summaryrefslogtreecommitdiffstats
path: root/kernel
diff options
context:
space:
mode:
Diffstat (limited to 'kernel')
-rw-r--r--kernel/events/Makefile2
-rw-r--r--kernel/events/core.c931
-rw-r--r--kernel/events/hw_breakpoint.c10
-rw-r--r--kernel/events/internal.h96
-rw-r--r--kernel/events/ring_buffer.c380
-rw-r--r--kernel/exit.c31
-rw-r--r--kernel/gcov/Kconfig3
-rw-r--r--kernel/kmod.c16
-rw-r--r--kernel/power/user.c4
-rw-r--r--kernel/rcutree.c398
-rw-r--r--kernel/rcutree.h12
-rw-r--r--kernel/rcutree_plugin.h419
-rw-r--r--kernel/rcutree_trace.c32
-rw-r--r--kernel/sched.c2
-rw-r--r--kernel/sched_rt.c6
-rw-r--r--kernel/signal.c2
-rw-r--r--kernel/smp.c5
-rw-r--r--kernel/softirq.c2
-rw-r--r--kernel/taskstats.c15
-rw-r--r--kernel/time/alarmtimer.c158
-rw-r--r--kernel/time/clocksource.c24
-rw-r--r--kernel/trace/trace_printk.c5
-rw-r--r--kernel/watchdog.c10
23 files changed, 1371 insertions, 1192 deletions
diff --git a/kernel/events/Makefile b/kernel/events/Makefile
index 1ce23d3d839..89e5e8aa4c3 100644
--- a/kernel/events/Makefile
+++ b/kernel/events/Makefile
@@ -2,5 +2,5 @@ ifdef CONFIG_FUNCTION_TRACER
CFLAGS_REMOVE_core.o = -pg
endif
-obj-y := core.o
+obj-y := core.o ring_buffer.o
obj-$(CONFIG_HAVE_HW_BREAKPOINT) += hw_breakpoint.o
diff --git a/kernel/events/core.c b/kernel/events/core.c
index 9efe7108cca..0567e32d71a 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -36,6 +36,8 @@
#include <linux/ftrace_event.h>
#include <linux/hw_breakpoint.h>
+#include "internal.h"
+
#include <asm/irq_regs.h>
struct remote_function_call {
@@ -200,6 +202,22 @@ __get_cpu_context(struct perf_event_context *ctx)
return this_cpu_ptr(ctx->pmu->pmu_cpu_context);
}
+static void perf_ctx_lock(struct perf_cpu_context *cpuctx,
+ struct perf_event_context *ctx)
+{
+ raw_spin_lock(&cpuctx->ctx.lock);
+ if (ctx)
+ raw_spin_lock(&ctx->lock);
+}
+
+static void perf_ctx_unlock(struct perf_cpu_context *cpuctx,
+ struct perf_event_context *ctx)
+{
+ if (ctx)
+ raw_spin_unlock(&ctx->lock);
+ raw_spin_unlock(&cpuctx->ctx.lock);
+}
+
#ifdef CONFIG_CGROUP_PERF
/*
@@ -340,11 +358,8 @@ void perf_cgroup_switch(struct task_struct *task, int mode)
rcu_read_lock();
list_for_each_entry_rcu(pmu, &pmus, entry) {
-
cpuctx = this_cpu_ptr(pmu->pmu_cpu_context);
- perf_pmu_disable(cpuctx->ctx.pmu);
-
/*
* perf_cgroup_events says at least one
* context on this CPU has cgroup events.
@@ -353,6 +368,8 @@ void perf_cgroup_switch(struct task_struct *task, int mode)
* events for a context.
*/
if (cpuctx->ctx.nr_cgroups > 0) {
+ perf_ctx_lock(cpuctx, cpuctx->task_ctx);
+ perf_pmu_disable(cpuctx->ctx.pmu);
if (mode & PERF_CGROUP_SWOUT) {
cpu_ctx_sched_out(cpuctx, EVENT_ALL);
@@ -372,9 +389,9 @@ void perf_cgroup_switch(struct task_struct *task, int mode)
cpuctx->cgrp = perf_cgroup_from_task(task);
cpu_ctx_sched_in(cpuctx, EVENT_ALL, task);
}
+ perf_pmu_enable(cpuctx->ctx.pmu);
+ perf_ctx_unlock(cpuctx, cpuctx->task_ctx);
}
-
- perf_pmu_enable(cpuctx->ctx.pmu);
}
rcu_read_unlock();
@@ -731,6 +748,7 @@ static u64 perf_event_time(struct perf_event *event)
/*
* Update the total_time_enabled and total_time_running fields for a event.
