* [U-Boot] [PATCH v2 0/5] ARM: at91: atmel_nand: check ONFI ecc minimum requirement
@ 2013-06-17 10:51 Josh Wu
2013-06-17 10:51 ` [U-Boot] [PATCH v2 1/5] ARM: at91: atmel_nand: pmecc driver will select the galois table by sector size Josh Wu
` (4 more replies)
0 siblings, 5 replies; 9+ messages in thread
From: Josh Wu @ 2013-06-17 10:51 UTC (permalink / raw)
To: u-boot
In this patch series, first we make the nand driver can dynamic change sector
size and ecc correct bits.
Then we enable ONFI ecc parameters check.
v1 --> v2: Fixes according to Scoot Wood's suggestion
- add a new patch which make all driver can use dev_err, dev_info, dev_dbg.
- replace all printk to dev_err or dev_info.
- fix pmecc data allocate error handler.
- fix the pmecc ecc requirement selection, should use <= instead of use <.
Josh Wu (5):
ARM: at91: atmel_nand: pmecc driver will select the galois table by
sector size
ARM: at91: sama5d3: remove unused definition about PMECC alpha table
offset
linux/compat.h: move dev_err, dev_info and dev_dbg from usb driver to
compat.h
mtd: atmel_nand: alloc memory instead of use static array for pmecc
data
ARM: at91: atmel_nand: add code to check the ONFI parameter ECC
requirement
arch/arm/include/asm/arch-at91/at91sam9x5.h | 6 +
arch/arm/include/asm/arch-at91/sama5d3.h | 2 -
doc/README.atmel_pmecc | 14 --
drivers/mtd/nand/atmel_nand.c | 197 +++++++++++++++++++++++++--
drivers/usb/musb-new/linux-compat.h | 16 ---
include/configs/at91sam9x5ek.h | 1 -
include/configs/sama5d3xek.h | 1 -
include/linux/compat.h | 8 ++
8 files changed, 201 insertions(+), 44 deletions(-)
--
1.7.9.5
^ permalink raw reply [flat|nested] 9+ messages in thread
* [U-Boot] [PATCH v2 1/5] ARM: at91: atmel_nand: pmecc driver will select the galois table by sector size
2013-06-17 10:51 [U-Boot] [PATCH v2 0/5] ARM: at91: atmel_nand: check ONFI ecc minimum requirement Josh Wu
@ 2013-06-17 10:51 ` Josh Wu
2013-06-17 10:51 ` [U-Boot] [PATCH v2 2/5] ARM: at91: sama5d3: remove unused definition about PMECC alpha table offset Josh Wu
` (3 subsequent siblings)
4 siblings, 0 replies; 9+ messages in thread
From: Josh Wu @ 2013-06-17 10:51 UTC (permalink / raw)
To: u-boot
Define the galois index table offset in chip head file. So user do not need
to set by himself. Driver will set it correctly according to sector_size.
Signed-off-by: Josh Wu <josh.wu@atmel.com>
---
arch/arm/include/asm/arch-at91/at91sam9x5.h | 6 ++++++
doc/README.atmel_pmecc | 14 --------------
drivers/mtd/nand/atmel_nand.c | 5 ++++-
include/configs/at91sam9x5ek.h | 1 -
include/configs/sama5d3xek.h | 1 -
5 files changed, 10 insertions(+), 17 deletions(-)
diff --git a/arch/arm/include/asm/arch-at91/at91sam9x5.h b/arch/arm/include/asm/arch-at91/at91sam9x5.h
index 85e42f5..de9fa50 100644
--- a/arch/arm/include/asm/arch-at91/at91sam9x5.h
+++ b/arch/arm/include/asm/arch-at91/at91sam9x5.h
@@ -165,6 +165,12 @@
#define ATMEL_ID_UHP ATMEL_ID_UHPHS
/*
+ * PMECC table in ROM
+ */
+#define ATMEL_PMECC_INDEX_OFFSET_512 0x8000
+#define ATMEL_PMECC_INDEX_OFFSET_1024 0x10000
+
+/*
* at91sam9x5 specific prototypes
*/
#ifndef __ASSEMBLY__
diff --git a/doc/README.atmel_pmecc b/doc/README.atmel_pmecc
index b483744..41f3bd7 100644
--- a/doc/README.atmel_pmecc
+++ b/doc/README.atmel_pmecc
@@ -19,17 +19,6 @@ To use PMECC in this driver, the user needs to set:
It can be 2, 4, 8, 12 or 24.
