drivers: hwspinlock: add framework

Add a platform-independent hwspinlock framework.

Hardware spinlock devices are needed, e.g., in order to access data
that is shared between remote processors, that otherwise have no
alternative mechanism to accomplish synchronization and mutual exclusion
operations.

Signed-off-by: Ohad Ben-Cohen <ohad@wizery.com>
Cc: Hari Kanigeri <h-kanigeri2@ti.com>
Cc: Benoit Cousson <b-cousson@ti.com>
Cc: Kevin Hilman <khilman@ti.com>
Cc: Grant Likely <grant.likely@secretlab.ca>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Russell King <linux@arm.linux.org.uk>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Tony Lindgren <tony@atomide.com>

8 files changed, 1216 insertions, 0 deletions

diff --git a/Documentation/hwspinlock.txt b/Documentation/hwspinlock.txt
new file mode 100644
index 000000000000..7dcd1a4e726c
--- /dev/null
+++ b/Documentation/hwspinlock.txt
@@ -0,0 +1,293 @@
+Hardware Spinlock Framework
+
+1. Introduction
+
+Hardware spinlock modules provide hardware assistance for synchronization
+and mutual exclusion between heterogeneous processors and those not operating
+under a single, shared operating system.
+
+For example, OMAP4 has dual Cortex-A9, dual Cortex-M3 and a C64x+ DSP,
+each of which is running a different Operating System (the master, A9,
+is usually running Linux and the slave processors, the M3 and the DSP,
+are running some flavor of RTOS).
+
+A generic hwspinlock framework allows platform-independent drivers to use
+the hwspinlock device in order to access data structures that are shared
+between remote processors, that otherwise have no alternative mechanism
+to accomplish synchronization and mutual exclusion operations.
+
+This is necessary, for example, for Inter-processor communications:
+on OMAP4, cpu-intensive multimedia tasks are offloaded by the host to the
+remote M3 and/or C64x+ slave processors (by an IPC subsystem called Syslink).
+
+To achieve fast message-based communications, a minimal kernel support
+is needed to deliver messages arriving from a remote processor to the
+appropriate user process.
+
+This communication is based on simple data structures that is shared between
+the remote processors, and access to it is synchronized using the hwspinlock
+module (remote processor directly places new messages in this shared data
+structure).
+
+A common hwspinlock interface makes it possible to have generic, platform-
+independent, drivers.
+
+2. User API
+
+  struct hwspinlock *hwspin_lock_request(void);
+   - dynamically assign an hwspinlock and return its address, or NULL
+     in case an unused hwspinlock isn't available. Users of this
+     API will usually want to communicate the lock's id to the remote core
+     before it can be used to achieve synchronization.
+     Can be called from an atomic context (this function will not sleep) but
+     not from within interrupt context.
+
+  struct hwspinlock *hwspin_lock_request_specific(unsigned int id);
+   - assign a specific hwspinlock id and return its address, or NULL
+     if that hwspinlock is already in use. Usually board code will
+     be calling this function in order to reserve specific hwspinlock
+     ids for predefined purposes.
+     Can be called from an atomic context (this function will not sleep) but
+     not from within interrupt context.
+
+  int hwspin_lock_free(struct hwspinlock *hwlock);
+   - free a previously-assigned hwspinlock; returns 0 on success, or an
+     appropriate error code on failure (e.g. -EINVAL if the hwspinlock
+     is already free).
+     Can be called from an atomic context (this function will not sleep) but
+     not from within interrupt context.
+
+  int hwspin_lock_timeout(struct hwspinlock *hwlock, unsigned int timeout);
+   - lock a previously-assigned hwspinlock with a timeout limit (specified in
+     msecs). If the hwspinlock is already taken, the function will busy loop
+     waiting for it to be released, but give up when the timeout elapses.
+     Upon a successful return from this function, preemption is disabled so
+     the caller must not sleep, and is advised to release the hwspinlock as
+     soon as possible, in order to minimize remote cores polling on the
+     hardware interconnect.
+     Returns 0 when successful and an appropriate error code otherwise (most
+     notably -ETIMEDOUT if the hwspinlock is still busy after timeout msecs).
+     The function will never sleep.
+
+  int hwspin_lock_timeout_irq(struct hwspinlock *hwlock, unsigned int timeout);
+   - lock a previously-assigned hwspinlock with a timeout limit (specified in
+     msecs). If the hwspinlock is already taken, the function will busy loop
+     waiting for it to be released, but give up when the timeout elapses.
+     Upon a successful return from this function, preemption and the local
+     interrupts are disabled, so the caller must not sleep, and is advised to
+     release the hwspinlock as soon as possible.
+     Returns 0 when successful and an appropriate error code otherwise (most
+     notably -ETIMEDOUT if the hwspinlock is still busy after timeout msecs).
+     The function will never sleep.
+
+  int hwspin_lock_timeout_irqsave(struct hwspinlock *hwlock, unsigned int to,
+                                                        unsigned long *flags);
+   - lock a previously-assigned hwspinlock with a timeout limit (specified in
+     msecs). If the hwspinlock is already taken, the function will busy loop
+     waiting for it to be released, but give up when the timeout elapses.
+     Upon a successful return from this function, preemption is disabled,
+     local interrupts are disabled and their previous state is saved at the
+     given flags placeholder. The caller must not sleep, and is advised to
+     release the hwspinlock as soon as possible.
+     Returns 0 when successful and an appropriate error code otherwise (most
+     notably -ETIMEDOUT if the hwspinlock is still busy after timeout msecs).
+     The function will never sleep.
+
+  int hwspin_trylock(struct hwspinlock *hwlock);
+   - attempt to lock a previously-assigned hwspinlock, but immediately fail if
+     it is already taken.
+     Upon a successful return from this function, preemption is disabled so
+     caller must not sleep, and is advised to release the hwspinlock as soon as
+     possible, in order to minimize remote cores polling on the hardware
+     interconnect.
+     Returns 0 on success and an appropriate error code otherwise (most
+     notably -EBUSY if the hwspinlock was already taken).
+     The function will never sleep.
