[PATCH v3 0/4] mtd: rawnand: Fixes nand infra delay setting and FSMC nand controller

[PATCH v3 0/4] mtd: rawnand: Fixes nand infra delay setting and FSMC nand controller

[Herve Codina]

Hi,

This patch series is mainly related to FSMC nand controller fixes except
for the two first patches which fix some issues in nand base.

Details are given in each patches.

Related to the v1 series:
- the patch related to 8bit accesses has been removed as it was
  not needed (I misunderstood 8bit accesses meaning).
- A new patch is introduced related to nand_choose_best_timings()
  issue I discovered during the v2.

Best regards,
Herve Codina

Herve Codina (4):
  mtd: rawnand: Fix nand_erase_op delay
  mtd: rawnand: Fix nand_choose_best_timings() on unsupported interface
  mtd: rawnand: fsmc: Take instruction delay into account
  mtd: rawnand: fsmc: Fix timing computation

 drivers/mtd/nand/raw/fsmc_nand.c | 36 +++++++++++++++++++++++++-------
 drivers/mtd/nand/raw/nand_base.c |  6 +++---
 2 files changed, 31 insertions(+), 11 deletions(-)

-- 
2.31.1

[PATCH v3 1/4] mtd: rawnand: Fix nand_erase_op delay

[Herve Codina]

NAND_OP_CMD() expects a delay parameter in nanoseconds.
The delay value is wrongly given in milliseconds.

Fix the conversion macro used in order to set this
delay in nanoseconds.

Fixes: d7a773e8812b ("mtd: rawnand: Access SDR and NV-DDR timings through a common macro")
Fixes: 8878b126df76 ("mtd: nand: add ->exec_op() implementation")
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
---
Changes v1 -> v2:
- Commit log reword

Changes v2 -> v3:
- Fix commit log typo
- Added 'Fixes' references

 drivers/mtd/nand/raw/nand_base.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/mtd/nand/raw/nand_base.c b/drivers/mtd/nand/raw/nand_base.c
index 3d6c6e880520..5c6b065837ef 100644
--- a/drivers/mtd/nand/raw/nand_base.c
+++ b/drivers/mtd/nand/raw/nand_base.c
@@ -1837,7 +1837,7 @@ int nand_erase_op(struct nand_chip *chip, unsigned int eraseblock)
 			NAND_OP_CMD(NAND_CMD_ERASE1, 0),
 			NAND_OP_ADDR(2, addrs, 0),
 			NAND_OP_CMD(NAND_CMD_ERASE2,
-				    NAND_COMMON_TIMING_MS(conf, tWB_max)),
+				    NAND_COMMON_TIMING_NS(conf, tWB_max)),
 			NAND_OP_WAIT_RDY(NAND_COMMON_TIMING_MS(conf, tBERS_max),
 					 0),
 		};
-- 
2.31.1

Re: [PATCH v3 1/4] mtd: rawnand: Fix nand_erase_op delay

[Miquel Raynal]

On Fri, 2021-11-19 at 15:03:13 UTC, Herve Codina wrote:
> NAND_OP_CMD() expects a delay parameter in nanoseconds.
> The delay value is wrongly given in milliseconds.
> 
> Fix the conversion macro used in order to set this
> delay in nanoseconds.
> 
> Fixes: d7a773e8812b ("mtd: rawnand: Access SDR and NV-DDR timings through a common macro")
> Fixes: 8878b126df76 ("mtd: nand: add ->exec_op() implementation")
> Signed-off-by: Herve Codina <herve.codina@bootlin.com>

Applied to https://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux.git mtd/fixes, thanks.

Miquel

[PATCH v3 2/4] mtd: rawnand: Fix nand_choose_best_timings() on unsupported interface

[Herve Codina]

When the NV-DDR interface is not supported by the NAND chip,
the value of onfi->nvddr_timing_modes is 0. In this case,
the best_mode variable value in nand_choose_best_nvddr_timings()
is -1. The last for-loop is skipped and the function returns an
uninitialized value.
If this returned value is 0, the nand_choose_best_sdr_timings()
is not executed and no 'best timing' are set. This leads the host
controller and the NAND chip working at default mode 0 timing
even if a better timing can be used.