+ * The caller of this function needs to hold the ctx->lock.
*/
static void update_event_times(struct perf_event *event)
{
@@ -1105,6 +1123,10 @@ static int __perf_remove_from_context(void *info)
raw_spin_lock(&ctx->lock);
event_sched_out(event, cpuctx, ctx);
list_del_event(event, ctx);
+ if (!ctx->nr_events && cpuctx->task_ctx == ctx) {
+ ctx->is_active = 0;
+ cpuctx->task_ctx = NULL;
+ }
raw_spin_unlock(&ctx->lock);
return 0;
@@ -1454,8 +1476,24 @@ static void add_event_to_ctx(struct perf_event *event,
event->tstamp_stopped = tstamp;
}
-static void perf_event_context_sched_in(struct perf_event_context *ctx,
- struct task_struct *tsk);
+static void task_ctx_sched_out(struct perf_event_context *ctx);
+static void
+ctx_sched_in(struct perf_event_context *ctx,
+ struct perf_cpu_context *cpuctx,
+ enum event_type_t event_type,
+ struct task_struct *task);
+
+static void perf_event_sched_in(struct perf_cpu_context *cpuctx,
+ struct perf_event_context *ctx,
+ struct task_struct *task)
+{
+ cpu_ctx_sched_in(cpuctx, EVENT_PINNED, task);
+ if (ctx)
+ ctx_sched_in(ctx, cpuctx, EVENT_PINNED, task);
+ cpu_ctx_sched_in(cpuctx, EVENT_FLEXIBLE, task);
+ if (ctx)
+ ctx_sched_in(ctx, cpuctx, EVENT_FLEXIBLE, task);
+}
/*
* Cross CPU call to install and enable a performance event
@@ -1466,20 +1504,37 @@ static int __perf_install_in_context(void *info)
{
struct perf_event *event = info;
struct perf_event_context *ctx = event->ctx;
- struct perf_event *leader = event->group_leader;
struct perf_cpu_context *cpuctx = __get_cpu_context(ctx);
- int err;
+ struct perf_event_context *task_ctx = cpuctx->task_ctx;
+ struct task_struct *task = current;
+
+ perf_ctx_lock(cpuctx, task_ctx);
+ perf_pmu_disable(cpuctx->ctx.pmu);
/*
- * In case we're installing a new context to an already running task,
- * could also happen before perf_event_task_sched_in() on architectures
- * which do context switches with IRQs enabled.
+ * If there was an active task_ctx schedule it out.
*/
- if (ctx->task && !cpuctx->task_ctx)
- perf_event_context_sched_in(ctx, ctx->task);
+ if (task_ctx)
+ task_ctx_sched_out(task_ctx);
+
+ /*
+ * If the context we're installing events in is not the
+ * active task_ctx, flip them.
+ */
+ if (ctx->task && task_ctx != ctx) {
+ if (task_ctx)
+ raw_spin_unlock(&task_ctx->lock);
+ raw_spin_lock(&ctx->lock);
+ task_ctx = ctx;
+ }
+
+ if (task_ctx) {
+ cpuctx->task_ctx = task_ctx;
+ task = task_ctx->task;
+ }
+
+ cpu_ctx_sched_out(cpuctx, EVENT_ALL);
- raw_spin_lock(&ctx->lock);
- ctx->is_active = 1;
update_context_time(ctx);
/*
* update cgrp time only if current cgrp
@@ -1490,43 +1545,13 @@ static int __perf_install_in_context(void *info)
add_event_to_ctx(event, ctx);
- if (!event_filter_match(event))
- goto unlock;
-
/*
- * Don't put the event on if it is disabled or if
- * it is in a group and the group isn't on.