+
+  int hwspin_trylock_irq(struct hwspinlock *hwlock);
+   - attempt to lock a previously-assigned hwspinlock, but immediately fail if
+     it is already taken.
+     Upon a successful return from this function, preemption and the local
+     interrupts are disabled so caller must not sleep, and is advised to
+     release the hwspinlock as soon as possible.
+     Returns 0 on success and an appropriate error code otherwise (most
+     notably -EBUSY if the hwspinlock was already taken).
+     The function will never sleep.
+
+  int hwspin_trylock_irqsave(struct hwspinlock *hwlock, unsigned long *flags);
+   - attempt to lock a previously-assigned hwspinlock, but immediately fail if
+     it is already taken.
+     Upon a successful return from this function, preemption is disabled,
+     the local interrupts are disabled and their previous state is saved
+     at the given flags placeholder. The caller must not sleep, and is advised
+     to release the hwspinlock as soon as possible.
+     Returns 0 on success and an appropriate error code otherwise (most
+     notably -EBUSY if the hwspinlock was already taken).
+     The function will never sleep.
+
+  void hwspin_unlock(struct hwspinlock *hwlock);
+   - unlock a previously-locked hwspinlock. Always succeed, and can be called
+     from any context (the function never sleeps). Note: code should _never_
+     unlock an hwspinlock which is already unlocked (there is no protection
+     against this).
+
+  void hwspin_unlock_irq(struct hwspinlock *hwlock);
+   - unlock a previously-locked hwspinlock and enable local interrupts.
+     The caller should _never_ unlock an hwspinlock which is already unlocked.
+     Doing so is considered a bug (there is no protection against this).
+     Upon a successful return from this function, preemption and local
+     interrupts are enabled. This function will never sleep.
+
+  void
+  hwspin_unlock_irqrestore(struct hwspinlock *hwlock, unsigned long *flags);
+   - unlock a previously-locked hwspinlock.
+     The caller should _never_ unlock an hwspinlock which is already unlocked.
+     Doing so is considered a bug (there is no protection against this).
+     Upon a successful return from this function, preemption is reenabled,
+     and the state of the local interrupts is restored to the state saved at
+     the given flags. This function will never sleep.
+
+  int hwspin_lock_get_id(struct hwspinlock *hwlock);
+   - retrieve id number of a given hwspinlock. This is needed when an
+     hwspinlock is dynamically assigned: before it can be used to achieve
+     mutual exclusion with a remote cpu, the id number should be communicated
+     to the remote task with which we want to synchronize.
+     Returns the hwspinlock id number, or -EINVAL if hwlock is null.
+
+3. Typical usage
+
+#include <linux/hwspinlock.h>
+#include <linux/err.h>
+
+int hwspinlock_example1(void)
+{
+        struct hwspinlock *hwlock;
+        int ret;
+
+        /* dynamically assign a hwspinlock */
+        hwlock = hwspin_lock_request();
+        if (!hwlock)
+                ...
+
+        id = hwspin_lock_get_id(hwlock);
+        /* probably need to communicate id to a remote processor now */
+
+        /* take the lock, spin for 1 sec if it's already taken */
+        ret = hwspin_lock_timeout(hwlock, 1000);
+        if (ret)
+                ...
+
+        /*
+         * we took the lock, do our thing now, but do NOT sleep
+         */
+
+        /* release the lock */
+        hwspin_unlock(hwlock);
+
+        /* free the lock */
+        ret = hwspin_lock_free(hwlock);
+        if (ret)
+                ...
+
+        return ret;
+}
+
+int hwspinlock_example2(void)
+{
+        struct hwspinlock *hwlock;
+        int ret;
+
+        /*
+         * assign a specific hwspinlock id - this should be called early
+         * by board init code.
+         */
+        hwlock = hwspin_lock_request_specific(PREDEFINED_LOCK_ID);
+        if (!hwlock)
+                ...
+
+        /* try to take it, but don't spin on it */
+        ret = hwspin_trylock(hwlock);
+        if (!ret) {
+                pr_info("lock is already taken\n");
+                return -EBUSY;
+        }
+
+        /*
+         * we took the lock, do our thing now, but do NOT sleep
+         */
+
+        /* release the lock */
+        hwspin_unlock(hwlock);
+
+        /* free the lock */
+        ret = hwspin_lock_free(hwlock);
+        if (ret)
+                ...
+
+        return ret;
+}
+
+
+4. API for implementors
+
+  int hwspin_lock_register(struct hwspinlock *hwlock);
+   - to be called from the underlying platform-specific implementation, in
+     order to register a new hwspinlock instance. Can be called from an atomic
+     context (this function will not sleep) but not from within interrupt
+     context. Returns 0 on success, or appropriate error code on failure.
+
+  struct hwspinlock *hwspin_lock_unregister(unsigned int id);
+   - to be called from the underlying vendor-specific implementation, in order
+     to unregister an existing (and unused) hwspinlock instance.
+     Can be called from an atomic context (will not sleep) but not from
+     within interrupt context.
+     Returns the address of hwspinlock on success, or NULL on error (e.g.
+     if the hwspinlock is sill in use).
+
+5. struct hwspinlock
+
+This struct represents an hwspinlock instance. It is registered by the
+underlying hwspinlock implementation using the hwspin_lock_register() API.
+
+/**
+ * struct hwspinlock - vendor-specific hwspinlock implementation
+ *
+ * @dev: underlying device, will be used with runtime PM api
+ * @ops: vendor-specific hwspinlock handlers
+ * @id: a global, unique, system-wide, index of the lock.
+ * @lock: initialized and used by hwspinlock core
+ * @owner: underlying implementation module, used to maintain module ref count
+ */
+struct hwspinlock {
+        struct device *dev;
+        const struct hwspinlock_ops *ops;
+        int id;
+        spinlock_t lock;
+        struct module *owner;
+};
+
+The underlying implementation is responsible to assign the dev, ops, id and
+owner members. The lock member, OTOH, is initialized and used by the hwspinlock
+core.
