MXS-NAND: Backport ecc.size from linux kernel

PATCH New Marek Vasut 3 participants 11 messages Dec 18, 2011 – Jan 4, 2012

Subsystems (3)

Arm i.MX and Layerscape Memory technology devices (MTD) NAND flash

lore

Share & tools

lore

Marek Vasut

author

2011-12-18 17:50 UTC #0 lore raw

PATCH 1/1 MXS-NAND: Backport ecc.size from linux kernel

The ecc.size for mxs NAND driver is set to 1 in Linux kernel and to 512 in U-Boot, which causes "ubi part" command malfunction due to wrong subpage size.

Signed-off-by Marek Vasut <[email protected]>

Cc Wolfgang Denk <[email protected]>

Cc Detlev Zundel <[email protected]>

Cc Stefan Roese <[email protected]>

Cc Scott Wood <[email protected]>

Cc Veli-Pekka Peltola <[email protected]>

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

drivers/mtd/nand/mxs_nand.c +1 -1

index ce2a326..26778ac 100644
@@ -1105,7 +1105,7 @@ int board_nand_init(struct nand_chip *nand)
 	nand->ecc.layout	= &fake_ecc_layout;
 	nand->ecc.mode		= NAND_ECC_HW;
 	nand->ecc.bytes		= 9;
-	nand->ecc.size		= 512;
+	nand->ecc.size		= 1;
 
 	return 0;

Signature · 1 line

1.7.7.3

Veli-Pekka Peltola

2011-12-19 14:33 UTC #1 lore raw

PATCH 1/1 MXS-NAND: Backport ecc.size from linux kernel

Hi Marek,

Quoted 2 lines · Marek Vasut

The ecc.size for mxs NAND driver is set to 1 in Linux kernel and to 512 in U-Boot, which causes "ubi part" command malfunction due to wrong subpage size.

[snip]

Subpage size is now reported correctly and ubifs works fine. Thanks!

Tested-by Veli-Pekka Peltola <[email protected]>

--
Veli-Pekka Peltola

Marek Vasut

author

2011-12-19 15:05 UTC #2 lore raw

PATCH 1/1 MXS-NAND: Backport ecc.size from linux kernel

Quoted 7 lines

Hi Marek,
Quoted 3 lines · Marek Vasut
The ecc.size for mxs NAND driver is set to 1 in Linux kernel and to 512 in U-Boot, which causes "ubi part" command malfunction due to wrong subpage size.

Quoted 5 lines

[snip]
Subpage size is now reported correctly and ubifs works fine. Thanks!
Tested-by Veli-Pekka Peltola <[email protected]>

You're the guy doing the bluegiga mx28 board ? I think I noticed in linux kernel, are you for example working on mx28/device tree support too?

Veli-Pekka Peltola

2011-12-19 15:30 UTC #3 lore raw

PATCH 1/1 MXS-NAND: Backport ecc.size from linux kernel

Quoted 2 lines · Marek Vasut

You're the guy doing the bluegiga mx28 board ? I think I noticed in linux kernel, are you for example working on mx28/device tree support too?

Yes, I have done something but most of kernel stuff related to our board is done by my colleague Lauri Hintsala. Currently device tree support is not in our task list. I don't know if someone is working on it. Shawn Guo (kernel maintainer for MX28) might know.

--
Veli-Pekka Peltola

Scott Wood

2011-12-19 20:19 UTC #4 lore raw

PATCH 1/1 MXS-NAND: Backport ecc.size from linux kernel

In reply to Marek Vasut

Quoted 18 lines · Marek Vasut

The ecc.size for mxs NAND driver is set to 1 in Linux kernel and to 512 in U-Boot, which causes "ubi part" command malfunction due to wrong subpage size.
Signed-off-by Marek Vasut <[email protected]>
Cc Wolfgang Denk <[email protected]>
Cc Detlev Zundel <[email protected]>
Cc Stefan Roese <[email protected]>
Cc Scott Wood <[email protected]>
Cc Veli-Pekka Peltola <[email protected]>
---
 drivers/mtd/nand/mxs_nand.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/mtd/nand/mxs_nand.c +0 -0
index ce2a326..26778ac 100644

drivers/mtd/nand/mxs_nand.c int board_nand_init(struct nand_chip *nand) · line 1111

Quoted 8 lines

drivers/mtd/nand/mxs_nand.c +1 -1

@@ -1105,7 +1105,7 @@ int board_nand_init(struct nand_chip *nand)
 	nand->ecc.layout	= &fake_ecc_layout;
 	nand->ecc.mode		= NAND_ECC_HW;
 	nand->ecc.bytes		= 9;
-	nand->ecc.size		= 512;
+	nand->ecc.size		= 1;
 
 	return 0;

ecc.size = 1 doesn't make sense -- this is the block size over which ecc is calculated.

Where is this Linux driver? I don't see mxs_nand.c in Linux.

What specifically is happening in "ubi part" with ecc.size = 512?

