Merge branch 'next' into for-linus

13 files changed, 630 insertions, 190 deletions

diff --git a/drivers/mmc/core/Kconfig b/drivers/mmc/core/Kconfig
index bb22ffd76ef8..ef103871517f 100644
--- a/drivers/mmc/core/Kconfig
+++ b/drivers/mmc/core/Kconfig
@@ -16,3 +16,14 @@ config MMC_UNSAFE_RESUME
 
 	  This option sets a default which can be overridden by the
 	  module parameter "removable=0" or "removable=1".
+
+config MMC_CLKGATE
+	bool "MMC host clock gating (EXPERIMENTAL)"
+	depends on EXPERIMENTAL
+	help
+	  This will attempt to aggressively gate the clock to the MMC card.
+	  This is done to save power due to gating off the logic and bus
+	  noise when the MMC card is not in use. Your host driver has to
+	  support handling this in order for it to be of any use.
+
+	  If unsure, say N.
diff --git a/drivers/mmc/core/bus.c b/drivers/mmc/core/bus.c
index af8dc6a2a317..63667a8f140c 100644
--- a/drivers/mmc/core/bus.c
+++ b/drivers/mmc/core/bus.c
@@ -303,14 +303,14 @@ int mmc_add_card(struct mmc_card *card)
 			type, card->rca);
 	}
 
-	ret = device_add(&card->dev);
-	if (ret)
-		return ret;
-
 #ifdef CONFIG_DEBUG_FS
 	mmc_add_card_debugfs(card);
 #endif
 
+	ret = device_add(&card->dev);
+	if (ret)
+		return ret;
+
 	mmc_card_set_present(card);
 
 	return 0;
diff --git a/drivers/mmc/core/core.c b/drivers/mmc/core/core.c
index 8f86d702e46e..6625c057be05 100644
--- a/drivers/mmc/core/core.c
+++ b/drivers/mmc/core/core.c
@@ -22,6 +22,7 @@
 #include <linux/scatterlist.h>
 #include <linux/log2.h>
 #include <linux/regulator/consumer.h>
+#include <linux/pm_runtime.h>
 
 #include <linux/mmc/card.h>
 #include <linux/mmc/host.h>
@@ -130,6 +131,8 @@ void mmc_request_done(struct mmc_host *host, struct mmc_request *mrq)
 
 		if (mrq->done)
 			mrq->done(mrq);
+
+		mmc_host_clk_gate(host);
 	}
 }
 
@@ -190,6 +193,7 @@ mmc_start_request(struct mmc_host *host, struct mmc_request *mrq)
 			mrq->stop->mrq = mrq;
 		}
 	}
+	mmc_host_clk_ungate(host);
 	host->ops->request(host, mrq);
 }
 
@@ -295,8 +299,9 @@ void mmc_set_data_timeout(struct mmc_data *data, const struct mmc_card *card)
 		unsigned int timeout_us, limit_us;
 
 		timeout_us = data->timeout_ns / 1000;
-		timeout_us += data->timeout_clks * 1000 /
-			(card->host->ios.clock / 1000);
+		if (mmc_host_clk_rate(card->host))
+			timeout_us += data->timeout_clks * 1000 /
+				(mmc_host_clk_rate(card->host) / 1000);
 
 		if (data->flags & MMC_DATA_WRITE)
 			/*
@@ -614,6 +619,8 @@ static inline void mmc_set_ios(struct mmc_host *host)
 		 ios->power_mode, ios->chip_select, ios->vdd,
 		 ios->bus_width, ios->timing);
 
+	if (ios->clock > 0)
+		mmc_set_ungated(host);
 	host->ops->set_ios(host, ios);
 }
 
@@ -641,6 +648,61 @@ void mmc_set_clock(struct mmc_host *host, unsigned int hz)
 	mmc_set_ios(host);
 }
 
+#ifdef CONFIG_MMC_CLKGATE
+/*
+ * This gates the clock by setting it to 0 Hz.
+ */
+void mmc_gate_clock(struct mmc_host *host)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->clk_lock, flags);
+	host->clk_old = host->ios.clock;
+	host->ios.clock = 0;
+	host->clk_gated = true;
+	spin_unlock_irqrestore(&host->clk_lock, flags);
+	mmc_set_ios(host);
+}
+
+/*
+ * This restores the clock from gating by using the cached
+ * clock value.
+ */
+void mmc_ungate_clock(struct mmc_host *host)
+{
+	/*
+	 * We should previously have gated the clock, so the clock shall
+	 * be 0 here! The clock may however be 0 during initialization,
+	 * when some request operations are performed before setting
+	 * the frequency. When ungate is requested in that situation
+	 * we just ignore the call.
+	 */
+	if (host->clk_old) {
+		BUG_ON(host->ios.clock);
+		/* This call will also set host->clk_gated to false */
+		mmc_set_clock(host, host->clk_old);
+	}
+}
+
+void mmc_set_ungated(struct mmc_host *host)
+{
+	unsigned long flags;
+
+	/*
+	 * We've been given a new frequency while the clock is gated,
+	 * so make sure we regard this as ungating it.
+	 */
+	spin_lock_irqsave(&host->clk_lock, flags);
+	host->clk_gated = false;
+	spin_unlock_irqrestore(&host->clk_lock, flags);
+}
+
+#else
+void mmc_set_ungated(struct mmc_host *host)
+{
+}
+#endif
+
 /*
  * Change the bus mode (open drain/push-pull) of a host.
  */
@@ -1424,35 +1486,57 @@ int mmc_set_blocklen(struct mmc_card *card, unsigned int blocklen)
 }
 EXPORT_SYMBOL(mmc_set_blocklen);
 