+
+6. Implementation callbacks
+
+There are three possible callbacks defined in 'struct hwspinlock_ops':
+
+struct hwspinlock_ops {
+        int (*trylock)(struct hwspinlock *lock);
+        void (*unlock)(struct hwspinlock *lock);
+        void (*relax)(struct hwspinlock *lock);
+};
+
+The first two callbacks are mandatory:
+
+The ->trylock() callback should make a single attempt to take the lock, and
+return 0 on failure and 1 on success. This callback may _not_ sleep.
+
+The ->unlock() callback releases the lock. It always succeed, and it, too,
+may _not_ sleep.
+
+The ->relax() callback is optional. It is called by hwspinlock core while
+spinning on a lock, and can be used by the underlying implementation to force
+a delay between two successive invocations of ->trylock(). It may _not_ sleep.
diff --git a/drivers/Kconfig b/drivers/Kconfig
index 9bfb71ff3a6a..177c7d156933 100644
--- a/drivers/Kconfig
+++ b/drivers/Kconfig
@@ -117,4 +117,6 @@ source "drivers/staging/Kconfig"
 source "drivers/platform/Kconfig"
 
 source "drivers/clk/Kconfig"
+
+source "drivers/hwspinlock/Kconfig"
 endmenu
diff --git a/drivers/Makefile b/drivers/Makefile
index b423bb16c3a8..3f135b6fb014 100644
--- a/drivers/Makefile
+++ b/drivers/Makefile
@@ -117,3 +117,5 @@ obj-y				+= platform/
 obj-y                           += ieee802154/
 #common clk code
 obj-y                           += clk/
+
+obj-$(CONFIG_HWSPINLOCK)        += hwspinlock/
diff --git a/drivers/hwspinlock/Kconfig b/drivers/hwspinlock/Kconfig
new file mode 100644
index 000000000000..9dd8db46606b
--- /dev/null
+++ b/drivers/hwspinlock/Kconfig
@@ -0,0 +1,13 @@
+#
+# Generic HWSPINLOCK framework
+#
+
+config HWSPINLOCK
+        tristate "Generic Hardware Spinlock framework"
+        help
+          Say y here to support the generic hardware spinlock framework.
+          You only need to enable this if you have hardware spinlock module
+          on your system (usually only relevant if your system has remote slave
+          coprocessors).
+
+          If unsure, say N.
diff --git a/drivers/hwspinlock/Makefile b/drivers/hwspinlock/Makefile
new file mode 100644
index 000000000000..b9d2b9f40491
--- /dev/null
+++ b/drivers/hwspinlock/Makefile
@@ -0,0 +1,5 @@
+#
+# Generic Hardware Spinlock framework
+#
+
+obj-$(CONFIG_HWSPINLOCK)                += hwspinlock_core.o
diff --git a/drivers/hwspinlock/hwspinlock_core.c b/drivers/hwspinlock/hwspinlock_core.c
new file mode 100644
index 000000000000..43a62714b4fb
--- /dev/null
+++ b/drivers/hwspinlock/hwspinlock_core.c
@@ -0,0 +1,548 @@
+/*
+ * Hardware spinlock framework
+ *
+ * Copyright (C) 2010 Texas Instruments Incorporated - http://www.ti.com
+ *
+ * Contact: Ohad Ben-Cohen <ohad@wizery.com>
+ *
+ * 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 published
+ * by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#define pr_fmt(fmt)    "%s: " fmt, __func__
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/spinlock.h>
+#include <linux/types.h>
+#include <linux/err.h>
+#include <linux/jiffies.h>
+#include <linux/radix-tree.h>
+#include <linux/hwspinlock.h>
+#include <linux/pm_runtime.h>
+
+#include "hwspinlock_internal.h"
+
+/* radix tree tags */
+#define HWSPINLOCK_UNUSED       (0) /* tags an hwspinlock as unused */
+
+/*
+ * A radix tree is used to maintain the available hwspinlock instances.
+ * The tree associates hwspinlock pointers with their integer key id,
+ * and provides easy-to-use API which makes the hwspinlock core code simple
+ * and easy to read.
+ *
+ * Radix trees are quick on lookups, and reasonably efficient in terms of
+ * storage, especially with high density usages such as this framework
+ * requires (a continuous range of integer keys, beginning with zero, is
+ * used as the ID's of the hwspinlock instances).
+ *
+ * The radix tree API supports tagging items in the tree, which this
+ * framework uses to mark unused hwspinlock instances (see the
+ * HWSPINLOCK_UNUSED tag above). As a result, the process of querying the
+ * tree, looking for an unused hwspinlock instance, is now reduced to a
+ * single radix tree API call.
+ */
+static RADIX_TREE(hwspinlock_tree, GFP_KERNEL);
+
+/*
+ * Synchronization of access to the tree is achieved using this spinlock,
+ * as the radix-tree API requires that users provide all synchronisation.
+ */
+static DEFINE_SPINLOCK(hwspinlock_tree_lock);
+
+/**
+ * __hwspin_trylock() - attempt to lock a specific hwspinlock
+ * @hwlock: an hwspinlock which we want to trylock
+ * @mode: controls whether local interrupts are disabled or not
+ * @flags: a pointer where the caller's interrupt state will be saved at (if
+ *         requested)
+ *
+ * This function attempts to lock an hwspinlock, and will immediately
+ * fail if the hwspinlock is already taken.
+ *
+ * Upon a successful return from this function, preemption (and possibly
+ * interrupts) is disabled, so the caller must not sleep, and is advised to
+ * release the hwspinlock as soon as possible. This is required in order to
+ * minimize remote cores polling on the hardware interconnect.
+ *
+ * The user decides whether local interrupts are disabled or not, and if yes,
+ * whether he wants their previous state to be saved. It is up to the user
+ * to choose the appropriate @mode of operation, exactly the same way users
+ * should decide between spin_trylock, spin_trylock_irq and
+ * spin_trylock_irqsave.
+ *
+ * Returns 0 if we successfully locked the hwspinlock or -EBUSY if
+ * the hwspinlock was already taken.
+ * This function will never sleep.
+ */
+int __hwspin_trylock(struct hwspinlock *hwlock, int mode, unsigned long *flags)
+{
+        int ret;
+
+        BUG_ON(!hwlock);
+        BUG_ON(!flags && mode == HWLOCK_IRQSTATE);
+
+        /*
+         * This spin_lock{_irq, _irqsave} serves three purposes:
+         *
+         * 1. Disable preemption, in order to minimize the period of time
+         *    in which the hwspinlock is taken. This is important in order
+         *    to minimize the possible polling on the hardware interconnect
+         *    by a remote user of this lock.
