+ Adrian
On Fri, 18 Jun 2021 at 10:23, Wolfram Sang
<wsa+renesas@xxxxxxxxxxxxxxxxxxxx> wrote:
>
> When we use the alive callback, we expect a command to fail if the card
> is not present. We should not trigger a retune then which will confuse
> users with a failed retune on a removed card:
>
> mmc2: tuning execution failed: -5
> mmc2: card 0001 removed
>
> Disable retuning in this code path.
>
> Reported-by: Ulrich Hecht <uli+renesas@xxxxxxxx>
> Signed-off-by: Wolfram Sang <wsa+renesas@xxxxxxxxxxxxxxxxxxxx>
> ---
> drivers/mmc/core/core.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/drivers/mmc/core/core.c b/drivers/mmc/core/core.c
> index 54f0814f110c..eb792dd845a3 100644
> --- a/drivers/mmc/core/core.c
> +++ b/drivers/mmc/core/core.c
> @@ -2088,6 +2088,9 @@ int _mmc_detect_card_removed(struct mmc_host *host)
> if (!host->card || mmc_card_removed(host->card))
> return 1;
>
> + /* we expect a failure if the card is removed */
> + mmc_retune_disable(host);
> +
Some controllers require a retune after it has been runtime suspended.
In the above path, when called via the bus_ops->detect() callback, it
could be that the controller may have been runtime suspended and then
got resumed by the call to mmc_get_card().
I think we need something more clever here, to make sure we don't end
up in that situation. I have looped in Adrian, to see if has some
ideas for how this can be fixed.
> ret = host->bus_ops->alive(host);
>
> /*
> @@ -2107,6 +2110,8 @@ int _mmc_detect_card_removed(struct mmc_host *host)
> pr_debug("%s: card remove detected\n", mmc_hostname(host));
> }
>
> + mmc_retune_enable(host);
> +
> return ret;
> }
>
> --
> 2.30.2
>
Kind regards
Uffe