2. The PMECC sector size: CONFIG_PMECC_SECTOR_SIZE.
It only can be 512 or 1024.
- 3. The PMECC index lookup table's offsets in ROM code: CONFIG_PMECC_INDEX_TABLE_OFFSET.
- In the chip datasheet section "Boot Stragegies", you can find
- two Galois Field Table in the ROM code. One table is for 512-bytes
- sector. Another is for 1024-byte sector. Each Galois Field includes
- two sub-table: indext table & alpha table.
- In the beginning of each Galois Field Table is the index table,
- Alpha table is in the following.
- So the index table's offset is same as the Galois Field Table.
-
- Please set CONFIG_PMECC_INDEX_TABLE_OFFSET correctly according the
- Galois Field Table's offset base on the sector size you used.
Take AT91SAM9X5EK as an example, the board definition file likes:
@@ -38,7 +27,4 @@ Take AT91SAM9X5EK as an example, the board definition file likes:
#define CONFIG_ATMEL_NAND_HW_PMECC 1
#define CONFIG_PMECC_CAP 2
#define CONFIG_PMECC_SECTOR_SIZE 512
-#define CONFIG_PMECC_INDEX_TABLE_OFFSET 0x8000
-NOTE: If you use 1024 as the sector size, then need set 0x10000 as the
- CONFIG_PMECC_INDEX_TABLE_OFFSET
diff --git a/drivers/mtd/nand/atmel_nand.c b/drivers/mtd/nand/atmel_nand.c
index 3bfbaf8..139a479 100644
--- a/drivers/mtd/nand/atmel_nand.c
+++ b/drivers/mtd/nand/atmel_nand.c
@@ -653,7 +653,10 @@ static int atmel_pmecc_nand_init_params(struct nand_chip *nand,
cap = host->pmecc_corr_cap = CONFIG_PMECC_CAP;
sector_size = host->pmecc_sector_size = CONFIG_PMECC_SECTOR_SIZE;
- host->pmecc_index_table_offset = CONFIG_PMECC_INDEX_TABLE_OFFSET;
+ if (host->pmecc_sector_size == 512)
+ host->pmecc_index_table_offset = ATMEL_PMECC_INDEX_OFFSET_512;
+ else
+ host->pmecc_index_table_offset = ATMEL_PMECC_INDEX_OFFSET_1024;
MTDDEBUG(MTD_DEBUG_LEVEL1,
"Initialize PMECC params, cap: %d, sector: %d\n",
diff --git a/include/configs/at91sam9x5ek.h b/include/configs/at91sam9x5ek.h
index ee6e3fc..4364026 100644
--- a/include/configs/at91sam9x5ek.h
+++ b/include/configs/at91sam9x5ek.h
@@ -137,7 +137,6 @@
#define CONFIG_ATMEL_NAND_HW_PMECC 1
#define CONFIG_PMECC_CAP 2
#define CONFIG_PMECC_SECTOR_SIZE 512
-#define CONFIG_PMECC_INDEX_TABLE_OFFSET 0x8000
#define CONFIG_MTD_DEVICE
#define CONFIG_CMD_MTDPARTS
diff --git a/include/configs/sama5d3xek.h b/include/configs/sama5d3xek.h
index c13e983..67f8180 100644
--- a/include/configs/sama5d3xek.h
+++ b/include/configs/sama5d3xek.h
@@ -142,7 +142,6 @@
#define CONFIG_ATMEL_NAND_HW_PMECC
#define CONFIG_PMECC_CAP 4
#define CONFIG_PMECC_SECTOR_SIZE 512
-#define CONFIG_PMECC_INDEX_TABLE_OFFSET ATMEL_PMECC_INDEX_OFFSET_512
#define CONFIG_CMD_NAND_TRIMFFS
#endif
--
1.7.9.5
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [U-Boot] [PATCH v2 2/5] ARM: at91: sama5d3: remove unused definition about PMECC alpha table offset
2013-06-17 10:51 [U-Boot] [PATCH v2 0/5] ARM: at91: atmel_nand: check ONFI ecc minimum requirement Josh Wu
2013-06-17 10:51 ` [U-Boot] [PATCH v2 1/5] ARM: at91: atmel_nand: pmecc driver will select the galois table by sector size Josh Wu
@ 2013-06-17 10:51 ` Josh Wu
2013-06-17 10:51 ` [U-Boot] [PATCH v2 3/5] linux/compat.h: move dev_err, dev_info and dev_dbg from usb driver to compat.h Josh Wu
` (2 subsequent siblings)
4 siblings, 0 replies; 9+ messages in thread
From: Josh Wu @ 2013-06-17 10:51 UTC (permalink / raw)
To: u-boot
Signed-off-by: Josh Wu <josh.wu@atmel.com>
---
arch/arm/include/asm/arch-at91/sama5d3.h | 2 --
1 file changed, 2 deletions(-)
diff --git a/arch/arm/include/asm/arch-at91/sama5d3.h b/arch/arm/include/asm/arch-at91/sama5d3.h