+ * Schedule everything back in
*/
- if (event->state != PERF_EVENT_STATE_INACTIVE ||
- (leader != event && leader->state != PERF_EVENT_STATE_ACTIVE))
- goto unlock;
+ perf_event_sched_in(cpuctx, task_ctx, task);
- /*
- * An exclusive event can't go on if there are already active
- * hardware events, and no hardware event can go on if there
- * is already an exclusive event on.
- */
- if (!group_can_go_on(event, cpuctx, 1))
- err = -EEXIST;
- else
- err = event_sched_in(event, cpuctx, ctx);
-
- if (err) {
- /*
- * This event couldn't go on. If it is in a group
- * then we have to pull the whole group off.
- * If the event group is pinned then put it in error state.
- */
- if (leader != event)
- group_sched_out(leader, cpuctx, ctx);
- if (leader->attr.pinned) {
- update_group_times(leader);
- leader->state = PERF_EVENT_STATE_ERROR;
- }
- }
-
-unlock:
- raw_spin_unlock(&ctx->lock);
+ perf_pmu_enable(cpuctx->ctx.pmu);
+ perf_ctx_unlock(cpuctx, task_ctx);
return 0;
}
@@ -1739,7 +1764,7 @@ out:
raw_spin_unlock_irq(&ctx->lock);
}
-static int perf_event_refresh(struct perf_event *event, int refresh)
+int perf_event_refresh(struct perf_event *event, int refresh)
{
/*
* not supported on inherited events
@@ -1752,36 +1777,35 @@ static int perf_event_refresh(struct perf_event *event, int refresh)
return 0;
}
+EXPORT_SYMBOL_GPL(perf_event_refresh);
static void ctx_sched_out(struct perf_event_context *ctx,
struct perf_cpu_context *cpuctx,
enum event_type_t event_type)
{
struct perf_event *event;
+ int is_active = ctx->is_active;
- raw_spin_lock(&ctx->lock);
- perf_pmu_disable(ctx->pmu);
- ctx->is_active = 0;
+ ctx->is_active &= ~event_type;
if (likely(!ctx->nr_events))
- goto out;
+ return;
+
update_context_time(ctx);
update_cgrp_time_from_cpuctx(cpuctx);
-
if (!ctx->nr_active)
- goto out;
+ return;
- if (event_type & EVENT_PINNED) {
+ perf_pmu_disable(ctx->pmu);
+ if ((is_active & EVENT_PINNED) && (event_type & EVENT_PINNED)) {
list_for_each_entry(event, &ctx->pinned_groups, group_entry)
group_sched_out(event, cpuctx, ctx);
}
- if (event_type & EVENT_FLEXIBLE) {
+ if ((is_active & EVENT_FLEXIBLE) && (event_type & EVENT_FLEXIBLE)) {
list_for_each_entry(event, &ctx->flexible_groups, group_entry)
group_sched_out(event, cpuctx, ctx);
}
-out:
perf_pmu_enable(ctx->pmu);
- raw_spin_unlock(&ctx->lock);
}
/*
@@ -1929,8 +1953,10 @@ static void perf_event_context_sched_out(struct task_struct *task, int ctxn,
rcu_read_unlock();
if (do_switch) {
+ raw_spin_lock(&ctx->lock);
ctx_sched_out(ctx, cpuctx, EVENT_ALL);
cpuctx->task_ctx = NULL;
+ raw_spin_unlock(&ctx->lock);
}
}
@@ -1965,8 +1991,7 @@ void __perf_event_task_sched_out(struct task_struct *task,
perf_cgroup_sched_out(task);
}
-static void task_ctx_sched_out(struct perf_event_context *ctx,
- enum event_type_t event_type)
+static void task_ctx_sched_out(struct perf_event_context *ctx)
{
struct perf_cpu_context *cpuctx = __get_cpu_context(ctx);
@@ -1976,7 +2001,7 @@ static void task_ctx_sched_out(struct perf_event_context *ctx,
if (WARN_ON_ONCE(ctx != cpuctx->task_ctx))
return;
- ctx_sched_out(ctx, cpuctx, event_type);
+ ctx_sched_out(ctx, cpuctx, EVENT_ALL);
cpuctx->task_ctx = NULL;
}
@@ -2055,11 +2080,11 @@ ctx_sched_in(struct perf_event_context *ctx,
struct task_struct *task)
{
u64 now;
+ int is_active = ctx->is_active;
- raw_spin_lock(&ctx->lock);
- ctx->is_active = 1;
+ ctx->is_active |= event_type;
if (likely(!ctx->nr_events))
- goto out;
+ return;
now = perf_clock();
ctx->timestamp = now;
@@ -2068,15 +2093,12 @@ ctx_sched_in(struct perf_event_context *ctx,
* First go through the list and put on any pinned groups
* in order to give them the best chance of going on.
