x86: split spinlock implementations out into their own files

ftrace requires certain low-level code, like spinlocks and timestamps,
to be compiled without -pg in order to avoid infinite recursion.  This
patch splits out the core paravirt spinlocks and the Xen spinlocks
into separate files which can be compiled without -pg.

Also do xen/time.c while we're about it.  As a result, we can now use
ftrace within a Xen domain.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

7 files changed, 226 insertions, 193 deletions

diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile
index 3db651fc8ec5..d679cb2c79b4 100644
--- a/arch/x86/kernel/Makefile
+++ b/arch/x86/kernel/Makefile
@@ -10,7 +10,7 @@ ifdef CONFIG_FTRACE
 # Do not profile debug and lowlevel utilities
 CFLAGS_REMOVE_tsc.o = -pg
 CFLAGS_REMOVE_rtc.o = -pg
-CFLAGS_REMOVE_paravirt.o = -pg
+CFLAGS_REMOVE_paravirt-spinlocks.o = -pg
 endif
 
 #
@@ -89,7 +89,7 @@ obj-$(CONFIG_DEBUG_NX_TEST)	+= test_nx.o
 obj-$(CONFIG_VMI)               += vmi_32.o vmiclock_32.o
 obj-$(CONFIG_KVM_GUEST)         += kvm.o
 obj-$(CONFIG_KVM_CLOCK)         += kvmclock.o
-obj-$(CONFIG_PARAVIRT)          += paravirt.o paravirt_patch_$(BITS).o
+obj-$(CONFIG_PARAVIRT)          += paravirt.o paravirt_patch_$(BITS).o paravirt-spinlocks.o
 obj-$(CONFIG_PARAVIRT_CLOCK)    += pvclock.o
 
 obj-$(CONFIG_PCSPKR_PLATFORM)   += pcspeaker.o
diff --git a/arch/x86/kernel/paravirt-spinlocks.c b/arch/x86/kernel/paravirt-spinlocks.c
new file mode 100644
index 000000000000..38d7f7f1dbc9
--- /dev/null
+++ b/arch/x86/kernel/paravirt-spinlocks.c
@@ -0,0 +1,31 @@
+/*
+ * Split spinlock implementation out into its own file, so it can be
+ * compiled in a FTRACE-compatible way.
+ */
+#include <linux/spinlock.h>
+#include <linux/module.h>
+
+#include <asm/paravirt.h>
+
+struct pv_lock_ops pv_lock_ops = {
+#ifdef CONFIG_SMP
+        .spin_is_locked = __ticket_spin_is_locked,
+        .spin_is_contended = __ticket_spin_is_contended,
+
+        .spin_lock = __ticket_spin_lock,
+        .spin_trylock = __ticket_spin_trylock,
+        .spin_unlock = __ticket_spin_unlock,
+#endif
+};
+EXPORT_SYMBOL_GPL(pv_lock_ops);
+
+void __init paravirt_use_bytelocks(void)
+{
+#ifdef CONFIG_SMP
+        pv_lock_ops.spin_is_locked = __byte_spin_is_locked;
+        pv_lock_ops.spin_is_contended = __byte_spin_is_contended;
+        pv_lock_ops.spin_lock = __byte_spin_lock;
+        pv_lock_ops.spin_trylock = __byte_spin_trylock;
+        pv_lock_ops.spin_unlock = __byte_spin_unlock;
+#endif
+}
diff --git a/arch/x86/kernel/paravirt.c b/arch/x86/kernel/paravirt.c
index 94da4d52d798..0d71de9ff56d 100644
--- a/arch/x86/kernel/paravirt.c
+++ b/arch/x86/kernel/paravirt.c
@@ -268,17 +268,6 @@ enum paravirt_lazy_mode paravirt_get_lazy_mode(void)
         return __get_cpu_var(paravirt_lazy_mode);
 }
 