index 883b932..aff12a6 100644
--- a/arch/arm/include/asm/arch-at91/sama5d3.h
+++ b/arch/arm/include/asm/arch-at91/sama5d3.h
@@ -194,8 +194,6 @@
*/
#define ATMEL_PMECC_INDEX_OFFSET_512 0x10000
#define ATMEL_PMECC_INDEX_OFFSET_1024 0x18000
-#define ATMEL_PMECC_ALPHA_OFFSET_512 0x10000
-#define ATMEL_PMECC_ALPHA_OFFSET_1024 0x18000
/*
* SAMA5D3 specific prototypes
--
1.7.9.5
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [U-Boot] [PATCH v2 3/5] linux/compat.h: move dev_err, dev_info and dev_dbg from usb driver to compat.h
2013-06-17 10:51 [U-Boot] [PATCH v2 0/5] ARM: at91: atmel_nand: check ONFI ecc minimum requirement Josh Wu
2013-06-17 10:51 ` [U-Boot] [PATCH v2 1/5] ARM: at91: atmel_nand: pmecc driver will select the galois table by sector size Josh Wu
2013-06-17 10:51 ` [U-Boot] [PATCH v2 2/5] ARM: at91: sama5d3: remove unused definition about PMECC alpha table offset Josh Wu
@ 2013-06-17 10:51 ` Josh Wu
2013-06-17 10:51 ` [U-Boot] [PATCH v2 4/5] mtd: atmel_nand: alloc memory instead of use static array for pmecc data Josh Wu
2013-06-17 10:51 ` [U-Boot] [PATCH v2 5/5] ARM: at91: atmel_nand: add code to check the ONFI parameter ECC requirement Josh Wu
4 siblings, 0 replies; 9+ messages in thread
From: Josh Wu @ 2013-06-17 10:51 UTC (permalink / raw)
To: u-boot
Since kernel code current use many dev_xxx() instead of using printk. To
compatible, move those dev_xxx from usb driver to linux/compat.h. Then all
driver code can use dev_err, dev_info and dev_vdbg.
This patch also removed duplicated macro definitions in usb driver.
Signed-off-by: Josh Wu <josh.wu@atmel.com>
---
drivers/usb/musb-new/linux-compat.h | 16 ----------------
include/linux/compat.h | 8 ++++++++
2 files changed, 8 insertions(+), 16 deletions(-)
diff --git a/drivers/usb/musb-new/linux-compat.h b/drivers/usb/musb-new/linux-compat.h
index 72c8c2b..d7a5663 100644
--- a/drivers/usb/musb-new/linux-compat.h
+++ b/drivers/usb/musb-new/linux-compat.h
@@ -39,15 +39,6 @@ typedef unsigned long dmaaddr_t;
#define cpu_relax() do {} while (0)
#define pr_debug(fmt, args...) debug(fmt, ##args)
-#define dev_dbg(dev, fmt, args...) \
- debug(fmt, ##args)
-#define dev_vdbg(dev, fmt, args...) \
- debug(fmt, ##args)
-#define dev_info(dev, fmt, args...) \
- printf(fmt, ##args)
-#define dev_err(dev, fmt, args...) \
- printf(fmt, ##args)
-#define printk printf
#define WARN(condition, fmt, args...) ({ \
int ret_warn = !!condition; \
@@ -55,13 +46,6 @@ typedef unsigned long dmaaddr_t;
printf(fmt, ##args); \
ret_warn; })
-#define KERN_DEBUG
-#define KERN_NOTICE
-#define KERN_WARNING
-#define KERN_ERR
-
-#define kfree(ptr) free(ptr)
-
#define pm_runtime_get_sync(dev) do {} while (0)
#define pm_runtime_put(dev) do {} while (0)
#define pm_runtime_put_sync(dev) do {} while (0)
diff --git a/include/linux/compat.h b/include/linux/compat.h
index e1338bf..3fdfb39 100644
--- a/include/linux/compat.h
+++ b/include/linux/compat.h
@@ -3,6 +3,14 @@
#define ndelay(x) udelay(1)
+#define dev_dbg(dev, fmt, args...) \
+ debug(fmt, ##args)
+#define dev_vdbg(dev, fmt, args...) \
+ debug(fmt, ##args)
+#define dev_info(dev, fmt, args...) \
+ printf(fmt, ##args)
+#define dev_err(dev, fmt, args...) \
+ printf(fmt, ##args)
#define printk printf
#define KERN_EMERG
--
1.7.9.5
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [U-Boot] [PATCH v2 4/5] mtd: atmel_nand: alloc memory instead of use static array for pmecc data
2013-06-17 10:51 [U-Boot] [PATCH v2 0/5] ARM: at91: atmel_nand: check ONFI ecc minimum requirement Josh Wu
` (2 preceding siblings ...)