*/
- if (event_type & EVENT_PINNED)
+ if (!(is_active & EVENT_PINNED) && (event_type & EVENT_PINNED))
ctx_pinned_sched_in(ctx, cpuctx);
/* Then walk through the lower prio flexible groups */
- if (event_type & EVENT_FLEXIBLE)
+ if (!(is_active & EVENT_FLEXIBLE) && (event_type & EVENT_FLEXIBLE))
ctx_flexible_sched_in(ctx, cpuctx);
-
-out:
- raw_spin_unlock(&ctx->lock);
}
static void cpu_ctx_sched_in(struct perf_cpu_context *cpuctx,
@@ -2088,19 +2110,6 @@ static void cpu_ctx_sched_in(struct perf_cpu_context *cpuctx,
ctx_sched_in(ctx, cpuctx, event_type, task);
}
-static void task_ctx_sched_in(struct perf_event_context *ctx,
- enum event_type_t event_type)
-{
- struct perf_cpu_context *cpuctx;
-
- cpuctx = __get_cpu_context(ctx);
- if (cpuctx->task_ctx == ctx)
- return;
-
- ctx_sched_in(ctx, cpuctx, event_type, NULL);
- cpuctx->task_ctx = ctx;
-}
-
static void perf_event_context_sched_in(struct perf_event_context *ctx,
struct task_struct *task)
{
@@ -2110,6 +2119,7 @@ static void perf_event_context_sched_in(struct perf_event_context *ctx,
if (cpuctx->task_ctx == ctx)
return;
+ perf_ctx_lock(cpuctx, ctx);
perf_pmu_disable(ctx->pmu);
/*
* We want to keep the following priority order:
@@ -2118,18 +2128,18 @@ static void perf_event_context_sched_in(struct perf_event_context *ctx,
*/
cpu_ctx_sched_out(cpuctx, EVENT_FLEXIBLE);
- ctx_sched_in(ctx, cpuctx, EVENT_PINNED, task);
- cpu_ctx_sched_in(cpuctx, EVENT_FLEXIBLE, task);
- ctx_sched_in(ctx, cpuctx, EVENT_FLEXIBLE, task);
+ perf_event_sched_in(cpuctx, ctx, task);
cpuctx->task_ctx = ctx;
+ perf_pmu_enable(ctx->pmu);
+ perf_ctx_unlock(cpuctx, ctx);
+
/*
* Since these rotations are per-cpu, we need to ensure the
* cpu-context we got scheduled on is actually rotating.
*/
perf_pmu_rotate_start(ctx->pmu);
- perf_pmu_enable(ctx->pmu);
}
/*
@@ -2269,7 +2279,6 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx, u64 period)
u64 interrupts, now;
s64 delta;
- raw_spin_lock(&ctx->lock);
list_for_each_entry_rcu(event, &ctx->event_list, event_entry) {
if (event->state != PERF_EVENT_STATE_ACTIVE)
continue;
@@ -2301,7 +2310,6 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx, u64 period)
if (delta > 0)
perf_adjust_period(event, period, delta);
}
- raw_spin_unlock(&ctx->lock);
}
/*
@@ -2309,16 +2317,12 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx, u64 period)
*/
static void rotate_ctx(struct perf_event_context *ctx)
{
- raw_spin_lock(&ctx->lock);
-
/*
* Rotate the first entry last of non-pinned groups. Rotation might be
* disabled by the inheritance code.