-Scott

Marek Vasut

author

2011-12-19 20:23 UTC #5 lore raw

PATCH 1/1 MXS-NAND: Backport ecc.size from linux kernel

Quoted 26 lines · Marek Vasut

Quoted 25 lines
The ecc.size for mxs NAND driver is set to 1 in Linux kernel and to 512 in U-Boot, which causes "ubi part" command malfunction due to wrong subpage size.
Signed-off-by Marek Vasut <[email protected]>
Cc Wolfgang Denk <[email protected]>
Cc Detlev Zundel <[email protected]>
Cc Stefan Roese <[email protected]>
Cc Scott Wood <[email protected]>
Cc Veli-Pekka Peltola <[email protected]>
---
 drivers/mtd/nand/mxs_nand.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/mtd/nand/mxs_nand.c +0 -0
index ce2a326..26778ac 100644
@@ -1105,7 +1105,7 @@ int board_nand_init(struct nand_chip *nand)
 
 	nand->ecc.layout	= &fake_ecc_layout;
 	nand->ecc.mode		= NAND_ECC_HW;
 	nand->ecc.bytes		= 9;
 

Quoted 7 lines

Quoted 4 lines
drivers/mtd/nand/mxs_nand.c +1 -1
-	nand->ecc.size		= 512;
+	nand->ecc.size		= 1;
 
return 0;
ecc.size = 1 doesn't make sense -- this is the block size over which ecc is calculated.

And in a way it forces the subpage shift to be 0 ... we need to fix it in both later, but for this release, let's make it this way (in sync with linux).

Quoted 2 lines

Where is this Linux driver? I don't see mxs_nand.c in Linux.

drivers/mtd/nand/gpmi-nand/*

Quoted 2 lines

What specifically is happening in "ubi part" with ecc.size = 512?

The driver doesn\t support subpage writes.

Quoted 2 lines

-Scott

Scott Wood

2011-12-19 20:34 UTC #6 lore raw

PATCH 1/1 MXS-NAND: Backport ecc.size from linux kernel

Quoted 4 lines · Marek Vasut

Quoted 2 lines · Marek Vasut
What specifically is happening in "ubi part" with ecc.size = 512?
The driver doesn\t support subpage writes.

Is this meant to be a workaround for the NAND layer ignoring a driver setting NAND_NO_SUBPAGE_WRITE? I'd rather just fix that...

-Scott

Marek Vasut

author

2011-12-19 20:56 UTC #7 lore raw

PATCH 1/1 MXS-NAND: Backport ecc.size from linux kernel

Quoted 8 lines · Marek Vasut

Quoted 4 lines
Quoted 2 lines · Marek Vasut
What specifically is happening in "ubi part" with ecc.size = 512?
The driver doesn\t support subpage writes.
Is this meant to be a workaround for the NAND layer ignoring a driver setting NAND_NO_SUBPAGE_WRITE? I'd rather just fix that...

Well it's like this in linux, it took me some time to find this out. I backported it, so probably.
M

Marek Vasut

author

2012-01-03 21:20 UTC #8 lore raw

PATCH 1/1 MXS-NAND: Backport ecc.size from linux kernel

In reply to Scott Wood

Quoted 10 lines · Marek Vasut

Quoted 4 lines
Quoted 2 lines · Marek Vasut
What specifically is happening in "ubi part" with ecc.size = 512?
The driver doesn\t support subpage writes.
Is this meant to be a workaround for the NAND layer ignoring a driver setting NAND_NO_SUBPAGE_WRITE? I'd rather just fix that...
-Scott

Any news here ?

Scott Wood

2012-01-03 21:30 UTC #9 lore raw

PATCH 1/1 MXS-NAND: Backport ecc.size from linux kernel

Quoted 4 lines · Marek Vasut

Quoted 4 lines · Marek Vasut
Quoted 3 lines
Quoted 2 lines · Marek Vasut
What specifically is happening in "ubi part" with ecc.size = 512?

Quoted 8 lines

Quoted 6 lines
Quoted 1 line
The driver doesn\t support subpage writes.
Is this meant to be a workaround for the NAND layer ignoring a driver setting NAND_NO_SUBPAGE_WRITE? I'd rather just fix that...
-Scott
Any news here ?

I just got back from end-of-year vacation; I'll look at it soon (will probably just remove the mask and trust the driver to not set flags that don't make sense). Thanks for the reminder.

-Scott

Marek Vasut

author

2012-01-04 00:39 UTC #10 lore raw

PATCH 1/1 MXS-NAND: Backport ecc.size from linux kernel

Quoted 7 lines · Marek Vasut

Quoted 6 lines
Quoted 6 lines · Marek Vasut
Quoted 4 lines
Quoted 2 lines · Marek Vasut
What specifically is happening in "ubi part" with ecc.size = 512?
The driver doesn\t support subpage writes.

Quoted 8 lines

Quoted 6 lines
Quoted 4 lines
Is this meant to be a workaround for the NAND layer ignoring a driver setting NAND_NO_SUBPAGE_WRITE? I'd rather just fix that...
-Scott
Any news here ?
I just got back from end-of-year vacation;

I hope you had good time :)

Quoted 3 lines

I'll look at it soon (will
probably just remove the mask and trust the driver to not set flags that don't make sense). Thanks for the reminder.

Naw, it should be fixed in linux too! We should probably start a discussion there.

Quoted 2 lines

-Scott