Re: [PATCH] mmc: core: fix __mmc_switch timeout caused by preempt

[Ulf Hansson]

On 26 November 2015 at 03:51, Chaotian Jing <chaotian.jing@xxxxxxxxxxxx> wrote:
> there is a time window between __mmc_send_status() and time_afer(),
> on some eMMC chip, the timeout_ms is only 10ms, if this thread was
> scheduled out during this period, then, even card has already changes
> to transfer state by the result of CMD13, this part of code also treat
> it to timeout error.
> So, need calculate timeout first, then call __mmc_send_status(), if
> already timeout and card still in programing state, then treat it to
> the real timeout error.
>
> Signed-off-by: Chaotian Jing <chaotian.jing@xxxxxxxxxxxx>
> ---
>  drivers/mmc/core/mmc_ops.c | 23 ++++++++++++++---------
>  1 file changed, 14 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/mmc/core/mmc_ops.c b/drivers/mmc/core/mmc_ops.c
> index 1f44426..0add634 100644
> --- a/drivers/mmc/core/mmc_ops.c
> +++ b/drivers/mmc/core/mmc_ops.c
> @@ -489,6 +489,7 @@ int __mmc_switch(struct mmc_card *card, u8 set, u8 index, u8 value,
>         unsigned long timeout;
>         u32 status = 0;
>         bool use_r1b_resp = use_busy_signal;
> +       bool expired = false;

No need to assign an initial value.

>
>         mmc_retune_hold(host);
>
> @@ -544,10 +545,22 @@ int __mmc_switch(struct mmc_card *card, u8 set, u8 index, u8 value,
>         /* Must check status to be sure of no errors. */
>         timeout = jiffies + msecs_to_jiffies(timeout_ms);
>         do {
> +               /* Timeout if the device never leaves the program state.

Please start mutli-row comment with an empty line.

> +                * must check it firstly to avoid problem caused by preempt.

Please rephrase this to:

Due to the possibility of being preempted after sending the status
command, check the expiration time first.

> +                */
> +               expired = time_after(jiffies, timeout);

I think we can move this inside the below "if (send_status) {"
.. but of course before sending the status command.

> +
>                 if (send_status) {
>                         err = __mmc_send_status(card, &status, ignore_crc);
> -                       if (err)
> +                       if (err) {
>                                 goto out;
> +                       } else if (expired &&

You may use a new "if" statement instead of an "else if", due to the
above "goto out".

> +                                  R1_CURRENT_STATE(status) == R1_STATE_PRG) {
> +                               pr_err("%s: Card stuck in programming state! %s\n",
> +                                      mmc_hostname(host), __func__);
> +                               err = -ETIMEDOUT;
> +                               goto out;
> +                       }
>                 }
>                 if ((host->caps & MMC_CAP_WAIT_WHILE_BUSY) && use_r1b_resp)
>                         break;
> @@ -563,14 +576,6 @@ int __mmc_switch(struct mmc_card *card, u8 set, u8 index, u8 value,
>                         mmc_delay(timeout_ms);
>                         goto out;
>                 }
> -
> -               /* Timeout if the device never leaves the program state. */
> -               if (time_after(jiffies, timeout)) {
> -                       pr_err("%s: Card stuck in programming state! %s\n",
> -                               mmc_hostname(host), __func__);
> -                       err = -ETIMEDOUT;
> -                       goto out;
> -               }
>         } while (R1_CURRENT_STATE(status) == R1_STATE_PRG);
>
>         err = mmc_switch_status_error(host, status);
> --
> 1.8.1.1.dirty
>

Kind regards
Uffe
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/