*/
if (!ctx->rotate_disable)
list_rotate_left(&ctx->flexible_groups);
-
- raw_spin_unlock(&ctx->lock);
}
/*
@@ -2345,6 +2349,7 @@ static void perf_rotate_context(struct perf_cpu_context *cpuctx)
rotate = 1;
}
+ perf_ctx_lock(cpuctx, cpuctx->task_ctx);
perf_pmu_disable(cpuctx->ctx.pmu);
perf_ctx_adjust_freq(&cpuctx->ctx, interval);
if (ctx)
@@ -2355,21 +2360,20 @@ static void perf_rotate_context(struct perf_cpu_context *cpuctx)
cpu_ctx_sched_out(cpuctx, EVENT_FLEXIBLE);
if (ctx)
- task_ctx_sched_out(ctx, EVENT_FLEXIBLE);
+ ctx_sched_out(ctx, cpuctx, EVENT_FLEXIBLE);
rotate_ctx(&cpuctx->ctx);
if (ctx)
rotate_ctx(ctx);
- cpu_ctx_sched_in(cpuctx, EVENT_FLEXIBLE, current);
- if (ctx)
- task_ctx_sched_in(ctx, EVENT_FLEXIBLE);
+ perf_event_sched_in(cpuctx, ctx, current);
done:
if (remove)
list_del_init(&cpuctx->rotation_list);
perf_pmu_enable(cpuctx->ctx.pmu);
+ perf_ctx_unlock(cpuctx, cpuctx->task_ctx);
}
void perf_event_task_tick(void)
@@ -2424,9 +2428,9 @@ static void perf_event_enable_on_exec(struct perf_event_context *ctx)
* in.
*/
perf_cgroup_sched_out(current);
- task_ctx_sched_out(ctx, EVENT_ALL);
raw_spin_lock(&ctx->lock);
+ task_ctx_sched_out(ctx);
list_for_each_entry(event, &ctx->pinned_groups, group_entry) {
ret = event_enable_on_exec(event, ctx);
@@ -2835,16 +2839,12 @@ retry:
unclone_ctx(ctx);
++ctx->pin_count;
raw_spin_unlock_irqrestore(&ctx->lock, flags);
- }
-
- if (!ctx) {
+ } else {
ctx = alloc_perf_context(pmu, task);
err = -ENOMEM;
if (!ctx)
goto errout;
- get_ctx(ctx);
-
err = 0;
mutex_lock(&task->perf_event_mutex);
/*
@@ -2856,14 +2856,14 @@ retry:
else if (task->perf_event_ctxp[ctxn])
err = -EAGAIN;
else {
+ get_ctx(ctx);
++ctx->pin_count;
rcu_assign_pointer(task->perf_event_ctxp[ctxn], ctx);
}
mutex_unlock(&task->perf_event_mutex);
if (unlikely(err)) {
- put_task_struct(task);
- kfree(ctx);
+ put_ctx(ctx);
if (err == -EAGAIN)
goto retry;
@@ -2890,7 +2890,7 @@ static void free_event_rcu(struct rcu_head *head)
kfree(event);
}
-static void perf_buffer_put(struct perf_buffer *buffer);
+static void ring_buffer_put(struct ring_buffer *rb);
static void free_event(struct perf_event *event)
{
@@ -2913,9 +2913,9 @@ static void free_event(struct perf_event *event)
}
}
- if (event->buffer) {
- perf_buffer_put(event->buffer);
- event->buffer = NULL;
+ if (event->rb) {
+ ring_buffer_put(event->rb);
+ event->rb = NULL;
}
if (is_cgroup_event(event))
@@ -2934,12 +2934,6 @@ int perf_event_release_kernel(struct perf_event *event)
{
struct perf_event_context *ctx = event->ctx;
- /*
- * Remove from the PMU, can't get re-enabled since we got
- * here because the last ref went.