2013-06-17 10:51 ` [U-Boot] [PATCH v2 3/5] linux/compat.h: move dev_err, dev_info and dev_dbg from usb driver to compat.h Josh Wu
@ 2013-06-17 10:51 ` Josh Wu
2013-06-17 10:51 ` [U-Boot] [PATCH v2 5/5] ARM: at91: atmel_nand: add code to check the ONFI parameter ECC requirement Josh Wu
4 siblings, 0 replies; 9+ messages in thread
From: Josh Wu @ 2013-06-17 10:51 UTC (permalink / raw)
To: u-boot
In this way, the pmecc corraction capbility can change in run time.
Signed-off-by: Josh Wu <josh.wu@atmel.com>
---
v1 --> v2:
replace printk with dev_err.
fix coding style.
free memory before return ENOMEM.
drivers/mtd/nand/atmel_nand.c | 63 ++++++++++++++++++++++++++++++++++++-----
1 file changed, 56 insertions(+), 7 deletions(-)
diff --git a/drivers/mtd/nand/atmel_nand.c b/drivers/mtd/nand/atmel_nand.c
index 139a479..3cd403f 100644
--- a/drivers/mtd/nand/atmel_nand.c
+++ b/drivers/mtd/nand/atmel_nand.c
@@ -32,6 +32,7 @@
#include <asm/arch/gpio.h>
#include <asm/arch/at91_pio.h>
+#include <malloc.h>
#include <nand.h>
#include <watchdog.h>
@@ -66,13 +67,13 @@ struct atmel_nand_host {
void __iomem *pmecc_index_of;
/* data for pmecc computation */
- int16_t pmecc_smu[(CONFIG_PMECC_CAP + 2) * (2 * CONFIG_PMECC_CAP + 1)];
- int16_t pmecc_partial_syn[2 * CONFIG_PMECC_CAP + 1];
- int16_t pmecc_si[2 * CONFIG_PMECC_CAP + 1];
- int16_t pmecc_lmu[CONFIG_PMECC_CAP + 1]; /* polynomal order */
- int pmecc_mu[CONFIG_PMECC_CAP + 1];
- int pmecc_dmu[CONFIG_PMECC_CAP + 1];
- int pmecc_delta[CONFIG_PMECC_CAP + 1];
+ int16_t *pmecc_smu;
+ int16_t *pmecc_partial_syn;
+ int16_t *pmecc_si;
+ int16_t *pmecc_lmu; /* polynomal order */
+ int *pmecc_mu;
+ int *pmecc_dmu;
+ int *pmecc_delta;
};
static struct atmel_nand_host pmecc_host;
@@ -125,6 +126,48 @@ static void __iomem *pmecc_get_alpha_to(struct atmel_nand_host *host)
table_size * sizeof(int16_t);
}
+static void pmecc_data_free(struct atmel_nand_host *host)
+{
+ free(host->pmecc_partial_syn);
+ free(host->pmecc_si);
+ free(host->pmecc_lmu);
+ free(host->pmecc_smu);
+ free(host->pmecc_mu);
+ free(host->pmecc_dmu);
+ free(host->pmecc_delta);
+}
+
+static int pmecc_data_alloc(struct atmel_nand_host *host)
+{
+ const int cap = host->pmecc_corr_cap;
+ int size;
+
+ size = (2 * cap + 1) * sizeof(int16_t);
+ host->pmecc_partial_syn = malloc(size);
+ host->pmecc_si = malloc(size);
+ host->pmecc_lmu = malloc((cap + 1) * sizeof(int16_t));
+ host->pmecc_smu = malloc((cap + 2) * size);
+
+ size = (cap + 1) * sizeof(int);
+ host->pmecc_mu = malloc(size);
+ host->pmecc_dmu = malloc(size);
+ host->pmecc_delta = malloc(size);
+
+ if (host->pmecc_partial_syn &&
+ host->pmecc_si &&
+ host->pmecc_lmu &&
+ host->pmecc_smu &&
+ host->pmecc_mu &&
+ host->pmecc_dmu &&
+ host->pmecc_delta)
+ return 0;
+
+ /* error happened */
+ pmecc_data_free(host);
+ return -ENOMEM;
+
+}
+
static void pmecc_gen_syndrome(struct mtd_info *mtd, int sector)
{
struct nand_chip *nand_chip = mtd->priv;
@@ -710,6 +753,12 @@ static int atmel_pmecc_nand_init_params(struct nand_chip *nand,
