diff --git a/drivers/mmc/core/mmc.c b/drivers/mmc/core/mmc.c
index 4ff08ea930a6..2510cebd670f 100644
--- a/drivers/mmc/core/mmc.c
+++ b/drivers/mmc/core/mmc.c
@@ -30,6 +30,7 @@
 #include "pwrseq.h"
 
 #define DEFAULT_CMD6_TIMEOUT_MS	500
+#define MIN_CACHE_EN_TIMEOUT_MS 1600
 
 static const unsigned int tran_exp[] = {
 	10000,		100000,		1000000,	10000000,
@@ -526,8 +527,7 @@ static int mmc_decode_ext_csd(struct mmc_card *card, u8 *ext_csd)
 			card->cid.year += 16;
 
 		/* check whether the eMMC card supports BKOPS */
-		if (!mmc_card_broken_hpi(card) &&
-		    ext_csd[EXT_CSD_BKOPS_SUPPORT] & 0x1) {
+		if (ext_csd[EXT_CSD_BKOPS_SUPPORT] & 0x1) {
 			card->ext_csd.bkops = 1;
 			card->ext_csd.man_bkops_en =
 					(ext_csd[EXT_CSD_BKOPS_EN] &
@@ -1785,20 +1785,26 @@ static int mmc_init_card(struct mmc_host *host, u32 ocr,
 		if (err) {
 			pr_warn("%s: Enabling HPI failed\n",
 				mmc_hostname(card->host));
+			card->ext_csd.hpi_en = 0;
 			err = 0;
-		} else
+		} else {
 			card->ext_csd.hpi_en = 1;
+		}
 	}
 
 	/*
-	 * If cache size is higher than 0, this indicates
-	 * the existence of cache and it can be turned on.
+	 * If cache size is higher than 0, this indicates the existence of cache
+	 * and it can be turned on. Note that some eMMCs from Micron has been
+	 * reported to need ~800 ms timeout, while enabling the cache after
+	 * sudden power failure tests. Let's extend the timeout to a minimum of
+	 * DEFAULT_CACHE_EN_TIMEOUT_MS and do it for all cards.
 	 */
-	if (!mmc_card_broken_hpi(card) &&
-	    card->ext_csd.cache_size > 0) {
+	if (card->ext_csd.cache_size > 0) {
+		unsigned int timeout_ms = MIN_CACHE_EN_TIMEOUT_MS;
+
+		timeout_ms = max(card->ext_csd.generic_cmd6_time, timeout_ms);
 		err = mmc_switch(card, EXT_CSD_CMD_SET_NORMAL,
-				EXT_CSD_CACHE_CTRL, 1,
-				card->ext_csd.generic_cmd6_time);
+				EXT_CSD_CACHE_CTRL, 1, timeout_ms);
 		if (err && err != -EBADMSG)
 			goto free_card;
 
diff --git a/drivers/mmc/host/omap_hsmmc.c b/drivers/mmc/host/omap_hsmmc.c
index 6094e9e687b1..29a1ddaa7466 100644
--- a/drivers/mmc/host/omap_hsmmc.c
+++ b/drivers/mmc/host/omap_hsmmc.c
@@ -1899,7 +1899,6 @@ static int omap_hsmmc_probe(struct platform_device *pdev)
 	mmc->max_blk_size = 512;       /* Block Length at max can be 1024 */
 	mmc->max_blk_count = 0xFFFF;    /* No. of Blocks is 16 bits */
 	mmc->max_req_size = mmc->max_blk_size * mmc->max_blk_count;
-	mmc->max_seg_size = mmc->max_req_size;
 
 	mmc->caps |= MMC_CAP_MMC_HIGHSPEED | MMC_CAP_SD_HIGHSPEED |
 		     MMC_CAP_WAIT_WHILE_BUSY | MMC_CAP_ERASE | MMC_CAP_CMD23;
@@ -1929,6 +1928,17 @@ static int omap_hsmmc_probe(struct platform_device *pdev)
 		goto err_irq;
 	}
 
+	/*
+	 * Limit the maximum segment size to the lower of the request size
+	 * and the DMA engine device segment size limits.  In reality, with
+	 * 32-bit transfers, the DMA engine can do longer segments than this
+	 * but there is no way to represent that in the DMA model - if we
+	 * increase this figure here, we get warnings from the DMA API debug.
+	 */
+	mmc->max_seg_size = min3(mmc->max_req_size,
+			dma_get_max_seg_size(host->rx_chan->device->dev),
+			dma_get_max_seg_size(host->tx_chan->device->dev));
+
 	/* Request IRQ for MMC operations */
 	ret = devm_request_irq(&pdev->dev, host->irq, omap_hsmmc_irq, 0,
 			mmc_hostname(mmc), host);
diff --git a/drivers/mmc/host/sdhci-tegra.c b/drivers/mmc/host/sdhci-tegra.c
index 7b95d088fdef..e6ace31e2a41 100644
--- a/drivers/mmc/host/sdhci-tegra.c
+++ b/drivers/mmc/host/sdhci-tegra.c
@@ -510,25 +510,25 @@ static void tegra_sdhci_parse_pad_autocal_dt(struct sdhci_host *host)
 
 	err = device_property_read_u32(host->mmc->parent,
 			"nvidia,pad-autocal-pull-up-offset-3v3-timeout",
-			&autocal->pull_up_3v3);
+			&autocal->pull_up_3v3_timeout);
 	if (err)
 		autocal->pull_up_3v3_timeout = 0;
 
 	err = device_property_read_u32(host->mmc->parent,
 			"nvidia,pad-autocal-pull-down-offset-3v3-timeout",
-			&autocal->pull_down_3v3);
+			&autocal->pull_down_3v3_timeout);
 	if (err)
 		autocal->pull_down_3v3_timeout = 0;
 
 	err = device_property_read_u32(host->mmc->parent,
 			"nvidia,pad-autocal-pull-up-offset-1v8-timeout",
-			&autocal->pull_up_1v8);
+			&autocal->pull_up_1v8_timeout);
 	if (err)
 		autocal->pull_up_1v8_timeout = 0;
 
 	err = device_property_read_u32(host->mmc->parent,
 			"nvidia,pad-autocal-pull-down-offset-1v8-timeout",
-			&autocal->pull_down_1v8);
+			&autocal->pull_down_1v8_timeout);
 	if (err)
 		autocal->pull_down_1v8_timeout = 0;
 
diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c
index 5de845b6b4d0..a22e11a65658 100644
--- a/drivers/mmc/host/sdhci.c
+++ b/drivers/mmc/host/sdhci.c
@@ -127,12 +127,12 @@ static void sdhci_do_enable_v4_mode(struct sdhci_host *host)
 {
 	u16 ctrl2;
 
-	ctrl2 = sdhci_readb(host, SDHCI_HOST_CONTROL2);
+	ctrl2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
 	if (ctrl2 & SDHCI_CTRL_V4_MODE)
 		return;
 
 	ctrl2 |= SDHCI_CTRL_V4_MODE;
-	sdhci_writeb(host, ctrl2, SDHCI_HOST_CONTROL);
+	sdhci_writew(host, ctrl2, SDHCI_HOST_CONTROL2);
 }
 
 /*