+static int mmc_rescan_try_freq(struct mmc_host *host, unsigned freq)
+{
+	host->f_init = freq;
+
+#ifdef CONFIG_MMC_DEBUG
+	pr_info("%s: %s: trying to init card at %u Hz\n",
+		mmc_hostname(host), __func__, host->f_init);
+#endif
+	mmc_power_up(host);
+	sdio_reset(host);
+	mmc_go_idle(host);
+
+	mmc_send_if_cond(host, host->ocr_avail);
+
+	/* Order's important: probe SDIO, then SD, then MMC */
+	if (!mmc_attach_sdio(host))
+		return 0;
+	if (!mmc_attach_sd(host))
+		return 0;
+	if (!mmc_attach_mmc(host))
+		return 0;
+
+	mmc_power_off(host);
+	return -EIO;
+}
+
 void mmc_rescan(struct work_struct *work)
 {
+	static const unsigned freqs[] = { 400000, 300000, 200000, 100000 };
 	struct mmc_host *host =
 		container_of(work, struct mmc_host, detect.work);
-	u32 ocr;
-	int err;
-	unsigned long flags;
 	int i;
-	const unsigned freqs[] = { 400000, 300000, 200000, 100000 };
-
-	spin_lock_irqsave(&host->lock, flags);
 
-	if (host->rescan_disable) {
-		spin_unlock_irqrestore(&host->lock, flags);
+	if (host->rescan_disable)
 		return;
-	}
-
-	spin_unlock_irqrestore(&host->lock, flags);
-
 
 	mmc_bus_get(host);
 
-	/* if there is a card registered, check whether it is still present */
-	if ((host->bus_ops != NULL) && host->bus_ops->detect && !host->bus_dead)
+	/*
+	 * if there is a _removable_ card registered, check whether it is
+	 * still present
+	 */
+	if (host->bus_ops && host->bus_ops->detect && !host->bus_dead
+	    && mmc_card_is_removable(host))
 		host->bus_ops->detect(host);
 
+	/*
+	 * Let mmc_bus_put() free the bus/bus_ops if we've found that
+	 * the card is no longer present.
+	 */
 	mmc_bus_put(host);
-
-
 	mmc_bus_get(host);
 
 	/* if there still is a card present, stop here */
@@ -1461,8 +1545,6 @@ void mmc_rescan(struct work_struct *work)
 		goto out;
 	}
 
-	/* detect a newly inserted card */
-
 	/*
 	 * Only we can add a new handler, so it's safe to
 	 * release the lock here.
@@ -1472,72 +1554,16 @@ void mmc_rescan(struct work_struct *work)
 	if (host->ops->get_cd && host->ops->get_cd(host) == 0)
 		goto out;
 
+	mmc_claim_host(host);
 	for (i = 0; i < ARRAY_SIZE(freqs); i++) {
-		mmc_claim_host(host);
-
-		if (freqs[i] >= host->f_min)
-			host->f_init = freqs[i];
-		else if (!i || freqs[i-1] > host->f_min)
-			host->f_init = host->f_min;
-		else {
-			mmc_release_host(host);
-			goto out;
-		}
-#ifdef CONFIG_MMC_DEBUG
-		pr_info("%s: %s: trying to init card at %u Hz\n",
-			mmc_hostname(host), __func__, host->f_init);
-#endif
-		mmc_power_up(host);
-		sdio_reset(host);
-		mmc_go_idle(host);
-
-		mmc_send_if_cond(host, host->ocr_avail);
-
-		/*
-		 * First we search for SDIO...
-		 */
-		err = mmc_send_io_op_cond(host, 0, &ocr);
-		if (!err) {
-			if (mmc_attach_sdio(host, ocr)) {
-				mmc_claim_host(host);
-				/*
-				 * Try SDMEM (but not MMC) even if SDIO
-				 * is broken.
-				 */
-				if (mmc_send_app_op_cond(host, 0, &ocr))
-					goto out_fail;
-
-				if (mmc_attach_sd(host, ocr))
-					mmc_power_off(host);
-			}
-			goto out;
-		}
-
-		/*
-		 * ...then normal SD...
-		 */
-		err = mmc_send_app_op_cond(host, 0, &ocr);
-		if (!err) {
-			if (mmc_attach_sd(host, ocr))
-				mmc_power_off(host);
-			goto out;
-		}
-
-		/*
-		 * ...and finally MMC.
-		 */
-		err = mmc_send_op_cond(host, 0, &ocr);
-		if (!err) {
-			if (mmc_attach_mmc(host, ocr))
-				mmc_power_off(host);
-			goto out;
-		}
-
-out_fail:
-		mmc_release_host(host);
-		mmc_power_off(host);
+		if (!mmc_rescan_try_freq(host, max(freqs[i], host->f_min)))
+			break;
+		if (freqs[i] < host->f_min)
+			break;
 	}
-out:
+	mmc_release_host(host);
+
+ out:
 	if (host->caps & MMC_CAP_NEEDS_POLL)
 		mmc_schedule_delayed_work(&host->detect, HZ);
 }
@@ -1559,7 +1585,7 @@ void mmc_stop_host(struct mmc_host *host)
 
 	if (host->caps & MMC_CAP_DISABLE)
 		cancel_delayed_work(&host->disable);
-	cancel_delayed_work(&host->detect);
+	cancel_delayed_work_sync(&host->detect);
 	mmc_flush_scheduled_work();
 
 	/* clear pm flags now and let card drivers set them as needed */
@@ -1721,6 +1747,18 @@ int mmc_resume_host(struct mmc_host *host)
 		if (!(host->pm_flags & MMC_PM_KEEP_POWER)) {
 			mmc_power_up(host);
 			mmc_select_voltage(host, host->ocr);
+			/*
+			 * Tell runtime PM core we just powered up the card,
+			 * since it still believes the card is powered off.
+			 * Note that currently runtime PM is only enabled
+			 * for SDIO cards that are MMC_CAP_POWER_OFF_CARD
+			 */
+			if (mmc_card_sdio(host->card) &&
+			    (host->caps & MMC_CAP_POWER_OFF_CARD)) {
+				pm_runtime_disable(&host->card->dev);
+				pm_runtime_set_active(&host->card->dev);
+				pm_runtime_enable(&host->card->dev);
+			}
 		}
 		BUG_ON(!host->bus_ops->resume);
 		err = host->bus_ops->resume(host);
@@ -1773,6 +1811,7 @@ int mmc_pm_notify(struct notifier_block *notify_block,
 
 	case PM_POST_SUSPEND:
 	case PM_POST_HIBERNATION:
+	case PM_POST_RESTORE:
 
 		spin_lock_irqsave(&host->lock, flags);
 		host->rescan_disable = 0;
@@ -1789,7 +1828,7 @@ static int __init mmc_init(void)
 {
 	int ret;
 