return 0;
}
+ /* Allocate data for PMECC computation */
+ if (pmecc_data_alloc(host)) {
+ dev_err(NULL, "Cannot allocate memory for PMECC computation!\n");
+ return -ENOMEM;
+ }
+
nand->ecc.read_page = atmel_nand_pmecc_read_page;
nand->ecc.write_page = atmel_nand_pmecc_write_page;
nand->ecc.strength = cap;
--
1.7.9.5
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [U-Boot] [PATCH v2 5/5] ARM: at91: atmel_nand: add code to check the ONFI parameter ECC requirement
2013-06-17 10:51 [U-Boot] [PATCH v2 0/5] ARM: at91: atmel_nand: check ONFI ecc minimum requirement Josh Wu
` (3 preceding siblings ...)
2013-06-17 10:51 ` [U-Boot] [PATCH v2 4/5] mtd: atmel_nand: alloc memory instead of use static array for pmecc data Josh Wu
@ 2013-06-17 10:51 ` Josh Wu
2013-06-20 20:57 ` Scott Wood
4 siblings, 1 reply; 9+ messages in thread
From: Josh Wu @ 2013-06-17 10:51 UTC (permalink / raw)
To: u-boot
1. if CONFIG_SYS_NAND_ONFI_DETECTION is defined, driver will check NAND flash's
ecc minimum requirement in ONFI parameter.
a) if CONFIG_PMECC_CAP, CONFIG_PMECC_SECTOR_SIZE are defined. then use it.
Driver will display a WARNING if the values are different from ONFI
parameters.
b) if CONFIG_PMECC_CAP, CONFIG_PMECC_SECTOR_SIZE are not defined, then use
the value from ONFI parameters.
* If ONFI ECC parameters are in ONFI extended parameter page, since we
are not support it, so assume the minimum ecc requirement is 2 bits
in 512 bytes.
* For non-ONFI support nand flash, also assume the minimum ecc
requirement is 2 bits in 512 bytes.
2. if CONFIG_SYS_NAND_ONFI_DETECTION is not defined, just use CONFIG_PMECC_CAP
and CONFIG_PMECC_SECTOR_SIZE.
Signed-off-by: Josh Wu <josh.wu@atmel.com>
---
v1 --> v2:
use dev_err/info to replace printk.
fix typo mistake, that should use <= instead of use <.
drivers/mtd/nand/atmel_nand.c | 129 ++++++++++++++++++++++++++++++++++++++++-
1 file changed, 127 insertions(+), 2 deletions(-)
diff --git a/drivers/mtd/nand/atmel_nand.c b/drivers/mtd/nand/atmel_nand.c
index 3cd403f..ceba562 100644
--- a/drivers/mtd/nand/atmel_nand.c
+++ b/drivers/mtd/nand/atmel_nand.c
@@ -681,6 +681,98 @@ static void atmel_pmecc_core_init(struct mtd_info *mtd)
pmecc_writel(host->pmecc, ctrl, PMECC_CTRL_ENABLE);
}
+#ifdef CONFIG_SYS_NAND_ONFI_DETECTION
+/*
+ * get_onfi_ecc_param - Get ECC requirement from ONFI parameters
+ * @ecc_bits: store the ONFI ECC correct bits capbility
+ * @sector_size: in how many bytes that ONFI require to correct @ecc_bits
+ *
+ * Returns -1 if ONFI parameters is not supported. In this case @ecc_bits,
+ * @sector_size are initialize to 0.
+ * Return 0 if success to get the ECC requirement.
+ */
+static int get_onfi_ecc_param(struct nand_chip *chip,
+ int *ecc_bits, int *sector_size)
+{
+ *ecc_bits = *sector_size = 0;
+
+ if (chip->onfi_params.ecc_bits == 0xff)
+ /* TODO: the sector_size and ecc_bits need to be find in
+ * extended ecc parameter, currently we don't support it.
