2020-08-14 21:01:10

by Tobias Schramm

[permalink] [raw]
Subject: [PATCH] mmc: mmc_spi: fix timeout calculation

Previously the cycle timeout was converted to a microsecond value but
then incorrectly treated as a nanosecond timeout. This patch changes
the code to convert both the nanosecond timeout and the cycle timeout
to a microsecond value and use that directly.

Signed-off-by: Tobias Schramm <[email protected]>
---
drivers/mmc/host/mmc_spi.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/mmc/host/mmc_spi.c b/drivers/mmc/host/mmc_spi.c
index 39bb1e30c2d7..f85e0ad896a9 100644
--- a/drivers/mmc/host/mmc_spi.c
+++ b/drivers/mmc/host/mmc_spi.c
@@ -882,9 +882,9 @@ mmc_spi_data_do(struct mmc_spi_host *host, struct mmc_command *cmd,
else
clock_rate = spi->max_speed_hz;

- timeout = data->timeout_ns +
+ timeout = data->timeout_ns / 1000 +
data->timeout_clks * 1000000 / clock_rate;
- timeout = usecs_to_jiffies((unsigned int)(timeout / 1000)) + 1;
+ timeout = usecs_to_jiffies((unsigned int)timeout) + 1;

/* Handle scatterlist segments one at a time, with synch for
* each 512-byte block
--
2.28.0


2020-08-21 08:47:23

by Ulf Hansson

[permalink] [raw]
Subject: Re: [PATCH] mmc: mmc_spi: fix timeout calculation

On Fri, 14 Aug 2020 at 20:50, Tobias Schramm <[email protected]> wrote:
>
> Previously the cycle timeout was converted to a microsecond value but
> then incorrectly treated as a nanosecond timeout. This patch changes
> the code to convert both the nanosecond timeout and the cycle timeout
> to a microsecond value and use that directly.
>
> Signed-off-by: Tobias Schramm <[email protected]>

This looks good to me, but before applying just wanted to check that
you tested this on some HW, to make sure it doesn't break anything?

Kind regards
Uffe


> ---
> drivers/mmc/host/mmc_spi.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/mmc/host/mmc_spi.c b/drivers/mmc/host/mmc_spi.c
> index 39bb1e30c2d7..f85e0ad896a9 100644
> --- a/drivers/mmc/host/mmc_spi.c
> +++ b/drivers/mmc/host/mmc_spi.c
> @@ -882,9 +882,9 @@ mmc_spi_data_do(struct mmc_spi_host *host, struct mmc_command *cmd,
> else
> clock_rate = spi->max_speed_hz;
>
> - timeout = data->timeout_ns +
> + timeout = data->timeout_ns / 1000 +
> data->timeout_clks * 1000000 / clock_rate;
> - timeout = usecs_to_jiffies((unsigned int)(timeout / 1000)) + 1;
> + timeout = usecs_to_jiffies((unsigned int)timeout) + 1;
>
> /* Handle scatterlist segments one at a time, with synch for
> * each 512-byte block
> --
> 2.28.0
>

2020-08-23 23:10:14

by Tobias Schramm

[permalink] [raw]
Subject: Re: [PATCH] mmc: mmc_spi: fix timeout calculation

Hi Uffe,

> This looks good to me, but before applying just wanted to check that
> you tested this on some HW, to make sure it doesn't break anything?

yes, I should have mentioned that. I tested the change on a custom
STM32H743 board with a microSD card connected to one of its SPIs. It
does still work just fine with this patch applied.

Tobias

2020-08-24 09:52:59

by Ulf Hansson

[permalink] [raw]
Subject: Re: [PATCH] mmc: mmc_spi: fix timeout calculation

On Fri, 14 Aug 2020 at 20:50, Tobias Schramm <[email protected]> wrote:
>
> Previously the cycle timeout was converted to a microsecond value but
> then incorrectly treated as a nanosecond timeout. This patch changes
> the code to convert both the nanosecond timeout and the cycle timeout
> to a microsecond value and use that directly.
>
> Signed-off-by: Tobias Schramm <[email protected]>

Applied for next, thanks!

Kind regards
Uffe


> ---
> drivers/mmc/host/mmc_spi.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/mmc/host/mmc_spi.c b/drivers/mmc/host/mmc_spi.c
> index 39bb1e30c2d7..f85e0ad896a9 100644
> --- a/drivers/mmc/host/mmc_spi.c
> +++ b/drivers/mmc/host/mmc_spi.c
> @@ -882,9 +882,9 @@ mmc_spi_data_do(struct mmc_spi_host *host, struct mmc_command *cmd,
> else
> clock_rate = spi->max_speed_hz;
>
> - timeout = data->timeout_ns +
> + timeout = data->timeout_ns / 1000 +
> data->timeout_clks * 1000000 / clock_rate;
> - timeout = usecs_to_jiffies((unsigned int)(timeout / 1000)) + 1;
> + timeout = usecs_to_jiffies((unsigned int)timeout) + 1;
>
> /* Handle scatterlist segments one at a time, with synch for
> * each 512-byte block
> --
> 2.28.0
>