2024-01-23 15:36:41

by Tudor Ambarus

Subject: [PATCH 08/21] spi: s3c64xx: move error check up to avoid rechecking

Check the return value of wait_for_completion_timeout() immediately
after the call so that we avoid checking the return value twice.

Signed-off-by: Tudor Ambarus <[email protected]>
---
drivers/spi/spi-s3c64xx.c | 13 ++++---------
1 file changed, 4 insertions(+), 9 deletions(-)

diff --git a/drivers/spi/spi-s3c64xx.c b/drivers/spi/spi-s3c64xx.c
index 16eea56892a2..128c3b8211ce 100644
--- a/drivers/spi/spi-s3c64xx.c
+++ b/drivers/spi/spi-s3c64xx.c
@@ -519,17 +519,17 @@ static int s3c64xx_wait_for_dma(struct s3c64xx_spi_driver_data *sdd,

val = msecs_to_jiffies(ms) + 10;
val = wait_for_completion_timeout(&sdd->xfer_completion, val);
-
+ if (!val)
+ return -EIO;
/*
- * If the previous xfer was completed within timeout, then
- * proceed further else return -EIO.
+ * If the previous xfer was completed within timeout proceed further.
* DmaTx returns after simply writing data in the FIFO,
* w/o waiting for real transmission on the bus to finish.
* DmaRx returns only after Dma read data from FIFO which
* needs bus transmission to finish, so we don't worry if
* Xfer involved Rx(with or without Tx).
*/
- if (val && !xfer->rx_buf) {
+ if (!xfer->rx_buf) {
val = msecs_to_loops(10);
status = readl(regs + S3C64XX_SPI_STATUS);
while ((TX_FIFO_LVL(status, sdd)
@@ -538,13 +538,8 @@ static int s3c64xx_wait_for_dma(struct s3c64xx_spi_driver_data *sdd,
cpu_relax();
status = readl(regs + S3C64XX_SPI_STATUS);
}
-
}

- /* If timed out while checking rx/tx status return error */
- if (!val)
- return -EIO;
-
return 0;
}

--
2.43.0.429.g432eaa2c6b-goog

2024-01-24 09:36:16

by Tudor Ambarus

Subject: Re: [PATCH 08/21] spi: s3c64xx: move error check up to avoid rechecking

On 1/24/24 09:21, André Draszik wrote:
> Hi Tudor,
>

Hi!

> On Tue, 2024-01-23 at 15:34 +0000, Tudor Ambarus wrote:
>> @@ -538,13 +538,8 @@ static int s3c64xx_wait_for_dma(struct s3c64xx_spi_driver_data *sdd,
>>   cpu_relax();
>>   status = readl(regs + S3C64XX_SPI_STATUS);
>>   }
>> -
>>   }
>>  
>> - /* If timed out while checking rx/tx status return error */
>> - if (!val)
>> - return -EIO;
>> -
>
> This change behaviour of this function. The loop just above adjusts val and it is used to
> determine if there was a timeout or not:
>
> if (val && !xfer->rx_buf) {
> val = msecs_to_loops(10);
> status = readl(regs + S3C64XX_SPI_STATUS);
> while ((TX_FIFO_LVL(status, sdd)
> || !S3C64XX_SPI_ST_TX_DONE(status, sdd))
> && --val) {
> cpu_relax();
> status = readl(regs + S3C64XX_SPI_STATUS);
> }
>
Oh, yes, the timeout in this block. You're right, I'll drop the patch.
Thanks!

ta

2024-01-24 09:51:31

by André Draszik

Subject: Re: [PATCH 08/21] spi: s3c64xx: move error check up to avoid rechecking

Hi Tudor,

On Tue, 2024-01-23 at 15:34 +0000, Tudor Ambarus wrote:
> @@ -538,13 +538,8 @@ static int s3c64xx_wait_for_dma(struct s3c64xx_spi_driver_data *sdd,
>   cpu_relax();
>   status = readl(regs + S3C64XX_SPI_STATUS);
>   }
> -
>   }
>  
> - /* If timed out while checking rx/tx status return error */
> - if (!val)
> - return -EIO;
> -

This change behaviour of this function. The loop just above adjusts val and it is used to
determine if there was a timeout or not:

if (val && !xfer->rx_buf) {
val = msecs_to_loops(10);
status = readl(regs + S3C64XX_SPI_STATUS);
while ((TX_FIFO_LVL(status, sdd)
|| !S3C64XX_SPI_ST_TX_DONE(status, sdd))
&& --val) {
cpu_relax();
status = readl(regs + S3C64XX_SPI_STATUS);
}


That doesn't work anymore now.

Cheers,
Andre'