+ */
+ return -1;
+
+ *ecc_bits = chip->onfi_params.ecc_bits;
+
+ /* The default sector size (ecc codeword size) is 512 */
+ *sector_size = 512;
+
+ return 0;
+}
+
+/*
+ * pmecc_choose_ecc - Get ecc requirement from ONFI parameters. If
+ * pmecc_corr_cap or pmecc_sector_size is 0, then set it as
+ * ONFI ECC parameters.
+ * @host: point to an atmel_nand_host structure.
+ * if host->pmecc_corr_cap is 0 then set it as the ONFI ecc_bits.
+ * if host->pmecc_sector_size is 0 then set it as the ONFI sector_size.
+ * @chip: point to an nand_chip structure.
+ * @cap: store the ONFI ECC correct bits capbility
+ * @sector_size: in how many bytes that ONFI require to correct @ecc_bits
+ *
+ * Return 0 if success. otherwise return the error code.
+ */
+static int pmecc_choose_ecc(struct atmel_nand_host *host,
+ struct nand_chip *chip,
+ int *cap, int *sector_size)
+{
+ /* Get ECC requirement from ONFI parameters */
+ *cap = *sector_size = 0;
+ if (chip->onfi_version) {
+ if (!get_onfi_ecc_param(chip, cap, sector_size))
+ pr_debug("ONFI params, minimum required ECC: %d bits in %d bytes\n",
+ *cap, *sector_size);
+ else
+ dev_info(NULL, "NAND chip ECC reqirement is in Extended ONFI parameter, we don't support yet.\n");
+ } else {
+ dev_info(NULL, "NAND chip is not ONFI compliant, assume ecc_bits is 2 in 512 bytes");
+ }
+ if (*cap == 0 && *sector_size == 0) {
+ /* Non-ONFI compliant or use extended ONFI parameters */
+ *cap = 2;
+ *sector_size = 512;
+ }
+
+ /* If head file doesn't specify then use the one in ONFI parameters */
+ if (host->pmecc_corr_cap == 0) {
+ /* use the most fitable ecc bits (the near bigger one ) */
+ if (*cap <= 2)
+ host->pmecc_corr_cap = 2;
+ else if (*cap <= 4)
+ host->pmecc_corr_cap = 4;
+ else if (*cap <= 8)
+ host->pmecc_corr_cap = 8;
+ else if (*cap <= 12)
+ host->pmecc_corr_cap = 12;
+ else if (*cap <= 24)
+ host->pmecc_corr_cap = 24;
+ else
+ return -EINVAL;
+ }
+ if (host->pmecc_sector_size == 0) {
+ /* use the most fitable sector size (the near smaller one ) */
+ if (*sector_size >= 1024)
+ host->pmecc_sector_size = 1024;
+ else if (*sector_size >= 512)
+ host->pmecc_sector_size = 512;
+ else
+ return -EINVAL;
+ }
+ return 0;
+}
+#endif
+
static int atmel_pmecc_nand_init_params(struct nand_chip *nand,
struct mtd_info *mtd)
{
@@ -694,8 +786,41 @@ static int atmel_pmecc_nand_init_params(struct nand_chip *nand,
nand->ecc.correct = NULL;
nand->ecc.hwctl = NULL;
- cap = host->pmecc_corr_cap = CONFIG_PMECC_CAP;
- sector_size = host->pmecc_sector_size = CONFIG_PMECC_SECTOR_SIZE;
+#ifdef CONFIG_SYS_NAND_ONFI_DETECTION
+ host->pmecc_corr_cap = host->pmecc_sector_size = 0;
+
+#ifdef CONFIG_PMECC_CAP
+ host->pmecc_corr_cap = CONFIG_PMECC_CAP;
+#endif
+#ifdef CONFIG_PMECC_SECTOR_SIZE
+ host->pmecc_sector_size = CONFIG_PMECC_SECTOR_SIZE;
+#endif
+ /* Get ECC requirement of ONFI parameters. And if CONFIG_PMECC_CAP or
+ * CONFIG_PMECC_SECTOR_SIZE not defined, then use ecc_bits, sector_size
+ * from ONFI.
