On Wed, Dec 12, 2012 at 04:50:21PM +0100, Ulf Hansson wrote: > diff --git a/drivers/mmc/host/mmci.c b/drivers/mmc/host/mmci.c > index edc3e9b..bf07823 100644 > --- a/drivers/mmc/host/mmci.c > +++ b/drivers/mmc/host/mmci.c > @@ -1147,6 +1147,15 @@ static void mmci_set_ios(struct mmc_host *mmc, struct mmc_ios *ios) > } > } > > + /* > + * If clock = 0 and the block needs a certain value in the clreg > + * to function, we need to gate the clock by removing MCI_PWR_ON. > + * This is a special case for ST Micro variants, since the power > + * register in the ARM legacy case is used for powering the cards. I wish you'd stop using "legacy" when talking about the ARM primecell. How can it be legacy when ARMs current development boards still use this primecell? > + */ > + if (!ios->clock && variant->clkreg) > + pwr &= ~MCI_PWR_ON; This is horrid. You're overloading the clkreg default value with another meaning here - "if we have any bits set in the default value for the MCICLOCK register, the MCIPOWER register has special meanings for the power control bits". When you think about it as I've described it above, do you think this is an acceptable solution to this problem? -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