-	workqueue = create_singlethread_workqueue("kmmcd");
+	workqueue = alloc_ordered_workqueue("kmmcd", 0);
 	if (!workqueue)
 		return -ENOMEM;
 
diff --git a/drivers/mmc/core/core.h b/drivers/mmc/core/core.h
index 77240cd11bcf..ca1fdde29df6 100644
--- a/drivers/mmc/core/core.h
+++ b/drivers/mmc/core/core.h
@@ -33,6 +33,9 @@ void mmc_init_erase(struct mmc_card *card);
 
 void mmc_set_chip_select(struct mmc_host *host, int mode);
 void mmc_set_clock(struct mmc_host *host, unsigned int hz);
+void mmc_gate_clock(struct mmc_host *host);
+void mmc_ungate_clock(struct mmc_host *host);
+void mmc_set_ungated(struct mmc_host *host);
 void mmc_set_bus_mode(struct mmc_host *host, unsigned int mode);
 void mmc_set_bus_width(struct mmc_host *host, unsigned int width);
 void mmc_set_bus_width_ddr(struct mmc_host *host, unsigned int width,
@@ -54,9 +57,9 @@ void mmc_rescan(struct work_struct *work);
 void mmc_start_host(struct mmc_host *host);
 void mmc_stop_host(struct mmc_host *host);
 
-int mmc_attach_mmc(struct mmc_host *host, u32 ocr);
-int mmc_attach_sd(struct mmc_host *host, u32 ocr);
-int mmc_attach_sdio(struct mmc_host *host, u32 ocr);
+int mmc_attach_mmc(struct mmc_host *host);
+int mmc_attach_sd(struct mmc_host *host);
+int mmc_attach_sdio(struct mmc_host *host);
 
 /* Module parameters */
 extern int use_spi_crc;
diff --git a/drivers/mmc/core/debugfs.c b/drivers/mmc/core/debugfs.c
index eed1405fd742..998797ed67a6 100644
--- a/drivers/mmc/core/debugfs.c
+++ b/drivers/mmc/core/debugfs.c
@@ -183,6 +183,11 @@ void mmc_add_host_debugfs(struct mmc_host *host)
 			&mmc_clock_fops))
 		goto err_node;
 
+#ifdef CONFIG_MMC_CLKGATE
+	if (!debugfs_create_u32("clk_delay", (S_IRUSR | S_IWUSR),
+				root, &host->clk_delay))
+		goto err_node;
+#endif
 	return;
 
 err_node:
diff --git a/drivers/mmc/core/host.c b/drivers/mmc/core/host.c
index 10b8af27e03a..b3ac6c5bc5c6 100644
--- a/drivers/mmc/core/host.c
+++ b/drivers/mmc/core/host.c
@@ -3,6 +3,7 @@
  *
  *  Copyright (C) 2003 Russell King, All Rights Reserved.
  *  Copyright (C) 2007-2008 Pierre Ossman
+ *  Copyright (C) 2010 Linus Walleij
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License version 2 as
@@ -20,6 +21,7 @@
 #include <linux/suspend.h>
 
 #include <linux/mmc/host.h>
+#include <linux/mmc/card.h>
 
 #include "core.h"
 #include "host.h"
@@ -50,6 +52,205 @@ void mmc_unregister_host_class(void)
 static DEFINE_IDR(mmc_host_idr);
 static DEFINE_SPINLOCK(mmc_host_lock);
 