+ */
+ if (pmecc_choose_ecc(host, nand, &cap, §or_size)) {
+ dev_err(NULL, "The NAND flash's ECC requirement(ecc_bits: %d, sector_size: %d) are not support!",
+ cap, sector_size);
+ return -EINVAL;
+ }
+
+ if (cap > host->pmecc_corr_cap)
+ dev_info(NULL, "WARNING: Using different ecc correct bits(%d bit) from Nand ONFI ECC reqirement (%d bit).\n",
+ host->pmecc_corr_cap, cap);
+ if (sector_size < host->pmecc_sector_size)
+ dev_info(NULL, "WARNING: Using different ecc correct sector size (%d bytes) from Nand ONFI ECC reqirement (%d bytes).\n",
+ host->pmecc_sector_size, sector_size);
+#else /* CONFIG_SYS_NAND_ONFI_DETECTION */
+ host->pmecc_corr_cap = CONFIG_PMECC_CAP;
+ host->pmecc_sector_size = CONFIG_PMECC_SECTOR_SIZE;
+#endif
+
+ cap = host->pmecc_corr_cap;
+ sector_size = host->pmecc_sector_size;
+
+ /* TODO: need check whether cap & sector_size is validate */
+
if (host->pmecc_sector_size == 512)
host->pmecc_index_table_offset = ATMEL_PMECC_INDEX_OFFSET_512;
else
--
1.7.9.5
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [U-Boot] [PATCH v2 5/5] ARM: at91: atmel_nand: add code to check the ONFI parameter ECC requirement
2013-06-17 10:51 ` [U-Boot] [PATCH v2 5/5] ARM: at91: atmel_nand: add code to check the ONFI parameter ECC requirement Josh Wu
@ 2013-06-20 20:57 ` Scott Wood
2013-06-25 6:33 ` Josh Wu
0 siblings, 1 reply; 9+ messages in thread
From: Scott Wood @ 2013-06-20 20:57 UTC (permalink / raw)
To: u-boot
On 06/17/2013 05:51:21 AM, Josh Wu wrote:
> +static int pmecc_choose_ecc(struct atmel_nand_host *host,
> + struct nand_chip *chip,
> + int *cap, int *sector_size)
> +{
> + /* Get ECC requirement from ONFI parameters */
> + *cap = *sector_size = 0;
> + if (chip->onfi_version) {
> + if (!get_onfi_ecc_param(chip, cap, sector_size))
> + pr_debug("ONFI params, minimum required ECC: %d
> bits in %d bytes\n",
> + *cap, *sector_size);
> + else
> + dev_info(NULL, "NAND chip ECC reqirement is in
> Extended ONFI parameter, we don't support yet.\n");
> + } else {
> + dev_info(NULL, "NAND chip is not ONFI compliant, assume
> ecc_bits is 2 in 512 bytes");
> + }
Don't pass NULL to dev_info(). Either pass host->dev as the Linux code
does (and use dev_dbg rather than pr_debug), or just use printf() (and
debug()) if there's no intent to sync this change back to Linux. It
doesn't matter if host->dev doesn't exist in U-Boot, as the macro
doesn't use that parameter, but the only reason to use Linux-style
prints is to keep the code common with Linux.
-Scott
^ permalink raw reply [flat|nested] 9+ messages in thread
* [U-Boot] [PATCH v2 5/5] ARM: at91: atmel_nand: add code to check the ONFI parameter ECC requirement
2013-06-20 20:57 ` Scott Wood
@ 2013-06-25 6:33 ` Josh Wu
2013-06-25 15:10 ` Scott Wood
0 siblings, 1 reply; 9+ messages in thread
From: Josh Wu @ 2013-06-25 6:33 UTC (permalink / raw)
To: u-boot
Hi, Dear Scott
On 6/21/2013 4:57 AM, Scott Wood wrote:
> On 06/17/2013 05:51:21 AM, Josh Wu wrote:
>> +static int pmecc_choose_ecc(struct atmel_nand_host *host,
>> + struct nand_chip *chip,
>> + int *cap, int *sector_size)
>> +{
>> + /* Get ECC requirement from ONFI parameters */
>> + *cap = *sector_size = 0;
>> + if (chip->onfi_version) {
>> + if (!get_onfi_ecc_param(chip, cap, sector_size))
>> + pr_debug("ONFI params, minimum required ECC: %d bits in
>> %d bytes\n",
>> + *cap, *sector_size);
>> + else
>> + dev_info(NULL, "NAND chip ECC reqirement is in Extended
>> ONFI parameter, we don't support yet.\n");
>> + } else {
>> + dev_info(NULL, "NAND chip is not ONFI compliant, assume
>> ecc_bits is 2 in 512 bytes");
>> + }
>
> Don't pass NULL to dev_info(). Either pass host->dev as the Linux
> code does (and use dev_dbg rather than pr_debug), or just use printf()
> (and debug()) if there's no intent to sync this change back to Linux.
