mmc: core: Restore parts of the polling policy when switch to HS/HS DDR

Regressions for not being able to detect an eMMC HS DDR mode card has been
reported for the sdhci-esdhc-imx driver, but potentially other sdhci
variants may suffer from the similar problem.

The commit e173f8911f09 ("mmc: core: Update CMD13 polling policy when
switch to HS DDR mode"), is causing the problem. It seems that change moved
one step to far, regarding changing the host's timing before polling for a
busy card.

To fix this, let's move back to the behaviour when the host's timing is
updated after the polling, but before the switch status is fetched and
validated.

In cases when polling with CMD13, we keep validating the switch status at
each attempt. However, to align with the other card busy detections
mechanism, let's fetch and validate the switch status also after the host's
timing is updated.

Reported-by: Clemens Gruber <clemens.gruber@pqgruber.com>
Reported-by: Gary Bisson <gary.bisson@boundarydevices.com>
Fixes: e173f8911f09 ("mmc: core: Update CMD13 polling policy when switch..")
Cc: Shawn Lin <shawn.lin@rock-chips.com>
Cc: Dong Aisheng <aisheng.dong@nxp.com>
Cc: Haibo Chen <haibo.chen@nxp.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Tested-by: Clemens Gruber <clemens.gruber@pqgruber.com>
Tested-by: Jagan Teki <jagan@amarulasolutions.com>
Reviewed-by: Shawn Lin <shawn.lin@rock-chips.com>
Tested-by: Haibo Chen <haibo.chen@nxp.com>
Reviewed-by: Dong Aisheng <aisheng.dong@nxp.com>

1 files changed, 12 insertions, 13 deletions

diff --git a/drivers/mmc/core/mmc_ops.c b/drivers/mmc/core/mmc_ops.c
index b11c3455b040..e6ea8503f40c 100644
--- a/drivers/mmc/core/mmc_ops.c
+++ b/drivers/mmc/core/mmc_ops.c
@@ -506,9 +506,6 @@ static int mmc_poll_for_busy(struct mmc_card *card, unsigned int timeout_ms,
                 }
         } while (busy);
 
-        if (host->ops->card_busy && send_status)
-                return mmc_switch_status(card);
-
         return 0;
 }
 
@@ -577,24 +574,26 @@ int __mmc_switch(struct mmc_card *card, u8 set, u8 index, u8 value,
         if (!use_busy_signal)
                 goto out;
 
-        /* Switch to new timing before poll and check switch status. */
-        if (timing)
-                mmc_set_timing(host, timing);
-
         /*If SPI or used HW busy detection above, then we don't need to poll. */
         if (((host->caps & MMC_CAP_WAIT_WHILE_BUSY) && use_r1b_resp) ||
-                mmc_host_is_spi(host)) {
-                if (send_status)
-                        err = mmc_switch_status(card);
+                mmc_host_is_spi(host))
                 goto out_tim;
-        }
 
         /* 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);
+        if (err)
+                goto out;
 
 out_tim:
-        if (err && timing)
-                mmc_set_timing(host, old_timing);
+        /* Switch to new timing before check switch status. */
+        if (timing)
+                mmc_set_timing(host, timing);
+
+        if (send_status) {
+                err = mmc_switch_status(card);
+                if (err && timing)
+                        mmc_set_timing(host, old_timing);
+        }
 out:
         mmc_retune_release(host);