+#ifdef CONFIG_MMC_CLKGATE
+
+/*
+ * Enabling clock gating will make the core call out to the host
+ * once up and once down when it performs a request or card operation
+ * intermingled in any fashion. The driver will see this through
+ * set_ios() operations with ios.clock field set to 0 to gate (disable)
+ * the block clock, and to the old frequency to enable it again.
+ */
+static void mmc_host_clk_gate_delayed(struct mmc_host *host)
+{
+	unsigned long tick_ns;
+	unsigned long freq = host->ios.clock;
+	unsigned long flags;
+
+	if (!freq) {
+		pr_debug("%s: frequency set to 0 in disable function, "
+			 "this means the clock is already disabled.\n",
+			 mmc_hostname(host));
+		return;
+	}
+	/*
+	 * New requests may have appeared while we were scheduling,
+	 * then there is no reason to delay the check before
+	 * clk_disable().
+	 */
+	spin_lock_irqsave(&host->clk_lock, flags);
+
+	/*
+	 * Delay n bus cycles (at least 8 from MMC spec) before attempting
+	 * to disable the MCI block clock. The reference count may have
+	 * gone up again after this delay due to rescheduling!
+	 */
+	if (!host->clk_requests) {
+		spin_unlock_irqrestore(&host->clk_lock, flags);
+		tick_ns = DIV_ROUND_UP(1000000000, freq);
+		ndelay(host->clk_delay * tick_ns);
+	} else {
+		/* New users appeared while waiting for this work */
+		spin_unlock_irqrestore(&host->clk_lock, flags);
+		return;
+	}
+	mutex_lock(&host->clk_gate_mutex);
+	spin_lock_irqsave(&host->clk_lock, flags);
+	if (!host->clk_requests) {
+		spin_unlock_irqrestore(&host->clk_lock, flags);
+		/* This will set host->ios.clock to 0 */
+		mmc_gate_clock(host);
+		spin_lock_irqsave(&host->clk_lock, flags);
+		pr_debug("%s: gated MCI clock\n", mmc_hostname(host));
+	}
+	spin_unlock_irqrestore(&host->clk_lock, flags);
+	mutex_unlock(&host->clk_gate_mutex);
+}
+
+/*
+ * Internal work. Work to disable the clock at some later point.
+ */
+static void mmc_host_clk_gate_work(struct work_struct *work)
+{
+	struct mmc_host *host = container_of(work, struct mmc_host,
+					      clk_gate_work);
+
+	mmc_host_clk_gate_delayed(host);
+}
+
+/**
+ *	mmc_host_clk_ungate - ungate hardware MCI clocks
+ *	@host: host to ungate.
+ *
+ *	Makes sure the host ios.clock is restored to a non-zero value
+ *	past this call.	Increase clock reference count and ungate clock
+ *	if we're the first user.
+ */
+void mmc_host_clk_ungate(struct mmc_host *host)
+{
+	unsigned long flags;
+
+	mutex_lock(&host->clk_gate_mutex);
+	spin_lock_irqsave(&host->clk_lock, flags);
+	if (host->clk_gated) {
+		spin_unlock_irqrestore(&host->clk_lock, flags);
+		mmc_ungate_clock(host);
+		spin_lock_irqsave(&host->clk_lock, flags);
+		pr_debug("%s: ungated MCI clock\n", mmc_hostname(host));
+	}
+	host->clk_requests++;
+	spin_unlock_irqrestore(&host->clk_lock, flags);
+	mutex_unlock(&host->clk_gate_mutex);
+}
+
+/**
+ *	mmc_host_may_gate_card - check if this card may be gated
+ *	@card: card to check.
+ */
+static bool mmc_host_may_gate_card(struct mmc_card *card)
+{
+	/* If there is no card we may gate it */
+	if (!card)
+		return true;
+	/*
+	 * Don't gate SDIO cards! These need to be clocked at all times
+	 * since they may be independent systems generating interrupts
+	 * and other events. The clock requests counter from the core will
+	 * go down to zero since the core does not need it, but we will not
+	 * gate the clock, because there is somebody out there that may still
+	 * be using it.
+	 */
+	if (mmc_card_sdio(card))
+		return false;
+
+	return true;
+}
+
+/**
+ *	mmc_host_clk_gate - gate off hardware MCI clocks
+ *	@host: host to gate.
+ *
+ *	Calls the host driver with ios.clock set to zero as often as possible
+ *	in order to gate off hardware MCI clocks. Decrease clock reference
+ *	count and schedule disabling of clock.
+ */
+void mmc_host_clk_gate(struct mmc_host *host)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->clk_lock, flags);
+	host->clk_requests--;
+	if (mmc_host_may_gate_card(host->card) &&
+	    !host->clk_requests)
+		schedule_work(&host->clk_gate_work);
+	spin_unlock_irqrestore(&host->clk_lock, flags);
+}
+
+/**
+ *	mmc_host_clk_rate - get current clock frequency setting
+ *	@host: host to get the clock frequency for.
+ *
+ *	Returns current clock frequency regardless of gating.
+ */
+unsigned int mmc_host_clk_rate(struct mmc_host *host)
+{
+	unsigned long freq;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->clk_lock, flags);
+	if (host->clk_gated)
+		freq = host->clk_old;
+	else
+		freq = host->ios.clock;
+	spin_unlock_irqrestore(&host->clk_lock, flags);
+	return freq;
+}
+
+/**
+ *	mmc_host_clk_init - set up clock gating code
+ *	@host: host with potential clock to control
+ */
+static inline void mmc_host_clk_init(struct mmc_host *host)
+{
+	host->clk_requests = 0;
+	/* Hold MCI clock for 8 cycles by default */
+	host->clk_delay = 8;
+	host->clk_gated = false;
+	INIT_WORK(&host->clk_gate_work, mmc_host_clk_gate_work);
+	spin_lock_init(&host->clk_lock);
+	mutex_init(&host->clk_gate_mutex);
+}
+
+/**
+ *	mmc_host_clk_exit - shut down clock gating code
+ *	@host: host with potential clock to control
+ */
+static inline void mmc_host_clk_exit(struct mmc_host *host)
+{
+	/*
+	 * Wait for any outstanding gate and then make sure we're
+	 * ungated before exiting.
+	 */
+	if (cancel_work_sync(&host->clk_gate_work))
+		mmc_host_clk_gate_delayed(host);
+	if (host->clk_gated)
+		mmc_host_clk_ungate(host);
+	/* There should be only one user now */
+	WARN_ON(host->clk_requests > 1);
+}
+
+#else
+
+static inline void mmc_host_clk_init(struct mmc_host *host)
+{
+}
+
+static inline void mmc_host_clk_exit(struct mmc_host *host)
+{
+}
+
+#endif
+
 /**
  *	mmc_alloc_host - initialise the per-host structure.
  *	@extra: sizeof private data structure
@@ -82,6 +283,8 @@ struct mmc_host *mmc_alloc_host(int extra, struct device *dev)
 	host->class_dev.class = &mmc_host_class;
 	device_initialize(&host->class_dev);
 
+	mmc_host_clk_init(host);
+
 	spin_lock_init(&host->lock);
 	init_waitqueue_head(&host->wq);
 	INIT_DELAYED_WORK(&host->detect, mmc_rescan);
@@ -163,6 +366,8 @@ void mmc_remove_host(struct mmc_host *host)
 	device_del(&host->class_dev);
 
 	led_trigger_unregister_simple(host->led);
+
+	mmc_host_clk_exit(host);
 }
 
 EXPORT_SYMBOL(mmc_remove_host);
@@ -183,4 +388,3 @@ void mmc_free_host(struct mmc_host *host)
 }
 
 EXPORT_SYMBOL(mmc_free_host);
-
diff --git a/drivers/mmc/core/host.h b/drivers/mmc/core/host.h
index 8c87e1109a34..de199f911928 100644
--- a/drivers/mmc/core/host.h
+++ b/drivers/mmc/core/host.h
@@ -10,10 +10,31 @@
  */
 #ifndef _MMC_CORE_HOST_H
 #define _MMC_CORE_HOST_H
+#include <linux/mmc/host.h>
 
 int mmc_register_host_class(void);
 void mmc_unregister_host_class(void);
 
