Re: [PATCH] mmc: core: increase the number of retries of the ACMD41

[Dragan Simic <dsimic@manjaro.org>]

Hello Felix,

On 2024-04-16 16:59, Felix Qin wrote:
> Extensive testing has shown that a few specific SD cards require
> an increase in the number of command retries to successfully
> initialize.

Any chances, please, to include the models of the affected  SD cards
in the patch description, together with the model of the used host
hardware?  That might be helpful to the people wanting to replicate
(or debug further) such card initialization issues.

> Signed-off-by: Felix Qin <xiaokeqinhealth@126.com>
> ---
>  drivers/mmc/core/sd_ops.c | 4 ++--
>  drivers/mmc/core/sd_ops.h | 8 ++++++++
>  2 files changed, 10 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/mmc/core/sd_ops.c b/drivers/mmc/core/sd_ops.c
> index a59cd592f06e..b3b6e8725b80 100644
> --- a/drivers/mmc/core/sd_ops.c
> +++ b/drivers/mmc/core/sd_ops.c
> @@ -127,7 +127,7 @@ int mmc_send_app_op_cond(struct mmc_host *host,
> u32 ocr, u32 *rocr)
>  		cmd.arg = ocr;
>  	cmd.flags = MMC_RSP_SPI_R1 | MMC_RSP_R3 | MMC_CMD_BCR;
> 
> -	for (i = 100; i; i--) {
> +	for (i = SD_APP_OP_COND_RETRIES; i; i--) {
>  		err = mmc_wait_for_app_cmd(host, NULL, &cmd);
>  		if (err)
>  			break;
> @@ -147,7 +147,7 @@ int mmc_send_app_op_cond(struct mmc_host *host,
> u32 ocr, u32 *rocr)
> 
>  		err = -ETIMEDOUT;
> 
> -		mmc_delay(10);
> +		mmc_delay(SD_APP_OP_COND_INTERVAL);
>  	}
> 
>  	if (!i)
> diff --git a/drivers/mmc/core/sd_ops.h b/drivers/mmc/core/sd_ops.h
> index 7667fc223b74..80c5c836ccd1 100644
> --- a/drivers/mmc/core/sd_ops.h
> +++ b/drivers/mmc/core/sd_ops.h
> @@ -10,6 +10,14 @@
> 
>  #include <linux/types.h>
> 
> +/*
> + * Extensive testing has shown that a few specific SD cards require
> + * an increase in the number of command retries to successfully
> + * initialize.
> + */
> +#define SD_APP_OP_COND_RETRIES  200
> +#define SD_APP_OP_COND_INTERVAL 10	/* millisecond */
> +
>  struct mmc_card;
>  struct mmc_host;