[PATCH v1 0/2] spi: stm32_qspi: flags management fixes

[PATCH v1 0/2] spi: stm32_qspi: flags management fixes

[Patrice Chotard]

This series update flags management in the following cases:
  - Always check TCF flag in stm32_qspi_wait_cmd()
  - Don't check BUSY flag when sending new command

Patrice Chotard (2):
  spi: stm32_qspi: Always check SR_TCF flags in stm32_qspi_wait_cmd()
  spi: stm32_qspi: Remove SR_BUSY bit check before sending command

 drivers/spi/stm32_qspi.c | 27 +++++++++++----------------
 1 file changed, 11 insertions(+), 16 deletions(-)

-- 
2.25.1

[PATCH v1 1/2] spi: stm32_qspi: Always check SR_TCF flags in stm32_qspi_wait_cmd()

[Patrice Chotard]

Currently, SR_TCF flag is checked in case there is data, this criteria
is not correct.

SR_TCF flags is set when programmed number of bytes have been transferred
to the memory device ("bytes" comprised command and data send to the
SPI device).
So even if there is no data, we must check SR_TCF flag.

Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
---

 drivers/spi/stm32_qspi.c | 23 +++++++++++------------
 1 file changed, 11 insertions(+), 12 deletions(-)

diff --git a/drivers/spi/stm32_qspi.c b/drivers/spi/stm32_qspi.c
index 8f4aabc3d1..3c8faecb54 100644
--- a/drivers/spi/stm32_qspi.c
+++ b/drivers/spi/stm32_qspi.c
@@ -150,20 +150,19 @@ static int _stm32_qspi_wait_cmd(struct stm32_qspi_priv *priv,
 	u32 sr;
 	int ret = 0;
 
-	if (op->data.nbytes) {
-		ret = readl_poll_timeout(&priv->regs->sr, sr,
-					 sr & STM32_QSPI_SR_TCF,
-					 STM32_QSPI_CMD_TIMEOUT_US);
-		if (ret) {
-			log_err("cmd timeout (stat:%#x)\n", sr);
-		} else if (readl(&priv->regs->sr) & STM32_QSPI_SR_TEF) {
-			log_err("transfer error (stat:%#x)\n", sr);
-			ret = -EIO;
-		}
-		/* clear flags */
-		writel(STM32_QSPI_FCR_CTCF | STM32_QSPI_FCR_CTEF, &priv->regs->fcr);
+	ret = readl_poll_timeout(&priv->regs->sr, sr,
+				 sr & STM32_QSPI_SR_TCF,
+				 STM32_QSPI_CMD_TIMEOUT_US);
+	if (ret) {
+		log_err("cmd timeout (stat:%#x)\n", sr);
+	} else if (readl(&priv->regs->sr) & STM32_QSPI_SR_TEF) {
+		log_err("transfer error (stat:%#x)\n", sr);
+		ret = -EIO;
 	}
 
+	/* clear flags */
+	writel(STM32_QSPI_FCR_CTCF | STM32_QSPI_FCR_CTEF, &priv->regs->fcr);
+
 	if (!ret)
 		ret = _stm32_qspi_wait_for_not_busy(priv);
 
-- 
2.25.1

[PATCH v1 2/2] spi: stm32_qspi: Remove SR_BUSY bit check before sending command

[Patrice Chotard]

Waiting for SR_BUSY bit when receiving a new command is not needed.
SR_BUSY bit is already managed in the previous command treatment.

Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>

---

 drivers/spi/stm32_qspi.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/drivers/spi/stm32_qspi.c b/drivers/spi/stm32_qspi.c
index 3c8faecb54..ceba413727 100644
--- a/drivers/spi/stm32_qspi.c
+++ b/drivers/spi/stm32_qspi.c
@@ -255,10 +255,6 @@ static int stm32_qspi_exec_op(struct spi_slave *slave,
 		op->dummy.buswidth, op->data.buswidth,
 		op->addr.val, op->data.nbytes);
 
-	ret = _stm32_qspi_wait_for_not_busy(priv);
-	if (ret)
-		return ret;
-
 	addr_max = op->addr.val + op->data.nbytes + 1;
 
 	if (op->data.dir == SPI_MEM_DATA_IN && op->data.nbytes) {
-- 
2.25.1

Re: [PATCH v1 1/2] spi: stm32_qspi: Always check SR_TCF flags in stm32_qspi_wait_cmd()

[Patrick DELAUNAY]