+         * 2. Make the hwspinlock SMP-safe (so we can take it from
+         *    additional contexts on the local host).
+         * 3. Ensure that in_atomic/might_sleep checks catch potential
+         *    problems with hwspinlock usage (e.g. scheduler checks like
+         *    'scheduling while atomic' etc.)
+         */
+        if (mode == HWLOCK_IRQSTATE)
+                ret = spin_trylock_irqsave(&hwlock->lock, *flags);
+        else if (mode == HWLOCK_IRQ)
+                ret = spin_trylock_irq(&hwlock->lock);
+        else
+                ret = spin_trylock(&hwlock->lock);
+
+        /* is lock already taken by another context on the local cpu ? */
+        if (!ret)
+                return -EBUSY;
+
+        /* try to take the hwspinlock device */
+        ret = hwlock->ops->trylock(hwlock);
+
+        /* if hwlock is already taken, undo spin_trylock_* and exit */
+        if (!ret) {
+                if (mode == HWLOCK_IRQSTATE)
+                        spin_unlock_irqrestore(&hwlock->lock, *flags);
+                else if (mode == HWLOCK_IRQ)
+                        spin_unlock_irq(&hwlock->lock);
+                else
+                        spin_unlock(&hwlock->lock);
+
+                return -EBUSY;
+        }
+
+        /*
+         * We can be sure the other core's memory operations
+         * are observable to us only _after_ we successfully take
+         * the hwspinlock, and we must make sure that subsequent memory
+         * operations (both reads and writes) will not be reordered before
+         * we actually took the hwspinlock.
+         *
+         * Note: the implicit memory barrier of the spinlock above is too
+         * early, so we need this additional explicit memory barrier.
+         */
+        mb();
+
+        return 0;
+}
+EXPORT_SYMBOL_GPL(__hwspin_trylock);
+
+/**
+ * __hwspin_lock_timeout() - lock an hwspinlock with timeout limit
+ * @hwlock: the hwspinlock to be locked
+ * @timeout: timeout value in msecs
+ * @mode: mode which controls whether local interrupts are disabled or not
+ * @flags: a pointer to where the caller's interrupt state will be saved at (if
+ *         requested)
+ *
+ * This function locks the given @hwlock. If the @hwlock
+ * is already taken, the function will busy loop waiting for it to
+ * be released, but give up after @timeout msecs have elapsed.
+ *
+ * Upon a successful return from this function, preemption is disabled
+ * (and possibly local interrupts, too), so the caller must not sleep,
+ * and is advised to release the hwspinlock as soon as possible.
+ * This is required in order to minimize remote cores polling on the
+ * hardware interconnect.
+ *
+ * The user decides whether local interrupts are disabled or not, and if yes,
+ * whether he wants their previous state to be saved. It is up to the user
+ * to choose the appropriate @mode of operation, exactly the same way users
+ * should decide between spin_lock, spin_lock_irq and spin_lock_irqsave.
+ *
+ * Returns 0 when the @hwlock was successfully taken, and an appropriate
+ * error code otherwise (most notably -ETIMEDOUT if the @hwlock is still
+ * busy after @timeout msecs). The function will never sleep.
+ */
+int __hwspin_lock_timeout(struct hwspinlock *hwlock, unsigned int to,
+                                        int mode, unsigned long *flags)
+{
+        int ret;
+        unsigned long expire;
+
+        expire = msecs_to_jiffies(to) + jiffies;
+
+        for (;;) {
+                /* Try to take the hwspinlock */
+                ret = __hwspin_trylock(hwlock, mode, flags);
+                if (ret != -EBUSY)
+                        break;
+
+                /*
+                 * The lock is already taken, let's check if the user wants
+                 * us to try again
+                 */
+                if (time_is_before_eq_jiffies(expire))
+                        return -ETIMEDOUT;
+
+                /*
+                 * Allow platform-specific relax handlers to prevent
+                 * hogging the interconnect (no sleeping, though)
+                 */
+                if (hwlock->ops->relax)
+                        hwlock->ops->relax(hwlock);
+        }
+
+        return ret;
+}
+EXPORT_SYMBOL_GPL(__hwspin_lock_timeout);
+
+/**
+ * __hwspin_unlock() - unlock a specific hwspinlock
+ * @hwlock: a previously-acquired hwspinlock which we want to unlock
+ * @mode: controls whether local interrupts needs to be restored or not
+ * @flags: previous caller's interrupt state to restore (if requested)
+ *
+ * This function will unlock a specific hwspinlock, enable preemption and
+ * (possibly) enable interrupts or restore their previous state.
+ * @hwlock must be already locked before calling this function: it is a bug
+ * to call unlock on a @hwlock that is already unlocked.
+ *
+ * The user decides whether local interrupts should be enabled or not, and
+ * if yes, whether he wants their previous state to be restored. It is up
+ * to the user to choose the appropriate @mode of operation, exactly the
+ * same way users decide between spin_unlock, spin_unlock_irq and
+ * spin_unlock_irqrestore.
+ *
+ * The function will never sleep.
+ */
+void __hwspin_unlock(struct hwspinlock *hwlock, int mode, unsigned long *flags)
+{
+        BUG_ON(!hwlock);
+        BUG_ON(!flags && mode == HWLOCK_IRQSTATE);
+
+        /*
+         * We must make sure that memory operations (both reads and writes),
+         * done before unlocking the hwspinlock, will not be reordered
+         * after the lock is released.
+         *
+         * That's the purpose of this explicit memory barrier.
+         *
+         * Note: the memory barrier induced by the spin_unlock below is too
+         * late; the other core is going to access memory soon after it will
+         * take the hwspinlock, and by then we want to be sure our memory
+         * operations are already observable.