-void __init paravirt_use_bytelocks(void)
-{
-#ifdef CONFIG_SMP
-        pv_lock_ops.spin_is_locked = __byte_spin_is_locked;
-        pv_lock_ops.spin_is_contended = __byte_spin_is_contended;
-        pv_lock_ops.spin_lock = __byte_spin_lock;
-        pv_lock_ops.spin_trylock = __byte_spin_trylock;
-        pv_lock_ops.spin_unlock = __byte_spin_unlock;
-#endif
-}
-
 struct pv_info pv_info = {
         .name = "bare hardware",
         .paravirt_enabled = 0,
@@ -461,18 +450,6 @@ struct pv_mmu_ops pv_mmu_ops = {
         .set_fixmap = native_set_fixmap,
 };
 
-struct pv_lock_ops pv_lock_ops = {
-#ifdef CONFIG_SMP
-        .spin_is_locked = __ticket_spin_is_locked,
-        .spin_is_contended = __ticket_spin_is_contended,
-
-        .spin_lock = __ticket_spin_lock,
-        .spin_trylock = __ticket_spin_trylock,
-        .spin_unlock = __ticket_spin_unlock,
-#endif
-};
-EXPORT_SYMBOL_GPL(pv_lock_ops);
-
 EXPORT_SYMBOL_GPL(pv_time_ops);
 EXPORT_SYMBOL    (pv_cpu_ops);
 EXPORT_SYMBOL    (pv_mmu_ops);
diff --git a/arch/x86/xen/Makefile b/arch/x86/xen/Makefile
index 59c1e539aed2..5bfee243cf9a 100644
--- a/arch/x86/xen/Makefile
+++ b/arch/x86/xen/Makefile
@@ -1,4 +1,10 @@
+ifdef CONFIG_FTRACE
+# Do not profile debug and lowlevel utilities
+CFLAGS_REMOVE_spinlock.o = -pg
+CFLAGS_REMOVE_time.o = -pg
+endif
+
 obj-y           := enlighten.o setup.o multicalls.o mmu.o \
                         time.o xen-asm_$(BITS).o grant-table.o suspend.o
 
-obj-$(CONFIG_SMP)       += smp.o
+obj-$(CONFIG_SMP)       += smp.o spinlock.o
diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c
index d8faf79a0a1d..baca7f2fbd8a 100644
--- a/arch/x86/xen/smp.c
+++ b/arch/x86/xen/smp.c
@@ -15,7 +15,6 @@
  * This does not handle HOTPLUG_CPU yet.
  */
 #include <linux/sched.h>
-#include <linux/kernel_stat.h>
 #include <linux/err.h>
 #include <linux/smp.h>
 
@@ -36,8 +35,6 @@
 #include "xen-ops.h"
 #include "mmu.h"
 
-static void __cpuinit xen_init_lock_cpu(int cpu);
-
 cpumask_t xen_cpu_initialized_map;
 
 static DEFINE_PER_CPU(int, resched_irq);
@@ -419,170 +416,6 @@ static irqreturn_t xen_call_function_single_interrupt(int irq, void *dev_id)
         return IRQ_HANDLED;
 }
 
