On Mon, Nov 26, 2012 at 01:06:26PM +0900, Magnus Damm wrote:

Hi Simon,

On Mon, Nov 26, 2012 at 9:16 AM, Simon Horman [off-list ref] wrote:

quoted

Use MMC_VDD_165_195 as the allowed voltages in the absence of
platform data. This should allow device tree to initialise MMCIF
for a number of boards.

A more complete solution would be to provide either MMCIF or MMC device
try bindings to set the allowed voltages.

Tested on the KZM-A9-GT board

Cc: Guennadi Liakhovetski <redacted>
Signed-off-by: Simon Horman <horms@verge.net.au>
---
 drivers/mmc/host/sh_mmcif.c |    7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

Thanks for your patch!

quoted

diff --git a/drivers/mmc/host/sh_mmcif.c b/drivers/mmc/host/sh_mmcif.c
index 11d2bc3..88d6866 100644
--- a/drivers/mmc/host/sh_mmcif.c
+++ b/drivers/mmc/host/sh_mmcif.c

@@ -1293,12 +1293,9 @@ static void sh_mmcif_init_ocr(struct sh_mmcif_host *host)

        mmc_regulator_get_supply(mmc);

-       if (!pd)
-               return;
-
        if (!mmc->ocr_avail)
-               mmc->ocr_avail = pd->ocr;
-       else if (pd->ocr)
+               mmc->ocr_avail = pd ? pd->ocr : MMC_VDD_165_195;
+       else if (pd && pd->ocr)
                dev_warn(mmc_dev(mmc), "Platform OCR mask is ignored\n");
 }

I may be mistaken, but I somehow thought that the code in
mmc_regulator_get_supply() would take care of this automatically for
us already. So this kind of patch isn't really needed as long as you
tie in regulators via DT as expected.

Thanks, I'll check that. I did write this patch before hooking
up the regulators correctly.