+         */
+        mb();
+
+        hwlock->ops->unlock(hwlock);
+
+        /* Undo the spin_trylock{_irq, _irqsave} called while locking */
+        if (mode == HWLOCK_IRQSTATE)
+                spin_unlock_irqrestore(&hwlock->lock, *flags);
+        else if (mode == HWLOCK_IRQ)
+                spin_unlock_irq(&hwlock->lock);
+        else
+                spin_unlock(&hwlock->lock);
+}
+EXPORT_SYMBOL_GPL(__hwspin_unlock);
+
+/**
+ * hwspin_lock_register() - register a new hw spinlock
+ * @hwlock: hwspinlock to register.
+ *
+ * This function should be called from the underlying platform-specific
+ * implementation, to register a new hwspinlock instance.
+ *
+ * Can be called from an atomic context (will not sleep) but not from
+ * within interrupt context.
+ *
+ * Returns 0 on success, or an appropriate error code on failure
+ */
+int hwspin_lock_register(struct hwspinlock *hwlock)
+{
+        struct hwspinlock *tmp;
+        int ret;
+
+        if (!hwlock || !hwlock->ops ||
+                !hwlock->ops->trylock || !hwlock->ops->unlock) {
+                pr_err("invalid parameters\n");
+                return -EINVAL;
+        }
+
+        spin_lock_init(&hwlock->lock);
+
+        spin_lock(&hwspinlock_tree_lock);
+
+        ret = radix_tree_insert(&hwspinlock_tree, hwlock->id, hwlock);
+        if (ret)
+                goto out;
+
+        /* mark this hwspinlock as available */
+        tmp = radix_tree_tag_set(&hwspinlock_tree, hwlock->id,
+                                                        HWSPINLOCK_UNUSED);
+
+        /* self-sanity check which should never fail */
+        WARN_ON(tmp != hwlock);
+
+out:
+        spin_unlock(&hwspinlock_tree_lock);
+        return ret;
+}
+EXPORT_SYMBOL_GPL(hwspin_lock_register);
+
+/**
+ * hwspin_lock_unregister() - unregister an hw spinlock
+ * @id: index of the specific hwspinlock to unregister
+ *
+ * This function should be called from the underlying platform-specific
+ * implementation, to unregister an existing (and unused) hwspinlock.
+ *
+ * Can be called from an atomic context (will not sleep) but not from
+ * within interrupt context.
+ *
+ * Returns the address of hwspinlock @id on success, or NULL on failure
+ */
+struct hwspinlock *hwspin_lock_unregister(unsigned int id)
+{
+        struct hwspinlock *hwlock = NULL;
+        int ret;
+
+        spin_lock(&hwspinlock_tree_lock);
+
+        /* make sure the hwspinlock is not in use (tag is set) */
+        ret = radix_tree_tag_get(&hwspinlock_tree, id, HWSPINLOCK_UNUSED);
+        if (ret == 0) {
+                pr_err("hwspinlock %d still in use (or not present)\n", id);
+                goto out;
+        }
+
+        hwlock = radix_tree_delete(&hwspinlock_tree, id);
+        if (!hwlock) {
+                pr_err("failed to delete hwspinlock %d\n", id);
+                goto out;
+        }
+
+out:
+        spin_unlock(&hwspinlock_tree_lock);
+        return hwlock;
+}
+EXPORT_SYMBOL_GPL(hwspin_lock_unregister);
+
+/**
+ * __hwspin_lock_request() - tag an hwspinlock as used and power it up
+ *
+ * This is an internal function that prepares an hwspinlock instance
+ * before it is given to the user. The function assumes that
+ * hwspinlock_tree_lock is taken.
+ *
+ * Returns 0 or positive to indicate success, and a negative value to
+ * indicate an error (with the appropriate error code)
+ */
+static int __hwspin_lock_request(struct hwspinlock *hwlock)
+{
+        struct hwspinlock *tmp;
+        int ret;
+
+        /* prevent underlying implementation from being removed */
+        if (!try_module_get(hwlock->owner)) {
+                dev_err(hwlock->dev, "%s: can't get owner\n", __func__);
+                return -EINVAL;
+        }
+
+        /* notify PM core that power is now needed */
+        ret = pm_runtime_get_sync(hwlock->dev);
+        if (ret < 0) {
+                dev_err(hwlock->dev, "%s: can't power on device\n", __func__);
+                return ret;
+        }
+
+        /* mark hwspinlock as used, should not fail */
+        tmp = radix_tree_tag_clear(&hwspinlock_tree, hwlock->id,
+                                                        HWSPINLOCK_UNUSED);
+
+        /* self-sanity check that should never fail */
+        WARN_ON(tmp != hwlock);
+
+        return ret;
+}
+
+/**
+ * hwspin_lock_get_id() - retrieve id number of a given hwspinlock
+ * @hwlock: a valid hwspinlock instance
+ *
+ * Returns the id number of a given @hwlock, or -EINVAL if @hwlock is invalid.
+ */
+int hwspin_lock_get_id(struct hwspinlock *hwlock)
+{
+        if (!hwlock) {
+                pr_err("invalid hwlock\n");
+                return -EINVAL;
+        }
+
+        return hwlock->id;
+}
+EXPORT_SYMBOL_GPL(hwspin_lock_get_id);
+
+/**
+ * hwspin_lock_request() - request an hwspinlock
+ *
+ * This function should be called by users of the hwspinlock device,
+ * in order to dynamically assign them an unused hwspinlock.
+ * Usually the user of this lock will then have to communicate the lock's id
+ * to the remote core before it can be used for synchronization (to get the
+ * id of a given hwlock, use hwspin_lock_get_id()).
+ *
+ * Can be called from an atomic context (will not sleep) but not from
+ * within interrupt context (simply because there is no use case for
+ * that yet).