-struct xen_spinlock {
-        unsigned char lock;             /* 0 -> free; 1 -> locked */
-        unsigned short spinners;        /* count of waiting cpus */
-};
-
-static int xen_spin_is_locked(struct raw_spinlock *lock)
-{
-        struct xen_spinlock *xl = (struct xen_spinlock *)lock;
-
-        return xl->lock != 0;
-}
-
-static int xen_spin_is_contended(struct raw_spinlock *lock)
-{
-        struct xen_spinlock *xl = (struct xen_spinlock *)lock;
-
-        /* Not strictly true; this is only the count of contended
-           lock-takers entering the slow path. */
-        return xl->spinners != 0;
-}
-
-static int xen_spin_trylock(struct raw_spinlock *lock)
-{
-        struct xen_spinlock *xl = (struct xen_spinlock *)lock;
-        u8 old = 1;
-
-        asm("xchgb %b0,%1"
-            : "+q" (old), "+m" (xl->lock) : : "memory");
-
-        return old == 0;
-}
-
-static DEFINE_PER_CPU(int, lock_kicker_irq) = -1;
-static DEFINE_PER_CPU(struct xen_spinlock *, lock_spinners);
-
-static inline void spinning_lock(struct xen_spinlock *xl)
-{
-        __get_cpu_var(lock_spinners) = xl;
-        wmb();                  /* set lock of interest before count */
-        asm(LOCK_PREFIX " incw %0"
-            : "+m" (xl->spinners) : : "memory");
-}
-
-static inline void unspinning_lock(struct xen_spinlock *xl)
-{
-        asm(LOCK_PREFIX " decw %0"
-            : "+m" (xl->spinners) : : "memory");
-        wmb();                  /* decrement count before clearing lock */
-        __get_cpu_var(lock_spinners) = NULL;
-}
-
-static noinline int xen_spin_lock_slow(struct raw_spinlock *lock)
-{
-        struct xen_spinlock *xl = (struct xen_spinlock *)lock;
-        int irq = __get_cpu_var(lock_kicker_irq);
-        int ret;
-
-        /* If kicker interrupts not initialized yet, just spin */
-        if (irq == -1)
-                return 0;
-
-        /* announce we're spinning */
-        spinning_lock(xl);
-
-        /* clear pending */
-        xen_clear_irq_pending(irq);
-
-        /* check again make sure it didn't become free while
-           we weren't looking  */
-        ret = xen_spin_trylock(lock);
-        if (ret)
-                goto out;
-
-        /* block until irq becomes pending */
-        xen_poll_irq(irq);
-        kstat_this_cpu.irqs[irq]++;
-
-out:
-        unspinning_lock(xl);
-        return ret;
-}
-
-static void xen_spin_lock(struct raw_spinlock *lock)
-{
-        struct xen_spinlock *xl = (struct xen_spinlock *)lock;
-        int timeout;
-        u8 oldval;
-
-        do {
-                timeout = 1 << 10;
-
-                asm("1: xchgb %1,%0\n"
-                    "   testb %1,%1\n"
-                    "   jz 3f\n"
-                    "2: rep;nop\n"
-                    "   cmpb $0,%0\n"
-                    "   je 1b\n"
-                    "   dec %2\n"
-                    "   jnz 2b\n"
-                    "3:\n"
-                    : "+m" (xl->lock), "=q" (oldval), "+r" (timeout)
-                    : "1" (1)
-                    : "memory");
-
-        } while (unlikely(oldval != 0 && !xen_spin_lock_slow(lock)));
-}
-
-static noinline void xen_spin_unlock_slow(struct xen_spinlock *xl)
-{
-        int cpu;
-
-        for_each_online_cpu(cpu) {
-                /* XXX should mix up next cpu selection */
-                if (per_cpu(lock_spinners, cpu) == xl) {
-                        xen_send_IPI_one(cpu, XEN_SPIN_UNLOCK_VECTOR);
-                        break;
-                }
-        }
-}
-
-static void xen_spin_unlock(struct raw_spinlock *lock)
-{
-        struct xen_spinlock *xl = (struct xen_spinlock *)lock;
-
-        smp_wmb();              /* make sure no writes get moved after unlock */
