Hi Love,

On 7/22/26 2:32 PM, Kumar, Love wrote:
Hi Bastien,

On 7/17/2026 8:21 PM, Bastien Curutchet wrote:
Hi Love

On 7/13/26 3:52 PM, Kumar, Love wrote:
Hi Bastien,

On 7/13/2026 3:18 PM, Bastien Curutchet wrote:
Hi Love,

On 7/11/26 8:36 AM, Kumar, Love wrote:
Hi Michal,

On 7/9/2026 11:23 AM, Michal Simek wrote:
Hi Love,

On 7/2/26 15:08, Bastien Curutchet wrote:
This driver is deeply broken in many ways. Miquèl has cleaned it up by:
- flagging the most hideous parts (where Micron on-die ECC support is
      hardcoded)
- enabling DT support, eg. for reading the ECC configuration
- fixing software ECC support
- aligning the software ECC OOB layout to the rest of the world and
      allow a working interoperability in Linux.

I then added a patch to fix writes when soft ECC is on.

Signed-off-by: Miquel Raynal (DAVE.eu) <[email protected]>
Signed-off-by: Bastien Curutchet (DAVE.eu) <[email protected]>
---
Changes in v3:
- Add patch 6 to fix write issues in soft ECC.
- Link to v2: 
https://lore.kernel.org/r/[email protected]

Changes in v2:
- Fix the sponsor: s/DAVE/DAVE.eu/, no other change.
- Link to v1: 
https://lore.kernel.org/r/[email protected]

---
Bastien Curutchet (DAVE.eu) (1):
          mtd: rawnand: nand_base: Don't overwrite read/write_page_raw() for 
soft ECC

Miquel Raynal (DAVE.eu) (5):
          mtd: rawnand: zynq: Align the layout structures with the driver 
namespace
          mtd: rawnand: zynq: Add comments about very bad practices
          mtd: rawnand: zynq: Make sure an of node is attached
          mtd: rawnand: zynq: Do not fail the probe for no reason
          mtd: rawnand: zynq: Fix Software ECC engine support

     drivers/mtd/nand/raw/nand_base.c |   6 +-
     drivers/mtd/nand/raw/zynq_nand.c | 115 
++++++++++-----------------------------
     2 files changed, 32 insertions(+), 89 deletions(-)
---
base-commit: 4433253ecf2041f9362a763bb6cb79960921ac7e
change-id: 20260522-dave-upstream-nand-fixes-60a5336517a2

Best regards,

Please test this series.

We are again seeing the NAND write issue with zynq-zc770-xm011 board.

Below is the log snippet:

Zynq> nand write 0x80000 0x0 0x126f80
nand write 0x80000 0x0 0x126f80
NAND write: device 0 offset 0x0, size 0x126f80
NAND write to offset 0 failed -5

The exact same command works in my setup:

```
Zynq> nand write 0x80000 0x0 0x126f80

NAND write: device 0 offset 0x0, size 0x126f80
   1208192 bytes written: OK
```

Can you please share your device-tree / defconfig ? I'd like to dig into this a 
bit.

I'd also be interested in seeing the output of 'mtd list' and 'nand info' on 
your side.

I have used the xilinx_zynq_virt_defconfig with zynq-zc770-xm011 DT.
Below are the log snippets for 'nand info' and 'mtd list' commands:
   Zynq> nand info
   Device 0: nand0, sector size 128 KiB
    Page size         2048 b
    OOB size            64 b
    Erase size      131072 b
    ecc strength         0 bits
    ecc step size     2048 b
    subpagesize       2048 b
    options       0x40000000
    bbt options   0x00028000
   Zynq> mtd list
List of MTD devices:
* nand0
    - type: NAND flash
    - block size: 0x20000 bytes
    - min I/O: 0x800 bytes
    - OOB size: 64 bytes
    - OOB available: 38 bytes
    - 0x000000000000-0x000008000000 : "nand0"
       - 0x000000000000-0x000001000000 : "nand-fsbl-uboot"
       - 0x000001000000-0x000003000000 : "nand-linux"
       - 0x000003000000-0x000003200000 : "nand-rootfs"
Zynq>


Thank for these information.

Have you explicitly enabled the soft ECC ? I don't find the nand-ecc-mode 
property in the upstream zynq-zc770-xm011 DT.


Yes, tested it by explicitly enabling it with `nand-ecc-mode = "soft";` in 
zynq-zc770-xm011 DT, observed the same behavior - NAND write fails. Same issue was 
observed without enabling ecc mode as well.

```
Zynq> nand write 0x80000 0x0 0x80000

NAND write: device 0 offset 0x0, size 0x80000
NAND write to offset 0 failed -5
0 bytes written: ERROR
Zynq>
```

Am I missing any other configurations?

I don't see any other configuration you'd be missing.

I had this error once with ECC-soft enabled but that was because I had forgotten to do a prior erase


Best regards,
Bastien

Reply via email to