Fix this uninitialized returned value.

nand_choose_best_sdr_timings() is pretty similar to
nand_choose_best_nvddr_timings(). Even if onfi->sdr_timing_modes
should never be seen as 0, nand_choose_best_sdr_timings() returned
value is fixed.

Fixes: a9ecc8c814e9 ("mtd: rawnand: Choose the best timings, NV-DDR included")
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
---
Changes v1 to v2:
- New patch in v2 series

Changes v2 -> v3:
- Fix commit log typo

 drivers/mtd/nand/raw/nand_base.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/mtd/nand/raw/nand_base.c b/drivers/mtd/nand/raw/nand_base.c
index 5c6b065837ef..a130320de412 100644
--- a/drivers/mtd/nand/raw/nand_base.c
+++ b/drivers/mtd/nand/raw/nand_base.c
@@ -926,7 +926,7 @@ int nand_choose_best_sdr_timings(struct nand_chip *chip,
 				 struct nand_sdr_timings *spec_timings)
 {
 	const struct nand_controller_ops *ops = chip->controller->ops;
-	int best_mode = 0, mode, ret;
+	int best_mode = 0, mode, ret = -EOPNOTSUPP;
 
 	iface->type = NAND_SDR_IFACE;
 
@@ -977,7 +977,7 @@ int nand_choose_best_nvddr_timings(struct nand_chip *chip,
 				   struct nand_nvddr_timings *spec_timings)
 {
 	const struct nand_controller_ops *ops = chip->controller->ops;
-	int best_mode = 0, mode, ret;
+	int best_mode = 0, mode, ret = -EOPNOTSUPP;
 
 	iface->type = NAND_NVDDR_IFACE;
 
-- 
2.31.1

Re: [PATCH v3 2/4] mtd: rawnand: Fix nand_choose_best_timings() on unsupported interface

[Miquel Raynal]

On Fri, 2021-11-19 at 15:03:14 UTC, Herve Codina wrote:
> When the NV-DDR interface is not supported by the NAND chip,
> the value of onfi->nvddr_timing_modes is 0. In this case,
> the best_mode variable value in nand_choose_best_nvddr_timings()
> is -1. The last for-loop is skipped and the function returns an
> uninitialized value.
> If this returned value is 0, the nand_choose_best_sdr_timings()
> is not executed and no 'best timing' are set. This leads the host
> controller and the NAND chip working at default mode 0 timing
> even if a better timing can be used.
> 
> Fix this uninitialized returned value.
> 
> nand_choose_best_sdr_timings() is pretty similar to
> nand_choose_best_nvddr_timings(). Even if onfi->sdr_timing_modes
> should never be seen as 0, nand_choose_best_sdr_timings() returned
> value is fixed.
> 
> Fixes: a9ecc8c814e9 ("mtd: rawnand: Choose the best timings, NV-DDR included")
> Signed-off-by: Herve Codina <herve.codina@bootlin.com>

Applied to https://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux.git mtd/fixes, thanks.

Miquel

[PATCH v3 3/4] mtd: rawnand: fsmc: Take instruction delay into account

[Herve Codina]

The FSMC NAND controller should apply a delay after the
instruction has been issued on the bus.
The FSMC NAND controller driver did not handle this delay.

Add this waiting delay in the FSMC NAND controller driver.