-        xl->lock = 0;           /* release lock */
-
-        /* make sure unlock happens before kick */
-        barrier();
-
-        if (unlikely(xl->spinners))
-                xen_spin_unlock_slow(xl);
-}
-
-static __cpuinit void xen_init_lock_cpu(int cpu)
-{
-        int irq;
-        const char *name;
-
-        name = kasprintf(GFP_KERNEL, "spinlock%d", cpu);
-        irq = bind_ipi_to_irqhandler(XEN_SPIN_UNLOCK_VECTOR,
-                                     cpu,
-                                     xen_reschedule_interrupt,
-                                     IRQF_DISABLED|IRQF_PERCPU|IRQF_NOBALANCING,
-                                     name,
-                                     NULL);
-
-        if (irq >= 0) {
-                disable_irq(irq); /* make sure it's never delivered */
-                per_cpu(lock_kicker_irq, cpu) = irq;
-        }
-
-        printk("cpu %d spinlock event irq %d\n", cpu, irq);
-}
-
-static void __init xen_init_spinlocks(void)
-{
-        pv_lock_ops.spin_is_locked = xen_spin_is_locked;
-        pv_lock_ops.spin_is_contended = xen_spin_is_contended;
-        pv_lock_ops.spin_lock = xen_spin_lock;
-        pv_lock_ops.spin_trylock = xen_spin_trylock;
-        pv_lock_ops.spin_unlock = xen_spin_unlock;
-}
-
 static const struct smp_ops xen_smp_ops __initdata = {
         .smp_prepare_boot_cpu = xen_smp_prepare_boot_cpu,
         .smp_prepare_cpus = xen_smp_prepare_cpus,
diff --git a/arch/x86/xen/spinlock.c b/arch/x86/xen/spinlock.c
new file mode 100644
index 000000000000..8dc4d31da67f
--- /dev/null
+++ b/arch/x86/xen/spinlock.c
@@ -0,0 +1,183 @@
+/*
+ * Split spinlock implementation out into its own file, so it can be
+ * compiled in a FTRACE-compatible way.
+ */
+#include <linux/kernel_stat.h>
+#include <linux/spinlock.h>
+
+#include <asm/paravirt.h>
+
+#include <xen/interface/xen.h>
+#include <xen/events.h>
+
+#include "xen-ops.h"
+
+struct xen_spinlock {
+        unsigned char lock;             /* 0 -> free; 1 -> locked */
+        unsigned short spinners;        /* count of waiting cpus */
+};
+
+static int xen_spin_is_locked(struct raw_spinlock *lock)
+{
+        struct xen_spinlock *xl = (struct xen_spinlock *)lock;
+
+        return xl->lock != 0;
+}
+
+static int xen_spin_is_contended(struct raw_spinlock *lock)
+{
+        struct xen_spinlock *xl = (struct xen_spinlock *)lock;
+
+        /* Not strictly true; this is only the count of contended
+           lock-takers entering the slow path. */
+        return xl->spinners != 0;
+}
+
+static int xen_spin_trylock(struct raw_spinlock *lock)
+{
+        struct xen_spinlock *xl = (struct xen_spinlock *)lock;
+        u8 old = 1;
+
+        asm("xchgb %b0,%1"
+            : "+q" (old), "+m" (xl->lock) : : "memory");
+
+        return old == 0;
+}
+
+static DEFINE_PER_CPU(int, lock_kicker_irq) = -1;
+static DEFINE_PER_CPU(struct xen_spinlock *, lock_spinners);
+
+static inline void spinning_lock(struct xen_spinlock *xl)
+{
+        __get_cpu_var(lock_spinners) = xl;
+        wmb();                  /* set lock of interest before count */
+        asm(LOCK_PREFIX " incw %0"
+            : "+m" (xl->spinners) : : "memory");
+}
+
+static inline void unspinning_lock(struct xen_spinlock *xl)
+{
+        asm(LOCK_PREFIX " decw %0"
+            : "+m" (xl->spinners) : : "memory");
+        wmb();                  /* decrement count before clearing lock */
+        __get_cpu_var(lock_spinners) = NULL;
+}
+
+static noinline int xen_spin_lock_slow(struct raw_spinlock *lock)
+{
+        struct xen_spinlock *xl = (struct xen_spinlock *)lock;
+        int irq = __get_cpu_var(lock_kicker_irq);
+        int ret;
+
+        /* If kicker interrupts not initialized yet, just spin */