+#ifdef CONFIG_MMC_CLKGATE
+void mmc_host_clk_ungate(struct mmc_host *host);
+void mmc_host_clk_gate(struct mmc_host *host);
+unsigned int mmc_host_clk_rate(struct mmc_host *host);
+
+#else
+static inline void mmc_host_clk_ungate(struct mmc_host *host)
+{
+}
+
+static inline void mmc_host_clk_gate(struct mmc_host *host)
+{
+}
+
+static inline unsigned int mmc_host_clk_rate(struct mmc_host *host)
+{
+	return host->ios.clock;
+}
+#endif
+
 void mmc_host_deeper_disable(struct work_struct *work);
 
 #endif
diff --git a/drivers/mmc/core/mmc.c b/drivers/mmc/core/mmc.c
index 995261f7fd70..16006ef153fe 100644
--- a/drivers/mmc/core/mmc.c
+++ b/drivers/mmc/core/mmc.c
@@ -375,7 +375,7 @@ static int mmc_init_card(struct mmc_host *host, u32 ocr,
 	struct mmc_card *oldcard)
 {
 	struct mmc_card *card;
-	int err, ddr = MMC_SDR_MODE;
+	int err, ddr = 0;
 	u32 cid[4];
 	unsigned int max_dtr;
 
@@ -534,34 +534,56 @@ static int mmc_init_card(struct mmc_host *host, u32 ocr,
 	 */
 	if ((card->csd.mmca_vsn >= CSD_SPEC_VER_4) &&
 	    (host->caps & (MMC_CAP_4_BIT_DATA | MMC_CAP_8_BIT_DATA))) {
-		unsigned ext_csd_bit, bus_width;
-
-		if (host->caps & MMC_CAP_8_BIT_DATA) {
-			if (ddr)
-				ext_csd_bit = EXT_CSD_DDR_BUS_WIDTH_8;
-			else
-				ext_csd_bit = EXT_CSD_BUS_WIDTH_8;
-			bus_width = MMC_BUS_WIDTH_8;
-		} else {
-			if (ddr)
-				ext_csd_bit = EXT_CSD_DDR_BUS_WIDTH_4;
-			else
-				ext_csd_bit = EXT_CSD_BUS_WIDTH_4;
-			bus_width = MMC_BUS_WIDTH_4;
+		static unsigned ext_csd_bits[][2] = {
+			{ EXT_CSD_BUS_WIDTH_8, EXT_CSD_DDR_BUS_WIDTH_8 },
+			{ EXT_CSD_BUS_WIDTH_4, EXT_CSD_DDR_BUS_WIDTH_4 },
+			{ EXT_CSD_BUS_WIDTH_1, EXT_CSD_BUS_WIDTH_1 },
+		};
+		static unsigned bus_widths[] = {
+			MMC_BUS_WIDTH_8,
+			MMC_BUS_WIDTH_4,
+			MMC_BUS_WIDTH_1
+		};
+		unsigned idx, bus_width = 0;
+
+		if (host->caps & MMC_CAP_8_BIT_DATA)
+			idx = 0;
+		else
+			idx = 1;
+		for (; idx < ARRAY_SIZE(bus_widths); idx++) {
+			bus_width = bus_widths[idx];
+			if (bus_width == MMC_BUS_WIDTH_1)
+				ddr = 0; /* no DDR for 1-bit width */
+			err = mmc_switch(card, EXT_CSD_CMD_SET_NORMAL,
+					 EXT_CSD_BUS_WIDTH,
+					 ext_csd_bits[idx][0]);
+			if (!err) {
+				mmc_set_bus_width_ddr(card->host,
+						      bus_width, MMC_SDR_MODE);
+				/*
+				 * If controller can't handle bus width test,
+				 * use the highest bus width to maintain
+				 * compatibility with previous MMC behavior.
+				 */
+				if (!(host->caps & MMC_CAP_BUS_WIDTH_TEST))
+					break;
+				err = mmc_bus_test(card, bus_width);
+				if (!err)
+					break;
+			}
 		}
 
-		err = mmc_switch(card, EXT_CSD_CMD_SET_NORMAL,
-				 EXT_CSD_BUS_WIDTH, ext_csd_bit);
-
-		if (err && err != -EBADMSG)
-			goto free_card;
-
+		if (!err && ddr) {
+			err = mmc_switch(card, EXT_CSD_CMD_SET_NORMAL,
+					EXT_CSD_BUS_WIDTH,
+					ext_csd_bits[idx][1]);
+		}
 		if (err) {
 			printk(KERN_WARNING "%s: switch to bus width %d ddr %d "
-			       "failed\n", mmc_hostname(card->host),
-			       1 << bus_width, ddr);
-			err = 0;
-		} else {
+				"failed\n", mmc_hostname(card->host),
+				1 << bus_width, ddr);
+			goto free_card;
+		} else if (ddr) {
 			mmc_card_set_ddr_mode(card);
 			mmc_set_bus_width_ddr(card->host, bus_width, ddr);
 		}
@@ -733,14 +755,21 @@ static void mmc_attach_bus_ops(struct mmc_host *host)
 /*
  * Starting point for MMC card init.
  */
-int mmc_attach_mmc(struct mmc_host *host, u32 ocr)
+int mmc_attach_mmc(struct mmc_host *host)
 {
 	int err;
+	u32 ocr;
 
 	BUG_ON(!host);
 	WARN_ON(!host->claimed);
 
+	err = mmc_send_op_cond(host, 0, &ocr);
+	if (err)
+		return err;
+
 	mmc_attach_bus_ops(host);
+	if (host->ocr_avail_mmc)
+		host->ocr_avail = host->ocr_avail_mmc;
 
