Hi, Alim

Sorry for late reply.

On 2015/2/16 07:28, Alim Akhtar wrote:

Hi Addy,

On Sat, Feb 14, 2015 at 11:47 AM, Addy Ke [off-list ref] wrote:

quoted

As show in mmc_power_up(), in MMC_POWER_UP state, the voltage isn't
stable and we may get 'data busy' which can't be cleaned by resetting
all blocks. So we should not send command to update clock in this state.

Signed-off-by: Addy Ke <redacted>
---
 drivers/mmc/host/dw_mmc.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/mmc/host/dw_mmc.c b/drivers/mmc/host/dw_mmc.c
index 4d2e3c2..3472f9b 100644
--- a/drivers/mmc/host/dw_mmc.c
+++ b/drivers/mmc/host/dw_mmc.c

@@ -1102,7 +1102,8 @@ static void dw_mci_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
                drv_data->set_ios(slot->host, ios);

        /* Slot specific timing and width adjustment */
-       dw_mci_setup_bus(slot, false);
+       if (ios->power_mode != MMC_POWER_UP)
+               dw_mci_setup_bus(slot, false);

This looks a HACK to me.
If stabilizing host voltage regulator is the problem, can you try out
below patch, and see if this resolve your issue?

I have test by:
cd /sys/bus/platform/drivers/dwmmc_rockchip
for i in $(seq 1 10000); do
  echo "========================" $i
  echo ff0c0000.dwmmc > unbind
  sleep .5
  echo ff0c0000.dwmmc > bind
  sleep 2
done

There is no error.
I think this patch can resolve my issue, thank you.

Do you send this patch upstream, or can I put it in my patch list?