+ *
+ * Returns the address of the assigned hwspinlock, or NULL on error
+ */
+struct hwspinlock *hwspin_lock_request(void)
+{
+        struct hwspinlock *hwlock;
+        int ret;
+
+        spin_lock(&hwspinlock_tree_lock);
+
+        /* look for an unused lock */
+        ret = radix_tree_gang_lookup_tag(&hwspinlock_tree, (void **)&hwlock,
+                                                0, 1, HWSPINLOCK_UNUSED);
+        if (ret == 0) {
+                pr_warn("a free hwspinlock is not available\n");
+                hwlock = NULL;
+                goto out;
+        }
+
+        /* sanity check that should never fail */
+        WARN_ON(ret > 1);
+
+        /* mark as used and power up */
+        ret = __hwspin_lock_request(hwlock);
+        if (ret < 0)
+                hwlock = NULL;
+
+out:
+        spin_unlock(&hwspinlock_tree_lock);
+        return hwlock;
+}
+EXPORT_SYMBOL_GPL(hwspin_lock_request);
+
+/**
+ * hwspin_lock_request_specific() - request for a specific hwspinlock
+ * @id: index of the specific hwspinlock that is requested
+ *
+ * This function should be called by users of the hwspinlock module,
+ * in order to assign them a specific hwspinlock.
+ * Usually early board code will be calling this function in order to
+ * reserve specific hwspinlock ids for predefined purposes.
+ *
+ * Can be called from an atomic context (will not sleep) but not from
+ * within interrupt context (simply because there is no use case for
+ * that yet).
+ *
+ * Returns the address of the assigned hwspinlock, or NULL on error
+ */
+struct hwspinlock *hwspin_lock_request_specific(unsigned int id)
+{
+        struct hwspinlock *hwlock;
+        int ret;
+
+        spin_lock(&hwspinlock_tree_lock);
+
+        /* make sure this hwspinlock exists */
+        hwlock = radix_tree_lookup(&hwspinlock_tree, id);
+        if (!hwlock) {
+                pr_warn("hwspinlock %u does not exist\n", id);
+                goto out;
+        }
+
+        /* sanity check (this shouldn't happen) */
+        WARN_ON(hwlock->id != id);
+
+        /* make sure this hwspinlock is unused */
+        ret = radix_tree_tag_get(&hwspinlock_tree, id, HWSPINLOCK_UNUSED);
+        if (ret == 0) {
+                pr_warn("hwspinlock %u is already in use\n", id);
+                hwlock = NULL;
+                goto out;
+        }
+
+        /* mark as used and power up */
+        ret = __hwspin_lock_request(hwlock);
+        if (ret < 0)
+                hwlock = NULL;
+
+out:
+        spin_unlock(&hwspinlock_tree_lock);
+        return hwlock;
+}
+EXPORT_SYMBOL_GPL(hwspin_lock_request_specific);
+
+/**
+ * hwspin_lock_free() - free a specific hwspinlock
+ * @hwlock: the specific hwspinlock to free
+ *
+ * This function mark @hwlock as free again.
+ * Should only be called with an @hwlock that was retrieved from
+ * an earlier call to omap_hwspin_lock_request{_specific}.
+ *
+ * Can be called from an atomic context (will not sleep) but not from
+ * within interrupt context (simply because there is no use case for
+ * that yet).
+ *
+ * Returns 0 on success, or an appropriate error code on failure
+ */
+int hwspin_lock_free(struct hwspinlock *hwlock)
+{
+        struct hwspinlock *tmp;
+        int ret;
+
+        if (!hwlock) {
+                pr_err("invalid hwlock\n");
+                return -EINVAL;
+        }
+
+        spin_lock(&hwspinlock_tree_lock);
+
+        /* make sure the hwspinlock is used */
+        ret = radix_tree_tag_get(&hwspinlock_tree, hwlock->id,
+                                                        HWSPINLOCK_UNUSED);
+        if (ret == 1) {
+                dev_err(hwlock->dev, "%s: hwlock is already free\n", __func__);
+                dump_stack();
+                ret = -EINVAL;
+                goto out;
+        }
+
+        /* notify the underlying device that power is not needed */
+        ret = pm_runtime_put(hwlock->dev);
+        if (ret < 0)
+                goto out;
+
+        /* mark this hwspinlock as available */
+        tmp = radix_tree_tag_set(&hwspinlock_tree, hwlock->id,
+                                                        HWSPINLOCK_UNUSED);
+
+        /* sanity check (this shouldn't happen) */
+        WARN_ON(tmp != hwlock);
+
+        module_put(hwlock->owner);
+
+out:
+        spin_unlock(&hwspinlock_tree_lock);
+        return ret;
+}
+EXPORT_SYMBOL_GPL(hwspin_lock_free);
+
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("Hardware spinlock interface");
+MODULE_AUTHOR("Ohad Ben-Cohen <ohad@wizery.com>");
diff --git a/drivers/hwspinlock/hwspinlock_internal.h b/drivers/hwspinlock/hwspinlock_internal.h
new file mode 100644
index 000000000000..69935e6b93e5
--- /dev/null
+++ b/drivers/hwspinlock/hwspinlock_internal.h
@@ -0,0 +1,61 @@
+/*
+ * Hardware spinlocks internal header
+ *
+ * Copyright (C) 2010 Texas Instruments Incorporated - http://www.ti.com
+ *
+ * Contact: Ohad Ben-Cohen <ohad@wizery.com>
+ *
+ * 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 published
+ * by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef __HWSPINLOCK_HWSPINLOCK_H
+#define __HWSPINLOCK_HWSPINLOCK_H
+
+#include <linux/spinlock.h>
+#include <linux/device.h>
+
+/**
+ * struct hwspinlock_ops - platform-specific hwspinlock handlers
+ *
+ * @trylock: make a single attempt to take the lock. returns 0 on
+ *           failure and true on success. may _not_ sleep.
+ * @unlock:  release the lock. always succeed. may _not_ sleep.
+ * @relax:   optional, platform-specific relax handler, called by hwspinlock
+ *           core while spinning on a lock, between two successive
+ *           invocations of @trylock. may _not_ sleep.
+ */
+struct hwspinlock_ops {
+        int (*trylock)(struct hwspinlock *lock);
+        void (*unlock)(struct hwspinlock *lock);
+        void (*relax)(struct hwspinlock *lock);
+};
+
+/**
+ * struct hwspinlock - this struct represents a single hwspinlock instance
+ *
+ * @dev: underlying device, will be used to invoke runtime PM api
+ * @ops: platform-specific hwspinlock handlers
+ * @id: a global, unique, system-wide, index of the lock.