 	/*
 	 * We need to get OCR a different way for SPI.
@@ -780,20 +809,20 @@ int mmc_attach_mmc(struct mmc_host *host, u32 ocr)
 		goto err;
 
 	mmc_release_host(host);
-
 	err = mmc_add_card(host->card);
+	mmc_claim_host(host);
 	if (err)
 		goto remove_card;
 
 	return 0;
 
 remove_card:
+	mmc_release_host(host);
 	mmc_remove_card(host->card);
-	host->card = NULL;
 	mmc_claim_host(host);
+	host->card = NULL;
 err:
 	mmc_detach_bus(host);
-	mmc_release_host(host);
 
 	printk(KERN_ERR "%s: error %d whilst initialising MMC card\n",
 		mmc_hostname(host), err);
diff --git a/drivers/mmc/core/mmc_ops.c b/drivers/mmc/core/mmc_ops.c
index 326447c9ede8..60842f878ded 100644
--- a/drivers/mmc/core/mmc_ops.c
+++ b/drivers/mmc/core/mmc_ops.c
@@ -462,3 +462,104 @@ int mmc_send_status(struct mmc_card *card, u32 *status)
 	return 0;
 }
 
+static int
+mmc_send_bus_test(struct mmc_card *card, struct mmc_host *host, u8 opcode,
+		  u8 len)
+{
+	struct mmc_request mrq;
+	struct mmc_command cmd;
+	struct mmc_data data;
+	struct scatterlist sg;
+	u8 *data_buf;
+	u8 *test_buf;
+	int i, err;
+	static u8 testdata_8bit[8] = { 0x55, 0xaa, 0, 0, 0, 0, 0, 0 };
+	static u8 testdata_4bit[4] = { 0x5a, 0, 0, 0 };
+
+	/* dma onto stack is unsafe/nonportable, but callers to this
+	 * routine normally provide temporary on-stack buffers ...
+	 */
+	data_buf = kmalloc(len, GFP_KERNEL);
+	if (!data_buf)
+		return -ENOMEM;
+
+	if (len == 8)
+		test_buf = testdata_8bit;
+	else if (len == 4)
+		test_buf = testdata_4bit;
+	else {
+		printk(KERN_ERR "%s: Invalid bus_width %d\n",
+		       mmc_hostname(host), len);
+		kfree(data_buf);
+		return -EINVAL;
+	}
+
+	if (opcode == MMC_BUS_TEST_W)
+		memcpy(data_buf, test_buf, len);
+
+	memset(&mrq, 0, sizeof(struct mmc_request));
+	memset(&cmd, 0, sizeof(struct mmc_command));
+	memset(&data, 0, sizeof(struct mmc_data));
+
+	mrq.cmd = &cmd;
+	mrq.data = &data;
+	cmd.opcode = opcode;
+	cmd.arg = 0;
+
+	/* NOTE HACK:  the MMC_RSP_SPI_R1 is always correct here, but we
+	 * rely on callers to never use this with "native" calls for reading
+	 * CSD or CID.  Native versions of those commands use the R2 type,
+	 * not R1 plus a data block.
+	 */
+	cmd.flags = MMC_RSP_SPI_R1 | MMC_RSP_R1 | MMC_CMD_ADTC;
+
+	data.blksz = len;
+	data.blocks = 1;
+	if (opcode == MMC_BUS_TEST_R)
+		data.flags = MMC_DATA_READ;
+	else
+		data.flags = MMC_DATA_WRITE;
+
+	data.sg = &sg;
+	data.sg_len = 1;
+	sg_init_one(&sg, data_buf, len);
+	mmc_wait_for_req(host, &mrq);
+	err = 0;
+	if (opcode == MMC_BUS_TEST_R) {
+		for (i = 0; i < len / 4; i++)
+			if ((test_buf[i] ^ data_buf[i]) != 0xff) {
+				err = -EIO;
+				break;
+			}
+	}
+	kfree(data_buf);
+
+	if (cmd.error)
+		return cmd.error;
+	if (data.error)
+		return data.error;
+
+	return err;
+}
+
+int mmc_bus_test(struct mmc_card *card, u8 bus_width)
+{
+	int err, width;
+
+	if (bus_width == MMC_BUS_WIDTH_8)
+		width = 8;
+	else if (bus_width == MMC_BUS_WIDTH_4)
+		width = 4;
+	else if (bus_width == MMC_BUS_WIDTH_1)
+		return 0; /* no need for test */
+	else
+		return -EINVAL;
+
+	/*
+	 * Ignore errors from BUS_TEST_W.  BUS_TEST_R will fail if there
+	 * is a problem.  This improves chances that the test will work.
+	 */
+	mmc_send_bus_test(card, card->host, MMC_BUS_TEST_W, width);
+	err = mmc_send_bus_test(card, card->host, MMC_BUS_TEST_R, width);
+	return err;
+}
diff --git a/drivers/mmc/core/mmc_ops.h b/drivers/mmc/core/mmc_ops.h
index 653eb8e84178..e6d44b8a18db 100644
--- a/drivers/mmc/core/mmc_ops.h
+++ b/drivers/mmc/core/mmc_ops.h
@@ -26,6 +26,7 @@ int mmc_send_cid(struct mmc_host *host, u32 *cid);
 int mmc_spi_read_ocr(struct mmc_host *host, int highcap, u32 *ocrp);
 int mmc_spi_set_crc(struct mmc_host *host, int use_crc);
 int mmc_card_sleepawake(struct mmc_host *host, int sleep);
+int mmc_bus_test(struct mmc_card *card, u8 bus_width);
 
 #endif
 
diff --git a/drivers/mmc/core/sd.c b/drivers/mmc/core/sd.c
index 49da4dffd28e..d18c32bca99b 100644
--- a/drivers/mmc/core/sd.c
+++ b/drivers/mmc/core/sd.c
@@ -764,14 +764,21 @@ static void mmc_sd_attach_bus_ops(struct mmc_host *host)
 /*
  * Starting point for SD card init.
  */
-int mmc_attach_sd(struct mmc_host *host, u32 ocr)
+int mmc_attach_sd(struct mmc_host *host)
 {
 	int err;
+	u32 ocr;
 
 	BUG_ON(!host);
 	WARN_ON(!host->claimed);
 