Hi Patrice,

On 5/12/22 09:17, Patrice Chotard wrote:
> Currently, SR_TCF flag is checked in case there is data, this criteria
> is not correct.
>
> SR_TCF flags is set when programmed number of bytes have been transferred
> to the memory device ("bytes" comprised command and data send to the
> SPI device).
> So even if there is no data, we must check SR_TCF flag.
>
> Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
> ---
>
>   drivers/spi/stm32_qspi.c | 23 +++++++++++------------
>   1 file changed, 11 insertions(+), 12 deletions(-)
>
> diff --git a/drivers/spi/stm32_qspi.c b/drivers/spi/stm32_qspi.c
> index 8f4aabc3d1..3c8faecb54 100644
> --- a/drivers/spi/stm32_qspi.c
> +++ b/drivers/spi/stm32_qspi.c
> @@ -150,20 +150,19 @@ static int _stm32_qspi_wait_cmd(struct stm32_qspi_priv *priv,
>   	u32 sr;
>   	int ret = 0;
>   
> -	if (op->data.nbytes) {
> -		ret = readl_poll_timeout(&priv->regs->sr, sr,
> -					 sr & STM32_QSPI_SR_TCF,
> -					 STM32_QSPI_CMD_TIMEOUT_US);
> -		if (ret) {
> -			log_err("cmd timeout (stat:%#x)\n", sr);
> -		} else if (readl(&priv->regs->sr) & STM32_QSPI_SR_TEF) {
> -			log_err("transfer error (stat:%#x)\n", sr);
> -			ret = -EIO;
> -		}
> -		/* clear flags */
> -		writel(STM32_QSPI_FCR_CTCF | STM32_QSPI_FCR_CTEF, &priv->regs->fcr);
> +	ret = readl_poll_timeout(&priv->regs->sr, sr,
> +				 sr & STM32_QSPI_SR_TCF,
> +				 STM32_QSPI_CMD_TIMEOUT_US);
> +	if (ret) {
> +		log_err("cmd timeout (stat:%#x)\n", sr);
> +	} else if (readl(&priv->regs->sr) & STM32_QSPI_SR_TEF) {
> +		log_err("transfer error (stat:%#x)\n", sr);
> +		ret = -EIO;
>   	}
>   
> +	/* clear flags */
> +	writel(STM32_QSPI_FCR_CTCF | STM32_QSPI_FCR_CTEF, &priv->regs->fcr);
> +
>   	if (!ret)
>   		ret = _stm32_qspi_wait_for_not_busy(priv);
>   


Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com>

Thanks
Patrick

Re: [PATCH v1 2/2] spi: stm32_qspi: Remove SR_BUSY bit check before sending command

[Patrick DELAUNAY]

Hi Patrice

On 5/12/22 09:17, Patrice Chotard wrote:
> Waiting for SR_BUSY bit when receiving a new command is not needed.
> SR_BUSY bit is already managed in the previous command treatment.
>
> Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
>
> ---
>
>   drivers/spi/stm32_qspi.c | 4 ----
>   1 file changed, 4 deletions(-)
>
> diff --git a/drivers/spi/stm32_qspi.c b/drivers/spi/stm32_qspi.c
> index 3c8faecb54..ceba413727 100644
> --- a/drivers/spi/stm32_qspi.c
> +++ b/drivers/spi/stm32_qspi.c
> @@ -255,10 +255,6 @@ static int stm32_qspi_exec_op(struct spi_slave *slave,
>   		op->dummy.buswidth, op->data.buswidth,
>   		op->addr.val, op->data.nbytes);
>   
> -	ret = _stm32_qspi_wait_for_not_busy(priv);
> -	if (ret)
> -		return ret;
> -
>   	addr_max = op->addr.val + op->data.nbytes + 1;
>   
>   	if (op->data.dir == SPI_MEM_DATA_IN && op->data.nbytes) {


Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com>

Thanks
Patrick

Re: [Uboot-stm32] [PATCH v1 1/2] spi: stm32_qspi: Always check SR_TCF flags in stm32_qspi_wait_cmd()

[Patrick DELAUNAY]