+ * @lock: initialized and used by hwspinlock core
+ * @owner: underlying implementation module, used to maintain module ref count
+ *
+ * Note: currently simplicity was opted for, but later we can squeeze some
+ * memory bytes by grouping the dev, ops and owner members in a single
+ * per-platform struct, and have all hwspinlocks point at it.
+ */
+struct hwspinlock {
+        struct device *dev;
+        const struct hwspinlock_ops *ops;
+        int id;
+        spinlock_t lock;
+        struct module *owner;
+};
+
+#endif /* __HWSPINLOCK_HWSPINLOCK_H */
diff --git a/include/linux/hwspinlock.h b/include/linux/hwspinlock.h
new file mode 100644
index 000000000000..8390efc457eb
--- /dev/null
+++ b/include/linux/hwspinlock.h
@@ -0,0 +1,292 @@
+/*
+ * Hardware spinlock public header
+ *
+ * Copyright (C) 2010 Texas Instruments Incorporated - http://www.ti.com
+ *
+ * Contact: Ohad Ben-Cohen <ohad@wizery.com>
+ *
+ * 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 published
+ * by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#ifndef __LINUX_HWSPINLOCK_H
+#define __LINUX_HWSPINLOCK_H
+
+#include <linux/err.h>
+#include <linux/sched.h>
+
+/* hwspinlock mode argument */
+#define HWLOCK_IRQSTATE 0x01    /* Disable interrupts, save state */
+#define HWLOCK_IRQ      0x02    /* Disable interrupts, don't save state */
+
+struct hwspinlock;
+
+#if defined(CONFIG_HWSPINLOCK) || defined(CONFIG_HWSPINLOCK_MODULE)
+
+int hwspin_lock_register(struct hwspinlock *lock);
+struct hwspinlock *hwspin_lock_unregister(unsigned int id);
+struct hwspinlock *hwspin_lock_request(void);
+struct hwspinlock *hwspin_lock_request_specific(unsigned int id);
+int hwspin_lock_free(struct hwspinlock *hwlock);
+int hwspin_lock_get_id(struct hwspinlock *hwlock);
+int __hwspin_lock_timeout(struct hwspinlock *, unsigned int, int,
+                                                        unsigned long *);
+int __hwspin_trylock(struct hwspinlock *, int, unsigned long *);
+void __hwspin_unlock(struct hwspinlock *, int, unsigned long *);
+
+#else /* !CONFIG_HWSPINLOCK */
+
+/*
+ * We don't want these functions to fail if CONFIG_HWSPINLOCK is not
+ * enabled. We prefer to silently succeed in this case, and let the
+ * code path get compiled away. This way, if CONFIG_HWSPINLOCK is not
+ * required on a given setup, users will still work.
+ *
+ * The only exception is hwspin_lock_register/hwspin_lock_unregister, with which
+ * we _do_ want users to fail (no point in registering hwspinlock instances if
+ * the framework is not available).
+ *
+ * Note: ERR_PTR(-ENODEV) will still be considered a success for NULL-checking
+ * users. Others, which care, can still check this with IS_ERR.
+ */
+static inline struct hwspinlock *hwspin_lock_request(void)
+{
+        return ERR_PTR(-ENODEV);
+}
+
+static inline struct hwspinlock *hwspin_lock_request_specific(unsigned int id)
+{
+        return ERR_PTR(-ENODEV);
+}
+
+static inline int hwspin_lock_free(struct hwspinlock *hwlock)
+{
+        return 0;
+}
+
+static inline
+int __hwspin_lock_timeout(struct hwspinlock *hwlock, unsigned int to,
+                                        int mode, unsigned long *flags)
+{
+        return 0;
+}
+
+static inline
+int __hwspin_trylock(struct hwspinlock *hwlock, int mode, unsigned long *flags)
+{
+        return 0;
+}
+
+static inline
+void __hwspin_unlock(struct hwspinlock *hwlock, int mode, unsigned long *flags)
+{
+        return 0;
+}
+
+static inline int hwspin_lock_get_id(struct hwspinlock *hwlock)
+{
+        return 0;
+}
+
+static inline int hwspin_lock_register(struct hwspinlock *hwlock)
+{
+        return -ENODEV;
+}
+
+static inline struct hwspinlock *hwspin_lock_unregister(unsigned int id)
+{
+        return NULL;
+}
+
+#endif /* !CONFIG_HWSPINLOCK */
+
+/**
+ * hwspin_trylock_irqsave() - try to lock an hwspinlock, disable interrupts
+ * @hwlock: an hwspinlock which we want to trylock
+ * @flags: a pointer to where the caller's interrupt state will be saved at
+ *
+ * This function attempts to lock the underlying hwspinlock, and will
+ * immediately fail if the hwspinlock is already locked.
+ *
+ * Upon a successful return from this function, preemption and local
+ * interrupts are disabled (previous interrupts state is saved at @flags),
+ * so the caller must not sleep, and is advised to release the hwspinlock
+ * as soon as possible.
+ *
+ * Returns 0 if we successfully locked the hwspinlock, -EBUSY if
+ * the hwspinlock was already taken, and -EINVAL if @hwlock is invalid.
+ */
+static inline
+int hwspin_trylock_irqsave(struct hwspinlock *hwlock, unsigned long *flags)
+{
+        return __hwspin_trylock(hwlock, HWLOCK_IRQSTATE, flags);
+}
+
+/**
+ * hwspin_trylock_irq() - try to lock an hwspinlock, disable interrupts
+ * @hwlock: an hwspinlock which we want to trylock
+ *
+ * This function attempts to lock the underlying hwspinlock, and will
+ * immediately fail if the hwspinlock is already locked.
+ *
+ * Upon a successful return from this function, preemption and local
+ * interrupts are disabled, so the caller must not sleep, and is advised
+ * to release the hwspinlock as soon as possible.
+ *
+ * Returns 0 if we successfully locked the hwspinlock, -EBUSY if
+ * the hwspinlock was already taken, and -EINVAL if @hwlock is invalid.
+ */
+static inline int hwspin_trylock_irq(struct hwspinlock *hwlock)
+{
+        return __hwspin_trylock(hwlock, HWLOCK_IRQ, NULL);
+}
+
+/**
+ * hwspin_trylock() - attempt to lock a specific hwspinlock
+ * @hwlock: an hwspinlock which we want to trylock
+ *
+ * This function attempts to lock an hwspinlock, and will immediately fail
+ * if the hwspinlock is already taken.