Fixes: 4da712e70294 ("mtd: nand: fsmc: use ->exec_op()")
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
---
Changes v1 -> v2:
- Commit log reword

Changes v2 -> v3:
- Added 'Fixes' reference

 drivers/mtd/nand/raw/fsmc_nand.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/mtd/nand/raw/fsmc_nand.c b/drivers/mtd/nand/raw/fsmc_nand.c
index 658f0cbe7ce8..0a6c9ef0ea8b 100644
--- a/drivers/mtd/nand/raw/fsmc_nand.c
+++ b/drivers/mtd/nand/raw/fsmc_nand.c
@@ -15,6 +15,7 @@
 
 #include <linux/clk.h>
 #include <linux/completion.h>
+#include <linux/delay.h>
 #include <linux/dmaengine.h>
 #include <linux/dma-direction.h>
 #include <linux/dma-mapping.h>
@@ -664,6 +665,9 @@ static int fsmc_exec_op(struct nand_chip *chip, const struct nand_operation *op,
 						instr->ctx.waitrdy.timeout_ms);
 			break;
 		}
+
+		if (instr->delay_ns)
+			ndelay(instr->delay_ns);
 	}
 
 	return ret;
-- 
2.31.1

Re: [PATCH v3 3/4] mtd: rawnand: fsmc: Take instruction delay into account

[Miquel Raynal]

On Fri, 2021-11-19 at 15:03:15 UTC, Herve Codina wrote:
> The FSMC NAND controller should apply a delay after the
> instruction has been issued on the bus.
> The FSMC NAND controller driver did not handle this delay.
> 
> Add this waiting delay in the FSMC NAND controller driver.
> 
> Fixes: 4da712e70294 ("mtd: nand: fsmc: use ->exec_op()")
> Signed-off-by: Herve Codina <herve.codina@bootlin.com>

Applied to https://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux.git mtd/fixes, thanks.

Miquel

[PATCH v3 4/4] mtd: rawnand: fsmc: Fix timing computation

[Herve Codina]

Under certain circumstances, the timing settings calculated by
the FSMC NAND controller driver were inaccurate.
These settings led to incorrect data reads or fallback to
timing mode 0 depending on the NAND chip used.

The timing computation did not take into account the following
constraint given in SPEAr3xx reference manual:
  twait >= tCEA - (tset * TCLK) + TOUTDEL + TINDEL

Enhance the timings calculation by taking into account this
additional constraint.

This change has no impact on slow timing modes such as mode 0.
Indeed, on mode 0, computed values are the same with and
without the patch.

NANDs which previously stayed in mode 0 because of fallback to
mode 0 can now work at higher speeds and NANDs which were not
working at all because of the corrupted data work at high
speeds without troubles.

Overall improvement on a Micron/MT29F1G08 (flash_speed tool):
                        mode0       mode3
eraseblock write speed  3220 KiB/s  4511 KiB/s
eraseblock read speed   4491 KiB/s  7529 KiB/s

Fixes: d9fb079571833 ("mtd: nand: fsmc: add support for SDR timings")
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
---
Changes v1 to v2:
- Commit log reword
- Added performance details in commit log
- Used #define for TOUTDEL and TINDEL and
  Fixed coding style
- Used max3()

Changes v2 -> v3:
- Improved commit log
- Added 'Fixes' reference
- Changed comment in source code

 drivers/mtd/nand/raw/fsmc_nand.c | 32 ++++++++++++++++++++++++--------
 1 file changed, 24 insertions(+), 8 deletions(-)

diff --git a/drivers/mtd/nand/raw/fsmc_nand.c b/drivers/mtd/nand/raw/fsmc_nand.c
index 0a6c9ef0ea8b..6b2bda815b88 100644
--- a/drivers/mtd/nand/raw/fsmc_nand.c
+++ b/drivers/mtd/nand/raw/fsmc_nand.c
@@ -94,6 +94,14 @@
 
 #define FSMC_BUSY_WAIT_TIMEOUT	(1 * HZ)
 