> It doesn't matter if host->dev doesn't exist in U-Boot, as the macro
> doesn't use that parameter, but the only reason to use Linux-style
> prints is to keep the code common with Linux.
understood. I prefer to use dev_xxx(host->dev, ...) to keep consistent
with kernel.
I use pr_debug because it is defined as MTDDEBUG in mtd/mtd.h, that is
exact what I want.
Since if use debug(), then when I enable debug, there are many mtd
non-related message shows up.
Best Regards,
Josh Wu
>
> -Scott
^ permalink raw reply [flat|nested] 9+ messages in thread
* [U-Boot] [PATCH v2 5/5] ARM: at91: atmel_nand: add code to check the ONFI parameter ECC requirement
2013-06-25 6:33 ` Josh Wu
@ 2013-06-25 15:10 ` Scott Wood
0 siblings, 0 replies; 9+ messages in thread
From: Scott Wood @ 2013-06-25 15:10 UTC (permalink / raw)
To: u-boot
On 06/25/2013 01:33:32 AM, Josh Wu wrote:
> Hi, Dear Scott
>
> On 6/21/2013 4:57 AM, Scott Wood wrote:
>> On 06/17/2013 05:51:21 AM, Josh Wu wrote:
>>> +static int pmecc_choose_ecc(struct atmel_nand_host *host,
>>> + struct nand_chip *chip,
>>> + int *cap, int *sector_size)
>>> +{
>>> + /* Get ECC requirement from ONFI parameters */
>>> + *cap = *sector_size = 0;
>>> + if (chip->onfi_version) {
>>> + if (!get_onfi_ecc_param(chip, cap, sector_size))
>>> + pr_debug("ONFI params, minimum required ECC: %d bits
>>> in %d bytes\n",
>>> + *cap, *sector_size);
>>> + else
>>> + dev_info(NULL, "NAND chip ECC reqirement is in
>>> Extended ONFI parameter, we don't support yet.\n");
>>> + } else {
>>> + dev_info(NULL, "NAND chip is not ONFI compliant, assume
>>> ecc_bits is 2 in 512 bytes");
>>> + }
>>
>> Don't pass NULL to dev_info(). Either pass host->dev as the Linux
>> code does (and use dev_dbg rather than pr_debug), or just use
>> printf() (and debug()) if there's no intent to sync this change back
>> to Linux. It doesn't matter if host->dev doesn't exist in U-Boot,
>> as the macro doesn't use that parameter, but the only reason to use
>> Linux-style prints is to keep the code common with Linux.
>
> understood. I prefer to use dev_xxx(host->dev, ...) to keep
> consistent with kernel.
>
> I use pr_debug because it is defined as MTDDEBUG in mtd/mtd.h, that
> is exact what I want.
> Since if use debug(), then when I enable debug, there are many mtd
> non-related message shows up.
If pr_debug() and dev_dbg() are behaving differently in U-Boot, that's
a bug that should be fixed.
-Scott
^ permalink raw reply [flat|nested] 9+ messages in thread
end of thread, other threads:[~2013-06-25 15:10 UTC | newest]
Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-06-17 10:51 [U-Boot] [PATCH v2 0/5] ARM: at91: atmel_nand: check ONFI ecc minimum requirement Josh Wu
2013-06-17 10:51 ` [U-Boot] [PATCH v2 1/5] ARM: at91: atmel_nand: pmecc driver will select the galois table by sector size Josh Wu
2013-06-17 10:51 ` [U-Boot] [PATCH v2 2/5] ARM: at91: sama5d3: remove unused definition about PMECC alpha table offset Josh Wu
2013-06-17 10:51 ` [U-Boot] [PATCH v2 3/5] linux/compat.h: move dev_err, dev_info and dev_dbg from usb driver to compat.h Josh Wu
2013-06-17 10:51 ` [U-Boot] [PATCH v2 4/5] mtd: atmel_nand: alloc memory instead of use static array for pmecc data Josh Wu
2013-06-17 10:51 ` [U-Boot] [PATCH v2 5/5] ARM: at91: atmel_nand: add code to check the ONFI parameter ECC requirement Josh Wu
2013-06-20 20:57 ` Scott Wood
2013-06-25 6:33 ` Josh Wu
2013-06-25 15:10 ` Scott Wood
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.