- */
- perf_event_disable(event);
-
WARN_ON_ONCE(ctx->parent_ctx);
/*
* There are two ways this annotation is useful:
@@ -2956,8 +2950,8 @@ int perf_event_release_kernel(struct perf_event *event)
mutex_lock_nested(&ctx->mutex, SINGLE_DEPTH_NESTING);
raw_spin_lock_irq(&ctx->lock);
perf_group_detach(event);
- list_del_event(event, ctx);
raw_spin_unlock_irq(&ctx->lock);
+ perf_remove_from_context(event);
mutex_unlock(&ctx->mutex);
free_event(event);
@@ -3149,13 +3143,13 @@ perf_read(struct file *file, char __user *buf, size_t count, loff_t *ppos)
static unsigned int perf_poll(struct file *file, poll_table *wait)
{
struct perf_event *event = file->private_data;
- struct perf_buffer *buffer;
+ struct ring_buffer *rb;
unsigned int events = POLL_HUP;
rcu_read_lock();
- buffer = rcu_dereference(event->buffer);
- if (buffer)
- events = atomic_xchg(&buffer->poll, 0);
+ rb = rcu_dereference(event->rb);
+ if (rb)
+ events = atomic_xchg(&rb->poll, 0);
rcu_read_unlock();
poll_wait(file, &event->waitq, wait);
@@ -3358,6 +3352,18 @@ static int perf_event_index(struct perf_event *event)
return event->hw.idx + 1 - PERF_EVENT_INDEX_OFFSET;
}
+static void calc_timer_values(struct perf_event *event,
+ u64 *running,
+ u64 *enabled)
+{
+ u64 now, ctx_time;
+
+ now = perf_clock();
+ ctx_time = event->shadow_ctx_time + now;
+ *enabled = ctx_time - event->tstamp_enabled;
+ *running = ctx_time - event->tstamp_running;
+}
+
/*
* Callers need to ensure there can be no nesting of this function, otherwise
* the seqlock logic goes bad. We can not serialize this because the arch
@@ -3366,14 +3372,25 @@ static int perf_event_index(struct perf_event *event)
void perf_event_update_userpage(struct perf_event *event)
{
struct perf_event_mmap_page *userpg;
- struct perf_buffer *buffer;
+ struct ring_buffer *rb;
+ u64 enabled, running;
rcu_read_lock();
- buffer = rcu_dereference(event->buffer);
- if (!buffer)
+ /*
+ * compute total_time_enabled, total_time_running
+ * based on snapshot values taken when the event
+ * was last scheduled in.
+ *
+ * we cannot simply called update_context_time()
+ * because of locking issue as we can be called in
+ * NMI context
+ */
+ calc_timer_values(event, &enabled, &running);
+ rb = rcu_dereference(event->rb);
+ if (!rb)
goto unlock;
- userpg = buffer->user_page;
+ userpg = rb->user_page;
/*
* Disable preemption so as to not let the corresponding user-space
@@ -3387,10 +3404,10 @@ void perf_event_update_userpage(struct perf_event *event)
if (event->state == PERF_EVENT_STATE_ACTIVE)
userpg->offset -= local64_read(&event->hw.prev_count);
- userpg->time_enabled = event->total_time_enabled +
+ userpg->time_enabled = enabled +
atomic64_read(&event->child_total_time_enabled);
- userpg->time_running = event->total_time_running +
+ userpg->time_running = running +
atomic64_read(&event->child_total_time_running);
barrier();
@@ -3400,220 +3417,10 @@ unlock:
rcu_read_unlock();
}
-static unsigned long perf_data_size(struct perf_buffer *buffer);
-
-static void
-perf_buffer_init(struct perf_buffer *buffer, long watermark, int flags)
-{
- long max_size = perf_data_size(buffer);
-
- if (watermark)
- buffer->watermark = min(max_size, watermark);
-
- if (!buffer->watermark)
- buffer->watermark = max_size / 2;
-
- if (flags & PERF_BUFFER_WRITABLE)
- buffer->writable = 1;
-
- atomic_set(&buffer->refcount, 1);
-}
-
-#ifndef CONFIG_PERF_USE_VMALLOC
-
-/*
- * Back perf_mmap() with regular GFP_KERNEL-0 pages.