+        if (irq == -1)
+                return 0;
+
+        /* announce we're spinning */
+        spinning_lock(xl);
+
+        /* clear pending */
+        xen_clear_irq_pending(irq);
+
+        /* check again make sure it didn't become free while
+           we weren't looking  */
+        ret = xen_spin_trylock(lock);
+        if (ret)
+                goto out;
+
+        /* block until irq becomes pending */
+        xen_poll_irq(irq);
+        kstat_this_cpu.irqs[irq]++;
+
+out:
+        unspinning_lock(xl);
+        return ret;
+}
+
+static void xen_spin_lock(struct raw_spinlock *lock)
+{
+        struct xen_spinlock *xl = (struct xen_spinlock *)lock;
+        int timeout;
+        u8 oldval;
+
+        do {
+                timeout = 1 << 10;
+
+                asm("1: xchgb %1,%0\n"
+                    "   testb %1,%1\n"
+                    "   jz 3f\n"
+                    "2: rep;nop\n"
+                    "   cmpb $0,%0\n"
+                    "   je 1b\n"
+                    "   dec %2\n"
+                    "   jnz 2b\n"
+                    "3:\n"
+                    : "+m" (xl->lock), "=q" (oldval), "+r" (timeout)
+                    : "1" (1)
+                    : "memory");
+
+        } while (unlikely(oldval != 0 && !xen_spin_lock_slow(lock)));
+}
+
+static noinline void xen_spin_unlock_slow(struct xen_spinlock *xl)
+{
+        int cpu;
+
+        for_each_online_cpu(cpu) {
+                /* XXX should mix up next cpu selection */
+                if (per_cpu(lock_spinners, cpu) == xl) {
+                        xen_send_IPI_one(cpu, XEN_SPIN_UNLOCK_VECTOR);
+                        break;
+                }
+        }
+}
+
+static void xen_spin_unlock(struct raw_spinlock *lock)
+{
+        struct xen_spinlock *xl = (struct xen_spinlock *)lock;
+
+        smp_wmb();              /* make sure no writes get moved after unlock */
+        xl->lock = 0;           /* release lock */
+
+        /* make sure unlock happens before kick */
+        barrier();
+
+        if (unlikely(xl->spinners))
+                xen_spin_unlock_slow(xl);
+}
+
+static irqreturn_t dummy_handler(int irq, void *dev_id)
+{
+        BUG();
+        return IRQ_HANDLED;
+}
+
+void __cpuinit xen_init_lock_cpu(int cpu)
+{
+        int irq;
+        const char *name;
+
+        name = kasprintf(GFP_KERNEL, "spinlock%d", cpu);
+        irq = bind_ipi_to_irqhandler(XEN_SPIN_UNLOCK_VECTOR,
+                                     cpu,
+                                     dummy_handler,
+                                     IRQF_DISABLED|IRQF_PERCPU|IRQF_NOBALANCING,
+                                     name,
+                                     NULL);
+
+        if (irq >= 0) {
+                disable_irq(irq); /* make sure it's never delivered */
+                per_cpu(lock_kicker_irq, cpu) = irq;
+        }
+
+        printk("cpu %d spinlock event irq %d\n", cpu, irq);
+}
+
+void __init xen_init_spinlocks(void)
+{
+        pv_lock_ops.spin_is_locked = xen_spin_is_locked;
+        pv_lock_ops.spin_is_contended = xen_spin_is_contended;
+        pv_lock_ops.spin_lock = xen_spin_lock;
+        pv_lock_ops.spin_trylock = xen_spin_trylock;
+        pv_lock_ops.spin_unlock = xen_spin_unlock;
+}
diff --git a/arch/x86/xen/xen-ops.h b/arch/x86/xen/xen-ops.h
index dd3c23152a2e..8847fb34f17e 100644
--- a/arch/x86/xen/xen-ops.h
+++ b/arch/x86/xen/xen-ops.h
@@ -50,6 +50,9 @@ void __init xen_setup_vcpu_info_placement(void);
 #ifdef CONFIG_SMP
 void xen_smp_init(void);
 
+void __init xen_init_spinlocks(void);
+__cpuinit void xen_init_lock_cpu(int cpu);
+
 extern cpumask_t xen_cpu_initialized_map;
 #else
 static inline void xen_smp_init(void) {}