+ *
+ * Upon a successful return from this function, preemption is disabled,
+ * so the caller must not sleep, and is advised to release the hwspinlock
+ * as soon as possible. This is required in order to minimize remote cores
+ * polling on the hardware interconnect.
+ *
+ * Returns 0 if we successfully locked the hwspinlock, -EBUSY if
+ * the hwspinlock was already taken, and -EINVAL if @hwlock is invalid.
+ */
+static inline int hwspin_trylock(struct hwspinlock *hwlock)
+{
+        return __hwspin_trylock(hwlock, 0, NULL);
+}
+
+/**
+ * hwspin_lock_timeout_irqsave() - lock hwspinlock, with timeout, disable irqs
+ * @hwlock: the hwspinlock to be locked
+ * @to: timeout value in msecs
+ * @flags: a pointer to where the caller's interrupt state will be saved at
+ *
+ * This function locks the underlying @hwlock. If the @hwlock
+ * is already taken, the function will busy loop waiting for it to
+ * be released, but give up when @timeout msecs have elapsed.
+ *
+ * Upon a successful return from this function, preemption and local interrupts
+ * are disabled (plus previous interrupt state is saved), so the caller must
+ * not sleep, and is advised to release the hwspinlock as soon as possible.
+ *
+ * Returns 0 when the @hwlock was successfully taken, and an appropriate
+ * error code otherwise (most notably an -ETIMEDOUT if the @hwlock is still
+ * busy after @timeout msecs). The function will never sleep.
+ */
+static inline int hwspin_lock_timeout_irqsave(struct hwspinlock *hwlock,
+                                unsigned int to, unsigned long *flags)
+{
+        return __hwspin_lock_timeout(hwlock, to, HWLOCK_IRQSTATE, flags);
+}
+
+/**
+ * hwspin_lock_timeout_irq() - lock hwspinlock, with timeout, disable irqs
+ * @hwlock: the hwspinlock to be locked
+ * @to: timeout value in msecs
+ *
+ * This function locks the underlying @hwlock. If the @hwlock
+ * is already taken, the function will busy loop waiting for it to
+ * be released, but give up when @timeout msecs have elapsed.
+ *
+ * Upon a successful return from this function, preemption and local interrupts
+ * are disabled so the caller must not sleep, and is advised to release the
+ * hwspinlock as soon as possible.
+ *
+ * Returns 0 when the @hwlock was successfully taken, and an appropriate
+ * error code otherwise (most notably an -ETIMEDOUT if the @hwlock is still
+ * busy after @timeout msecs). The function will never sleep.
+ */
+static inline
+int hwspin_lock_timeout_irq(struct hwspinlock *hwlock, unsigned int to)
+{
+        return __hwspin_lock_timeout(hwlock, to, HWLOCK_IRQ, NULL);
+}
+
+/**
+ * hwspin_lock_timeout() - lock an hwspinlock with timeout limit
+ * @hwlock: the hwspinlock to be locked
+ * @to: timeout value in msecs
+ *
+ * This function locks the underlying @hwlock. If the @hwlock
+ * is already taken, the function will busy loop waiting for it to
+ * be released, but give up when @timeout msecs have elapsed.
+ *
+ * Upon a successful return from this function, preemption is disabled
+ * so the caller must not sleep, and is advised to release the hwspinlock
+ * as soon as possible.
+ * This is required in order to minimize remote cores polling on the
+ * hardware interconnect.
+ *
+ * Returns 0 when the @hwlock was successfully taken, and an appropriate
+ * error code otherwise (most notably an -ETIMEDOUT if the @hwlock is still
+ * busy after @timeout msecs). The function will never sleep.
+ */
+static inline
+int hwspin_lock_timeout(struct hwspinlock *hwlock, unsigned int to)
+{
+        return __hwspin_lock_timeout(hwlock, to, 0, NULL);
+}
+
+/**
+ * hwspin_unlock_irqrestore() - unlock hwspinlock, restore irq state
+ * @hwlock: a previously-acquired hwspinlock which we want to unlock
+ * @flags: previous caller's interrupt state to restore
+ *
+ * This function will unlock a specific hwspinlock, enable preemption and
+ * restore the previous state of the local interrupts. It should be used
+ * to undo, e.g., hwspin_trylock_irqsave().
+ *
+ * @hwlock must be already locked before calling this function: it is a bug
+ * to call unlock on a @hwlock that is already unlocked.
+ */
+static inline void hwspin_unlock_irqrestore(struct hwspinlock *hwlock,
+                                                        unsigned long *flags)
+{
+        __hwspin_unlock(hwlock, HWLOCK_IRQSTATE, flags);
+}
+
+/**
+ * hwspin_unlock_irq() - unlock hwspinlock, enable interrupts
+ * @hwlock: a previously-acquired hwspinlock which we want to unlock
+ *
+ * This function will unlock a specific hwspinlock, enable preemption and
+ * enable local interrupts. Should be used to undo hwspin_lock_irq().
+ *
+ * @hwlock must be already locked (e.g. by hwspin_trylock_irq()) before
+ * calling this function: it is a bug to call unlock on a @hwlock that is
+ * already unlocked.
+ */
+static inline void hwspin_unlock_irq(struct hwspinlock *hwlock)
+{
+        __hwspin_unlock(hwlock, HWLOCK_IRQ, NULL);
+}
+
+/**
+ * hwspin_unlock() - unlock hwspinlock
+ * @hwlock: a previously-acquired hwspinlock which we want to unlock
+ *
+ * This function will unlock a specific hwspinlock and enable preemption
+ * back.
+ *
+ * @hwlock must be already locked (e.g. by hwspin_trylock()) before calling
+ * this function: it is a bug to call unlock on a @hwlock that is already
+ * unlocked.
+ */
+static inline void hwspin_unlock(struct hwspinlock *hwlock)
+{
+        __hwspin_unlock(hwlock, 0, NULL);
+}
+
+#endif /* __LINUX_HWSPINLOCK_H */