, part_probes, &info->parts, 0);
 	if (err > 0)
So no comments? Is any rework needed? Without this patch I am not able to mount partions with OMAP_ECC_HAMMING_CODE_HW_ROMCODE.

Sorry, I missed this patch. Could you please update the commit message and make it more verbose and state clearly:

1. which problem you solve
2. how you solve it
3. why is this the right solution

Thanks!

Signature · 2 lines

Best Regards,
Artem Bityutskiy (Артём Битюцкий)

Jan Weitzel

author

2011-04-18 06:50 UTC #3 lore raw

PATCH v2 1/1 ARM: omap2: mtd split nand_scan in ident and tail

nand_scan calls nand_scan_tail and here we got a ecc.layout and calculate oobavail for this layout. After calling nand_scan, we change the layout pointer if OMAP_ECC_HAMMING_CODE_HW_ROMCODE is set. This results in not calcluated oobavail. Mountig as jffs2 is not possible.

To fix that nand_scan has to split up in nand_scan_ident and nand_scan_tail setting ecc.layout between these calls. So nand_scan_tail calculates oobvail for the used layout. This is also done in serveral other platforms.

v2:
update commit message

Signed-off-by Jan Weitzel <[email protected]>

---
 drivers/mtd/nand/omap2.c |   10 ++++++++--
 1 files changed, 8 insertions(+), 2 deletions(-)

drivers/mtd/nand/omap2.c +8 -2

index da9a351..288423f 100644
@@ -1073,9 +1073,9 @@ static int __devinit omap_nand_probe(struct platform_device *pdev)
 	/* DIP switches on some boards change between 8 and 16 bit
 	 * bus widths for flash.  Try the other width if the first try fails.
 	 */
-	if (nand_scan(&info->mtd, 1)) {
+	if (nand_scan_ident(&info->mtd, 1, NULL)) {
 		info->nand.options ^= NAND_BUSWIDTH_16;
-		if (nand_scan(&info->mtd, 1)) {
+		if (nand_scan_ident(&info->mtd, 1, NULL)) {
 			err = -ENXIO;
 			goto out_release_mem_region;
 		}

@@ -1101,6 +1101,12 @@ static int __devinit omap_nand_probe(struct platform_device *pdev)
 		info->nand.ecc.layout = &omap_oobinfo;
 	}
 
+	/* second phase scan */
+	if (nand_scan_tail(&info->mtd)) {
+		err = -ENXIO;
+		goto out_release_mem_region;
+	}
+
 #ifdef CONFIG_MTD_PARTITIONS
 	err = parse_mtd_partitions(&info->mtd, part_probes, &info->parts, 0);
 	if (err > 0)

Signature · 1 line

1.7.0.4

Vimal Singh

Reviewed-by on v2 1/1

2011-04-18 07:00 UTC #4 lore raw

Quoted 11 lines · Jan Weitzel

nand_scan calls nand_scan_tail and here we got a ecc.layout and calculate oobavail for this layout. After calling nand_scan, we change the layout pointer if OMAP_ECC_HAMMING_CODE_HW_ROMCODE is set. This results in not calcluated oobavail. Mountig as jffs2 is not possible.
To fix that nand_scan has to split up in nand_scan_ident and nand_scan_tail setting ecc.layout between these calls. So nand_scan_tail calculates oobvail for the used layout. This is also done in serveral other platforms.
v2:
update commit message

Move above two lines to below (as there are not part of commit log):

Quoted 3 lines

Signed-off-by Jan Weitzel <[email protected]>
---

here.

Feel free to add:

Reviewed-by Vimal Singh <[email protected]>

Regards,
Vimal

Jan Weitzel

author

2011-04-19 14:15 UTC #5 lore raw

PATCH v3 1/1 ARM: omap2: mtd split nand_scan in ident and tail

Signed-off-by Jan Weitzel <[email protected]>

Reviewed-by Vimal Singh <[email protected]>

---
v2:
update commit message
v3:
Add Reviewed-by

 drivers/mtd/nand/omap2.c |   10 ++++++++--
 1 files changed, 8 insertions(+), 2 deletions(-)

drivers/mtd/nand/omap2.c +8 -2

index da9a351..288423f 100644
@@ -1073,9 +1073,9 @@ static int __devinit omap_nand_probe(struct platform_device *pdev)
 	/* DIP switches on some boards change between 8 and 16 bit
 	 * bus widths for flash.  Try the other width if the first try fails.
 	 */
-	if (nand_scan(&info->mtd, 1)) {
+	if (nand_scan_ident(&info->mtd, 1, NULL)) {
 		info->nand.options ^= NAND_BUSWIDTH_16;
-		if (nand_scan(&info->mtd, 1)) {
+		if (nand_scan_ident(&info->mtd, 1, NULL)) {
 			err = -ENXIO;
 			goto out_release_mem_region;
 		}

@@ -1101,6 +1101,12 @@ static int __devinit omap_nand_probe(struct platform_device *pdev)
 		info->nand.ecc.layout = &omap_oobinfo;
 	}
 
+	/* second phase scan */
+	if (nand_scan_tail(&info->mtd)) {
+		err = -ENXIO;
+		goto out_release_mem_region;
+	}
+
 #ifdef CONFIG_MTD_PARTITIONS
 	err = parse_mtd_partitions(&info->mtd, part_probes, &info->parts, 0);
 	if (err > 0)

Signature · 1 line

1.7.0.4

Artem Bityutskiy

2011-04-20 09:00 UTC #6 lore raw

PATCH v3 1/1 ARM: omap2: mtd split nand_scan in ident and tail

Quoted 11 lines · Jan Weitzel

nand_scan calls nand_scan_tail and here we got a ecc.layout and calculate oobavail for this layout. After calling nand_scan, we change the layout pointer if OMAP_ECC_HAMMING_CODE_HW_ROMCODE is set. This results in not calcluated oobavail. Mountig as jffs2 is not possible.
To fix that nand_scan has to split up in nand_scan_ident and nand_scan_tail setting ecc.layout between these calls. So nand_scan_tail calculates oobvail for the used layout. This is also done in serveral other platforms.
Signed-off-by Jan Weitzel <[email protected]>
Reviewed-by Vimal Singh <[email protected]>

Jan,

thanks for the patch, pushed to l2-mtd-2.6.git. And apologies that I missed.

Signature · 2 lines

Best Regards,
Artem Bityutskiy (????? ????????)