On 30 May 2018 at 04:11, Shawn Lin <shawn.lin@xxxxxxxxxxxxxx> wrote:
> In preparation for reusing mmc_poll_for_busy() to avoid duplication
> of code for polling busy.
>
> No functional change intended.
>
> Signed-off-by: Shawn Lin <shawn.lin@xxxxxxxxxxxxxx>
> ---
>
> Changes in v3: None
> Changes in v2: None
>
> drivers/mmc/core/mmc_ops.c | 18 +++++++++++++++---
> 1 file changed, 15 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/mmc/core/mmc_ops.c b/drivers/mmc/core/mmc_ops.c
> index ee5f5ea..7e8be97 100644
> --- a/drivers/mmc/core/mmc_ops.c
> +++ b/drivers/mmc/core/mmc_ops.c
> @@ -447,7 +447,8 @@ int mmc_switch_status(struct mmc_card *card)
> }
>
> static int mmc_poll_for_busy(struct mmc_card *card, unsigned int timeout_ms,
> - bool send_status, bool retry_crc_err, bool use_r1b_resp)
> + bool send_status, bool retry_crc_err, bool use_r1b_resp,
> + u32 *resp_status, bool check_busy(u32 device_status))
Okay, so I like the idea that the caller should be give the
responsibility of checking error/status codes via a callback.
An option, could be to add that callback to the card struct. So,
instead of providing it as a parameter to a function the caller will
have to set the callback prior invoking mmc_poll_for_busy().
I think the code may become a bit nicer, but I guess it depends on
what one prefer. :-)
> {
> struct mmc_host *host = card->host;
> int err;
> @@ -461,6 +462,9 @@ static int mmc_poll_for_busy(struct mmc_card *card, unsigned int timeout_ms,
> mmc_host_is_spi(host))
> return 0;
>
> + if (WARN_ON(!check_busy))
> + return 0;
> +
> /* We have an unspecified cmd timeout, use the fallback value. */
> if (!timeout_ms)
> timeout_ms = MMC_OPS_TIMEOUT_MS;
> @@ -487,6 +491,9 @@ static int mmc_poll_for_busy(struct mmc_card *card, unsigned int timeout_ms,
> busy = host->ops->card_busy(host);
> } else {
> err = mmc_send_status(card, &status);
> + /* Accumulate any response error bits seen */
> + if (resp_status)
> + *resp_status |= status;
This makes sense, particular if we expect the caller to care about
several status bits, which may be set while iterating through the
loops. I assume that's the intent, right?
> if (retry_crc_err && err == -EILSEQ) {
> busy = true;
> } else if (err) {
> @@ -495,7 +502,7 @@ static int mmc_poll_for_busy(struct mmc_card *card, unsigned int timeout_ms,
> err = mmc_switch_status_error(host, status);
> if (err)
> return err;
> - busy = R1_CURRENT_STATE(status) == R1_STATE_PRG;
> + busy = check_busy(status);
I think this can be made more flexible, and also with the benefit of
keeping mmc_poll_for_busy() as simple as possible.
Instead of making the callback to check for the card status only,
wouldn't it be better if had the responsibility to check the entire
error/busy thingy?
In principle it should mean we get lesser in-parameters to
mmc_poll_for_busy() rather than more.
Thoughts?
> }
> }
>
> @@ -510,6 +517,11 @@ static int mmc_poll_for_busy(struct mmc_card *card, unsigned int timeout_ms,
> return 0;
> }
>
> +static inline bool mmc_switch_in_prg_state(u32 status)
> +{
> + return R1_CURRENT_STATE(status) == R1_STATE_PRG;
> +}
> +
> /**
> * __mmc_switch - modify EXT_CSD register
> * @card: the MMC card associated with the data transfer
> @@ -577,7 +589,7 @@ int __mmc_switch(struct mmc_card *card, u8 set, u8 index, u8 value,
>
> /* Let's try to poll to find out when the command is completed. */
> err = mmc_poll_for_busy(card, timeout_ms, send_status, retry_crc_err,
> - use_r1b_resp);
> + use_r1b_resp, NULL, &mmc_switch_in_prg_state);
> if (err)
> goto out;
>
Kind regards
Uffe
--
To unsubscribe from this list: send the line "unsubscribe linux-mmc" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