+/*
+ * According to SPEAr300 Reference Manual (RM0082)
+ *  TOUDEL = 7ns (Output delay from the flip-flops to the board)
+ *  TINDEL = 5ns (Input delay from the board to the flipflop)
+ */
+#define TOUTDEL	7000
+#define TINDEL	5000
+
 struct fsmc_nand_timings {
 	u8 tclr;
 	u8 tar;
@@ -278,7 +286,7 @@ static int fsmc_calc_timings(struct fsmc_nand_data *host,
 {
 	unsigned long hclk = clk_get_rate(host->clk);
 	unsigned long hclkn = NSEC_PER_SEC / hclk;
-	u32 thiz, thold, twait, tset;
+	u32 thiz, thold, twait, tset, twait_min;
 
 	if (sdrt->tRC_min < 30000)
 		return -EOPNOTSUPP;
@@ -310,13 +318,6 @@ static int fsmc_calc_timings(struct fsmc_nand_data *host,
 	else if (tims->thold > FSMC_THOLD_MASK)
 		tims->thold = FSMC_THOLD_MASK;
 
-	twait = max(sdrt->tRP_min, sdrt->tWP_min);
-	tims->twait = DIV_ROUND_UP(twait / 1000, hclkn) - 1;
-	if (tims->twait == 0)
-		tims->twait = 1;
-	else if (tims->twait > FSMC_TWAIT_MASK)
-		tims->twait = FSMC_TWAIT_MASK;
-
 	tset = max(sdrt->tCS_min - sdrt->tWP_min,
 		   sdrt->tCEA_max - sdrt->tREA_max);
 	tims->tset = DIV_ROUND_UP(tset / 1000, hclkn) - 1;
@@ -325,6 +326,21 @@ static int fsmc_calc_timings(struct fsmc_nand_data *host,
 	else if (tims->tset > FSMC_TSET_MASK)
 		tims->tset = FSMC_TSET_MASK;
 
+	/*
+	 * According to SPEAr300 Reference Manual (RM0082) which gives more
+	 * information related to FSMSC timings than the SPEAr600 one (RM0305),
+	 *   twait >= tCEA - (tset * TCLK) + TOUTDEL + TINDEL
+	 */
+	twait_min = sdrt->tCEA_max - ((tims->tset + 1) * hclkn * 1000)
+		    + TOUTDEL + TINDEL;
+	twait = max3(sdrt->tRP_min, sdrt->tWP_min, twait_min);
+
+	tims->twait = DIV_ROUND_UP(twait / 1000, hclkn) - 1;
+	if (tims->twait == 0)
+		tims->twait = 1;
+	else if (tims->twait > FSMC_TWAIT_MASK)
+		tims->twait = FSMC_TWAIT_MASK;
+
 	return 0;
 }
 
-- 
2.31.1

Re: [PATCH v3 4/4] mtd: rawnand: fsmc: Fix timing computation

[Miquel Raynal]

On Fri, 2021-11-19 at 15:03:16 UTC, Herve Codina wrote:
> Under certain circumstances, the timing settings calculated by
> the FSMC NAND controller driver were inaccurate.
> These settings led to incorrect data reads or fallback to
> timing mode 0 depending on the NAND chip used.
> 
> The timing computation did not take into account the following
> constraint given in SPEAr3xx reference manual:
>   twait >= tCEA - (tset * TCLK) + TOUTDEL + TINDEL
> 
> Enhance the timings calculation by taking into account this
> additional constraint.
> 
> This change has no impact on slow timing modes such as mode 0.
> Indeed, on mode 0, computed values are the same with and
> without the patch.
> 
> NANDs which previously stayed in mode 0 because of fallback to
> mode 0 can now work at higher speeds and NANDs which were not
> working at all because of the corrupted data work at high
> speeds without troubles.
> 
> Overall improvement on a Micron/MT29F1G08 (flash_speed tool):
>                         mode0       mode3
> eraseblock write speed  3220 KiB/s  4511 KiB/s
> eraseblock read speed   4491 KiB/s  7529 KiB/s
> 
> Fixes: d9fb079571833 ("mtd: nand: fsmc: add support for SDR timings")
> Signed-off-by: Herve Codina <herve.codina@bootlin.com>

Applied to https://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux.git mtd/fixes, thanks.

Miquel