From: David Hendricks <[email protected]>
To avoid spamming the EC we calculate the time between the previous
transfer and the current transfer and force a delay if the time delta
is too small.
However, a small miscalculation causes the delay period to be
far too short. Most noticably this impacts commands with a long
turnaround time such as EC firmware reads and writes.
Signed-off-by: David Hendricks <[email protected]>
Signed-off-by: Doug Anderson <[email protected]>
---
drivers/mfd/cros_ec_spi.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/mfd/cros_ec_spi.c b/drivers/mfd/cros_ec_spi.c
index 84af8d7..c185eb6 100644
--- a/drivers/mfd/cros_ec_spi.c
+++ b/drivers/mfd/cros_ec_spi.c
@@ -219,7 +219,7 @@ static int cros_ec_command_spi_xfer(struct cros_ec_device *ec_dev,
ktime_get_ts(&ts);
delay = timespec_to_ns(&ts) - ec_spi->last_transfer_ns;
if (delay < EC_SPI_RECOVERY_TIME_NS)
- ndelay(delay);
+ ndelay(EC_SPI_RECOVERY_TIME_NS - delay);
}
/* Transmit phase - send our message */
--
1.9.1.423.g4596e3a
On 17 April 2014 11:59, Doug Anderson <[email protected]> wrote:
> From: David Hendricks <[email protected]>
>
> To avoid spamming the EC we calculate the time between the previous
> transfer and the current transfer and force a delay if the time delta
> is too small.
>
> However, a small miscalculation causes the delay period to be
> far too short. Most noticably this impacts commands with a long
> turnaround time such as EC firmware reads and writes.
>
> Signed-off-by: David Hendricks <[email protected]>
> Signed-off-by: Doug Anderson <[email protected]>
Reviewed-by: Simon Glass <[email protected]>