+	err = mmc_send_app_op_cond(host, 0, &ocr);
+	if (err)
+		return err;
+
 	mmc_sd_attach_bus_ops(host);
+	if (host->ocr_avail_sd)
+		host->ocr_avail = host->ocr_avail_sd;
 
 	/*
 	 * We need to get OCR a different way for SPI.
@@ -795,7 +802,8 @@ int mmc_attach_sd(struct mmc_host *host, u32 ocr)
 		ocr &= ~0x7F;
 	}
 
-	if (ocr & MMC_VDD_165_195) {
+	if ((ocr & MMC_VDD_165_195) &&
+	    !(host->ocr_avail_sd & MMC_VDD_165_195)) {
 		printk(KERN_WARNING "%s: SD card claims to support the "
 		       "incompletely defined 'low voltage range'. This "
 		       "will be ignored.\n", mmc_hostname(host));
@@ -820,20 +828,20 @@ int mmc_attach_sd(struct mmc_host *host, u32 ocr)
 		goto err;
 
 	mmc_release_host(host);
-
 	err = mmc_add_card(host->card);
+	mmc_claim_host(host);
 	if (err)
 		goto remove_card;
 
 	return 0;
 
 remove_card:
+	mmc_release_host(host);
 	mmc_remove_card(host->card);
 	host->card = NULL;
 	mmc_claim_host(host);
 err:
 	mmc_detach_bus(host);
-	mmc_release_host(host);
 
 	printk(KERN_ERR "%s: error %d whilst initialising SD card\n",
 		mmc_hostname(host), err);
diff --git a/drivers/mmc/core/sdio.c b/drivers/mmc/core/sdio.c
index c3ad1058cd31..5c4a54d9b6a4 100644
--- a/drivers/mmc/core/sdio.c
+++ b/drivers/mmc/core/sdio.c
@@ -547,9 +547,11 @@ static void mmc_sdio_detect(struct mmc_host *host)
 	BUG_ON(!host->card);
 
 	/* Make sure card is powered before detecting it */
-	err = pm_runtime_get_sync(&host->card->dev);
-	if (err < 0)
-		goto out;
+	if (host->caps & MMC_CAP_POWER_OFF_CARD) {
+		err = pm_runtime_get_sync(&host->card->dev);
+		if (err < 0)
+			goto out;
+	}
 
 	mmc_claim_host(host);
 
@@ -560,6 +562,20 @@ static void mmc_sdio_detect(struct mmc_host *host)
 
 	mmc_release_host(host);
 
+	/*
+	 * Tell PM core it's OK to power off the card now.
+	 *
+	 * The _sync variant is used in order to ensure that the card
+	 * is left powered off in case an error occurred, and the card
+	 * is going to be removed.
+	 *
+	 * Since there is no specific reason to believe a new user
+	 * is about to show up at this point, the _sync variant is
+	 * desirable anyway.
+	 */
+	if (host->caps & MMC_CAP_POWER_OFF_CARD)
+		pm_runtime_put_sync(&host->card->dev);
+
 out:
 	if (err) {
 		mmc_sdio_remove(host);
@@ -568,9 +584,6 @@ out:
 		mmc_detach_bus(host);
 		mmc_release_host(host);
 	}
-
-	/* Tell PM core that we're done */
-	pm_runtime_put(&host->card->dev);
 }
 
 /*
@@ -614,15 +627,27 @@ static int mmc_sdio_suspend(struct mmc_host *host)
 
 static int mmc_sdio_resume(struct mmc_host *host)
 {
-	int i, err;
+	int i, err = 0;
 
 	BUG_ON(!host);
 	BUG_ON(!host->card);
 
 	/* Basic card reinitialization. */
 	mmc_claim_host(host);
-	err = mmc_sdio_init_card(host, host->ocr, host->card,
+
+	/* No need to reinitialize powered-resumed nonremovable cards */
+	if (mmc_card_is_removable(host) || !mmc_card_is_powered_resumed(host))
+		err = mmc_sdio_init_card(host, host->ocr, host->card,
 				 (host->pm_flags & MMC_PM_KEEP_POWER));
+	else if (mmc_card_is_powered_resumed(host)) {
+		/* We may have switched to 1-bit mode during suspend */
+		err = sdio_enable_4bit_bus(host->card);
+		if (err > 0) {
+			mmc_set_bus_width(host, MMC_BUS_WIDTH_4);
+			err = 0;
+		}
+	}
+
 	if (!err && host->sdio_irqs)
 		mmc_signal_sdio_irq(host);
 	mmc_release_host(host);
@@ -677,16 +702,22 @@ static const struct mmc_bus_ops mmc_sdio_ops = {
 /*
  * Starting point for SDIO card init.
  */
-int mmc_attach_sdio(struct mmc_host *host, u32 ocr)
+int mmc_attach_sdio(struct mmc_host *host)
 {
-	int err;
-	int i, funcs;
+	int err, i, funcs;
+	u32 ocr;
 	struct mmc_card *card;
 
 	BUG_ON(!host);
 	WARN_ON(!host->claimed);
 
+	err = mmc_send_io_op_cond(host, 0, &ocr);
+	if (err)
+		return err;
+
 	mmc_attach_bus(host, &mmc_sdio_ops);
+	if (host->ocr_avail_sdio)
+		host->ocr_avail = host->ocr_avail_sdio;
 
 	/*
 	 * Sanity check the voltages that the card claims to
@@ -718,16 +749,21 @@ int mmc_attach_sdio(struct mmc_host *host, u32 ocr)
 	card = host->card;
 
 	/*
-	 * Let runtime PM core know our card is active
+	 * Enable runtime PM only if supported by host+card+board
 	 */
-	err = pm_runtime_set_active(&card->dev);
-	if (err)
-		goto remove;
+	if (host->caps & MMC_CAP_POWER_OFF_CARD) {
+		/*
+		 * Let runtime PM core know our card is active
+		 */
+		err = pm_runtime_set_active(&card->dev);
+		if (err)
+			goto remove;
 