Hi,

On 5/17/22 10:23, Patrick DELAUNAY wrote:
> Hi Patrice,
>
> On 5/12/22 09:17, Patrice Chotard wrote:
>> Currently, SR_TCF flag is checked in case there is data, this criteria
>> is not correct.
>>
>> SR_TCF flags is set when programmed number of bytes have been 
>> transferred
>> to the memory device ("bytes" comprised command and data send to the
>> SPI device).
>> So even if there is no data, we must check SR_TCF flag.
>>
>> Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
>> ---
>>
>>   drivers/spi/stm32_qspi.c | 23 +++++++++++------------
>>   1 file changed, 11 insertions(+), 12 deletions(-)
>>
>> diff --git a/drivers/spi/stm32_qspi.c b/drivers/spi/stm32_qspi.c
>> index 8f4aabc3d1..3c8faecb54 100644
>> --- a/drivers/spi/stm32_qspi.c
>> +++ b/drivers/spi/stm32_qspi.c
>> @@ -150,20 +150,19 @@ static int _stm32_qspi_wait_cmd(struct 
>> stm32_qspi_priv *priv,
>>       u32 sr;
>>       int ret = 0;
>>   -    if (op->data.nbytes) {
>> -        ret = readl_poll_timeout(&priv->regs->sr, sr,
>> -                     sr & STM32_QSPI_SR_TCF,
>> -                     STM32_QSPI_CMD_TIMEOUT_US);
>> -        if (ret) {
>> -            log_err("cmd timeout (stat:%#x)\n", sr);
>> -        } else if (readl(&priv->regs->sr) & STM32_QSPI_SR_TEF) {
>> -            log_err("transfer error (stat:%#x)\n", sr);
>> -            ret = -EIO;
>> -        }
>> -        /* clear flags */
>> -        writel(STM32_QSPI_FCR_CTCF | STM32_QSPI_FCR_CTEF, 
>> &priv->regs->fcr);
>> +    ret = readl_poll_timeout(&priv->regs->sr, sr,
>> +                 sr & STM32_QSPI_SR_TCF,
>> +                 STM32_QSPI_CMD_TIMEOUT_US);
>> +    if (ret) {
>> +        log_err("cmd timeout (stat:%#x)\n", sr);
>> +    } else if (readl(&priv->regs->sr) & STM32_QSPI_SR_TEF) {
>> +        log_err("transfer error (stat:%#x)\n", sr);
>> +        ret = -EIO;
>>       }
>>   +    /* clear flags */
>> +    writel(STM32_QSPI_FCR_CTCF | STM32_QSPI_FCR_CTEF, 
>> &priv->regs->fcr);
>> +
>>       if (!ret)
>>           ret = _stm32_qspi_wait_for_not_busy(priv);
>
>
> Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
>
> Thanks
> Patrick
>
> _______________________________________________
> Uboot-stm32 mailing list
> Uboot-stm32@st-md-mailman.stormreply.com
> https://st-md-mailman.stormreply.com/mailman/listinfo/uboot-stm32



Applied to u-boot-stm/master, thanks!

Regards
Patrick

Re: [PATCH v1 2/2] spi: stm32_qspi: Remove SR_BUSY bit check before sending command

[Patrick DELAUNAY]

Hi,

On 5/17/22 10:24, Patrick DELAUNAY wrote:
> Hi Patrice
>
> On 5/12/22 09:17, Patrice Chotard wrote:
>> Waiting for SR_BUSY bit when receiving a new command is not needed.
>> SR_BUSY bit is already managed in the previous command treatment.
>>
>> Signed-off-by: Patrice Chotard <patrice.chotard@foss.st.com>
>>
>> ---
>>
>>   drivers/spi/stm32_qspi.c | 4 ----
>>   1 file changed, 4 deletions(-)
>>
>> diff --git a/drivers/spi/stm32_qspi.c b/drivers/spi/stm32_qspi.c
>> index 3c8faecb54..ceba413727 100644
>> --- a/drivers/spi/stm32_qspi.c
>> +++ b/drivers/spi/stm32_qspi.c
>> @@ -255,10 +255,6 @@ static int stm32_qspi_exec_op(struct spi_slave 
>> *slave,
>>           op->dummy.buswidth, op->data.buswidth,
>>           op->addr.val, op->data.nbytes);
>>   -    ret = _stm32_qspi_wait_for_not_busy(priv);
>> -    if (ret)
>> -        return ret;
>> -
>>       addr_max = op->addr.val + op->data.nbytes + 1;
>>         if (op->data.dir == SPI_MEM_DATA_IN && op->data.nbytes) {
>
>
> Reviewed-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
>
> Thanks
> Patrick
>
>

Applied to u-boot-stm/master, thanks!

Regards
Patrick