Hi Marek, > Marek Vasut <marex@xxxxxxx> hat am 6. August 2016 um 15:13 geschrieben: > > > On 08/06/2016 02:55 PM, Stefan Wahren wrote: > > This patch implements driver support for 3.3V DDR eMMCs. > > > > Signed-off-by: Stefan Wahren <stefan.wahren@xxxxxxxx> > > --- > > drivers/mmc/host/mxs-mmc.c | 22 ++++++++++++++++++++++ > > 1 file changed, 22 insertions(+) > > > > diff --git a/drivers/mmc/host/mxs-mmc.c b/drivers/mmc/host/mxs-mmc.c > > index d839147..84019d5 100644 > > --- a/drivers/mmc/host/mxs-mmc.c > > +++ b/drivers/mmc/host/mxs-mmc.c > > @@ -71,6 +71,7 @@ struct mxs_mmc_host { > > spinlock_t lock; > > int sdio_irq_en; > > bool broken_cd; > > + bool is_ddr; > > }; > > > > static int mxs_mmc_get_cd(struct mmc_host *mmc) > > @@ -411,6 +412,9 @@ static void mxs_mmc_adtc(struct mxs_mmc_host *host) > > cmd0 |= BF_SSP(log2_blksz, CMD0_BLOCK_SIZE) | > > BF_SSP(blocks - 1, CMD0_BLOCK_COUNT); > > } else { > > + if (host->is_ddr) > > + cmd0 |= BM_SSP_CMD0_DBL_DATA_RATE_EN; > > + > > writel(data_size, ssp->base + HW_SSP_XFER_SIZE); > > writel(BF_SSP(log2_blksz, BLOCK_SIZE_BLOCK_SIZE) | > > BF_SSP(blocks - 1, BLOCK_SIZE_BLOCK_COUNT), > > @@ -499,6 +503,7 @@ static void mxs_mmc_request(struct mmc_host *mmc, struct > > mmc_request *mrq) > > static void mxs_mmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios) > > { > > struct mxs_mmc_host *host = mmc_priv(mmc); > > + struct mxs_ssp *ssp = &host->ssp; > > > > if (ios->bus_width == MMC_BUS_WIDTH_8) > > host->bus_width = 2; > > @@ -509,6 +514,23 @@ static void mxs_mmc_set_ios(struct mmc_host *mmc, > > struct mmc_ios *ios) > > > > if (ios->clock) > > mxs_ssp_set_clk_rate(&host->ssp, ios->clock); > > + > > + if (ssp_is_old(ssp)) > > + return; > > + > > + if (ios->timing == MMC_TIMING_MMC_DDR52) { > > Shouldn't you validate that the clock are set to at least 52MHz before > neabling the DDR mode ? according to the i.MX28 datasheet IMX28CEC 3.5.14.2 MMC4.4 (Dual Data Rate) AC Timing the minimum clock frequency is 0 and maximum is 52 MHz. So i assume the core take care of the right clock rate. > > > + /* > > + * ENGR00133481-1: In DDR mode the host send the data at > > + * negative edge and the MMC receive the data at positive edge. > > + */ > > + host->is_ddr = true; > > + writel(BM_SSP_CTRL1_POLARITY, ssp->base + > > + HW_SSP_CTRL1(ssp) + STMP_OFFSET_REG_CLR); > > + } else { > > Is it by any chance possible that this else branch will catch not only > the non-DDR options, but also some DDR ones , thus causing problems ? No, because the polarity bit has been set all the time by mxs_mmc_reset(). But it's possible that i didn't catch all DDR cases. > > > + host->is_ddr = false; > > + writel(BM_SSP_CTRL1_POLARITY, ssp->base + HW_SSP_CTRL1(ssp) + > > + STMP_OFFSET_REG_SET); > > + } > > } > > > > static void mxs_mmc_enable_sdio_irq(struct mmc_host *mmc, int enable) > > > > > -- > Best regards, > Marek Vasut -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html