-	/*
-	 * Enable runtime PM for this card
-	 */
-	pm_runtime_enable(&card->dev);
+		/*
+		 * Enable runtime PM for this card
+		 */
+		pm_runtime_enable(&card->dev);
+	}
 
 	/*
 	 * The number of functions on the card is encoded inside
@@ -745,17 +781,18 @@ int mmc_attach_sdio(struct mmc_host *host, u32 ocr)
 			goto remove;
 
 		/*
-		 * Enable Runtime PM for this func
+		 * Enable Runtime PM for this func (if supported)
 		 */
-		pm_runtime_enable(&card->sdio_func[i]->dev);
+		if (host->caps & MMC_CAP_POWER_OFF_CARD)
+			pm_runtime_enable(&card->sdio_func[i]->dev);
 	}
 
-	mmc_release_host(host);
-
 	/*
 	 * First add the card to the driver model...
 	 */
+	mmc_release_host(host);
 	err = mmc_add_card(host->card);
+	mmc_claim_host(host);
 	if (err)
 		goto remove_added;
 
@@ -773,15 +810,17 @@ int mmc_attach_sdio(struct mmc_host *host, u32 ocr)
 
 remove_added:
 	/* Remove without lock if the device has been added. */
+	mmc_release_host(host);
 	mmc_sdio_remove(host);
 	mmc_claim_host(host);
 remove:
 	/* And with lock if it hasn't been added. */
+	mmc_release_host(host);
 	if (host->card)
 		mmc_sdio_remove(host);
+	mmc_claim_host(host);
 err:
 	mmc_detach_bus(host);
-	mmc_release_host(host);
 
 	printk(KERN_ERR "%s: error %d whilst initialising SDIO card\n",
 		mmc_hostname(host), err);
diff --git a/drivers/mmc/core/sdio_bus.c b/drivers/mmc/core/sdio_bus.c
index 2716c7ab6bbf..d29b9c36919a 100644
--- a/drivers/mmc/core/sdio_bus.c
+++ b/drivers/mmc/core/sdio_bus.c
@@ -17,6 +17,7 @@
 #include <linux/pm_runtime.h>
 
 #include <linux/mmc/card.h>
+#include <linux/mmc/host.h>
 #include <linux/mmc/sdio_func.h>
 
 #include "sdio_cis.h"
@@ -132,9 +133,11 @@ static int sdio_bus_probe(struct device *dev)
 	 * it should call pm_runtime_put_noidle() in its probe routine and
 	 * pm_runtime_get_noresume() in its remove routine.
 	 */
-	ret = pm_runtime_get_sync(dev);
-	if (ret < 0)
-		goto out;
+	if (func->card->host->caps & MMC_CAP_POWER_OFF_CARD) {
+		ret = pm_runtime_get_sync(dev);
+		if (ret < 0)
+			goto out;
+	}
 
 	/* Set the default block size so the driver is sure it's something
 	 * sensible. */
@@ -151,7 +154,8 @@ static int sdio_bus_probe(struct device *dev)
 	return 0;
 
 disable_runtimepm:
-	pm_runtime_put_noidle(dev);
+	if (func->card->host->caps & MMC_CAP_POWER_OFF_CARD)
+		pm_runtime_put_noidle(dev);
 out:
 	return ret;
 }
@@ -160,12 +164,14 @@ static int sdio_bus_remove(struct device *dev)
 {
 	struct sdio_driver *drv = to_sdio_driver(dev->driver);
 	struct sdio_func *func = dev_to_sdio_func(dev);
-	int ret;
+	int ret = 0;
 
 	/* Make sure card is powered before invoking ->remove() */
-	ret = pm_runtime_get_sync(dev);
-	if (ret < 0)
-		goto out;
+	if (func->card->host->caps & MMC_CAP_POWER_OFF_CARD) {
+		ret = pm_runtime_get_sync(dev);
+		if (ret < 0)
+			goto out;
+	}
 
 	drv->remove(func);
 
@@ -178,10 +184,12 @@ static int sdio_bus_remove(struct device *dev)
 	}
 
 	/* First, undo the increment made directly above */
-	pm_runtime_put_noidle(dev);
+	if (func->card->host->caps & MMC_CAP_POWER_OFF_CARD)
+		pm_runtime_put_noidle(dev);
 
 	/* Then undo the runtime PM settings in sdio_bus_probe() */
-	pm_runtime_put_noidle(dev);
+	if (func->card->host->caps & MMC_CAP_POWER_OFF_CARD)
+		pm_runtime_put_noidle(dev);
 
 out:
 	return ret;
@@ -189,41 +197,12 @@ out:
 
 #ifdef CONFIG_PM_RUNTIME
 
-static int sdio_bus_pm_prepare(struct device *dev)
-{
-	/*
-	 * Resume an SDIO device which was suspended at run time at this
-	 * point, in order to allow standard SDIO suspend/resume paths
-	 * to keep working as usual.
-	 *
-	 * Ultimately, the SDIO driver itself will decide (in its
-	 * suspend handler, or lack thereof) whether the card should be
-	 * removed or kept, and if kept, at what power state.
-	 *
-	 * At this point, PM core have increased our use count, so it's
-	 * safe to directly resume the device. After system is resumed
-	 * again, PM core will drop back its runtime PM use count, and if
-	 * needed device will be suspended again.
-	 *
-	 * The end result is guaranteed to be a power state that is
-	 * coherent with the device's runtime PM use count.
-	 *
-	 * The return value of pm_runtime_resume is deliberately unchecked
-	 * since there is little point in failing system suspend if a
-	 * device can't be resumed.
-	 */
-	pm_runtime_resume(dev);
-
-	return 0;
-}
-
 static const struct dev_pm_ops sdio_bus_pm_ops = {
 	SET_RUNTIME_PM_OPS(
 		pm_generic_runtime_suspend,
 		pm_generic_runtime_resume,
 		pm_generic_runtime_idle
 	)
-	.prepare = sdio_bus_pm_prepare,
 };
 
 #define SDIO_PM_OPS_PTR	(&sdio_bus_pm_ops)