Hi Grygorii,

On Wednesday 29 July 2015 05:39 PM, Grygorii Strashko wrote:

On 07/29/2015 02:09 PM, Kishon Vijay Abraham I wrote:

quoted

Since vmmc can be optional for some platforms, use
devm_regulator_get_optional() for vmmc. Now return error only
in the case of -EPROBE_DEFER and for all other cases set
host->vcc to NULL.

Signed-off-by: Kishon Vijay Abraham I <redacted>
---
  drivers/mmc/host/omap_hsmmc.c |    8 +++++---
  1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/mmc/host/omap_hsmmc.c b/drivers/mmc/host/omap_hsmmc.c
index 4d12032..b673e59 100644
--- a/drivers/mmc/host/omap_hsmmc.c
+++ b/drivers/mmc/host/omap_hsmmc.c

@@ -344,11 +344,13 @@ static int omap_hsmmc_reg_get(struct omap_hsmmc_host *host)
  	struct regulator *reg;
  	int ocr_value = 0;
  
-	reg = devm_regulator_get(host->dev, "vmmc");
+	reg = devm_regulator_get_optional(host->dev, "vmmc");
  	if (IS_ERR(reg)) {
-		dev_err(host->dev, "unable to get vmmc regulator %ld\n",
+		if (PTR_ERR(reg) == -EPROBE_DEFER)
+			return -EPROBE_DEFER;
+		host->vcc = NULL;
+		dev_dbg(host->dev, "unable to get vmmc regulator %ld\n",
  			PTR_ERR(reg));
-		return PTR_ERR(reg);

I think, It could be unsafe to just drop this return.
regulator_get_optional may return:
1 valid pointer on regulator : success;
2 ERR_PTR(-ENODEV)           : regulator is not assigned, can proceed.
3 ERR_PTR(-EPROBE_DEFER)     : regulator is assigned, but not ready yet, retry.
4 ERR_PTR(<other error codes>: regulator is assigned, but can't be retrieved, failure, can't proceed

So, It's allowed to continue with host->vcc = NULL; only in case [2], while
in case [4] probe should fail.

You are right. Will fix it and resend the patch.

Thanks
Kishon