mmc: sdhci: fix IS_ERR() checking of regulator_get()

There are two problems here:

The check for vmmc was printing an unnecessary pr_info() when
host->vmmc is NULL.

The intent of the check for vqmmc was to only remove UHS if we have a
regulator that doesn't support the required voltage, but since IS_ERR()
doesn't catch NULL, we were actually removing UHS modes if vqmmc isn't
present at all -- since it isn't present for most users, this breaks
UHS for them.  This patch fixes that UHS regression in 3.7-rc1.

Signed-off-by: Kevin Liu <[email protected]>
Signed-off-by: Bin Wang <[email protected]>
Reviewed-by: Philip Rakity <[email protected]>
Signed-off-by: Chris Ball <[email protected]>

0 files changed, 0 insertions, 0 deletions