- */
-
-static struct page *
-perf_mmap_to_page(struct perf_buffer *buffer, unsigned long pgoff)
-{
- if (pgoff > buffer->nr_pages)
- return NULL;
-
- if (pgoff == 0)
- return virt_to_page(buffer->user_page);
-
- return virt_to_page(buffer->data_pages[pgoff - 1]);
-}
-
-static void *perf_mmap_alloc_page(int cpu)
-{
- struct page *page;
- int node;
-
- node = (cpu == -1) ? cpu : cpu_to_node(cpu);
- page = alloc_pages_node(node, GFP_KERNEL | __GFP_ZERO, 0);
- if (!page)
- return NULL;
-
- return page_address(page);
-}
-
-static struct perf_buffer *
-perf_buffer_alloc(int nr_pages, long watermark, int cpu, int flags)
-{
- struct perf_buffer *buffer;
- unsigned long size;
- int i;
-
- size = sizeof(struct perf_buffer);
- size += nr_pages * sizeof(void *);
-
- buffer = kzalloc(size, GFP_KERNEL);
- if (!buffer)
- goto fail;
-
- buffer->user_page = perf_mmap_alloc_page(cpu);
- if (!buffer->user_page)
- goto fail_user_page;
-
- for (i = 0; i < nr_pages; i++) {
- buffer->data_pages[i] = perf_mmap_alloc_page(cpu);
- if (!buffer->data_pages[i])
- goto fail_data_pages;
- }
-
- buffer->nr_pages = nr_pages;
-
- perf_buffer_init(buffer, watermark, flags);
-
- return buffer;
-
-fail_data_pages:
- for (i--; i >= 0; i--)
- free_page((unsigned long)buffer->data_pages[i]);
-
- free_page((unsigned long)buffer->user_page);
-
-fail_user_page:
- kfree(buffer);
-
-fail:
- return NULL;
-}
-
-static void perf_mmap_free_page(unsigned long addr)
-{
- struct page *page = virt_to_page((void *)addr);
-
- page->mapping = NULL;
- __free_page(page);
-}
-
-static void perf_buffer_free(struct perf_buffer *buffer)
-{
- int i;
-
- perf_mmap_free_page((unsigned long)buffer->user_page);
- for (i = 0; i < buffer->nr_pages; i++)
- perf_mmap_free_page((unsigned long)buffer->data_pages[i]);
- kfree(buffer);
-}
-
-static inline int page_order(struct perf_buffer *buffer)
-{
- return 0;
-}
-
-#else
-
-/*
- * Back perf_mmap() with vmalloc memory.
- *
- * Required for architectures that have d-cache aliasing issues.
- */
-
-static inline int page_order(struct perf_buffer *buffer)
-{
- return buffer->page_order;
-}
-
-static struct page *
-perf_mmap_to_page(struct perf_buffer *buffer, unsigned long pgoff)
-{
- if (pgoff > (1UL << page_order(buffer)))
- return NULL;
-
- return vmalloc_to_page((void *)buffer->user_page + pgoff * PAGE_SIZE);
-}
-
-static void perf_mmap_unmark_page(void *addr)
-{
- struct page *page = vmalloc_to_page(addr);
-
- page->mapping = NULL;
-}
-
-static void perf_buffer_free_work(struct work_struct *work)
-{
- struct perf_buffer *buffer;
- void *base;
- int i, nr;
-
- buffer = container_of(work, struct perf_buffer, work);
- nr = 1 << page_order(buffer);
-
- base = buffer->user_page;
- for (i = 0; i < nr + 1; i++)
- perf_mmap_unmark_page(base + (i * PAGE_SIZE));
-
- vfree(base);
- kfree(buffer);
-}
-
-static void perf_buffer_free(struct perf_buffer *buffer)
-{
- schedule_work(&buffer->work);
-}
-
-static struct perf_buffer *
-perf_buffer_alloc(int nr_pages, long watermark, int cpu, int flags)
-{
- struct perf_buffer *buffer;
- unsigned long size;
- void *all_buf;
-
- size = sizeof(struct perf_buffer);
- size += sizeof(void *);
-
- buffer = kzalloc(size, GFP_KERNEL);
- if (!buffer)
- goto fail;
-
- INIT_WORK(&buffer->work, perf_buffer_free_work);
-
- all_buf = vmalloc_user((nr_pages + 1) * PAGE_SIZE);
- if (!all_buf)
- goto fail_all_buf;
-
- buffer->user_page = all_buf;
- buffer->data_pages[0] = all_buf + PAGE_SIZE;
- buffer->page_order = ilog2(nr_pages);
- buffer->nr_pages = 1;
-
- perf_buffer_init(buffer, watermark, flags);
-
- return buffer;
-
-fail_all_buf:
- kfree(buffer);
-
-fail:
- return NULL;
-}
-
-#endif
-
-static unsigned long perf_data_size(struct perf_buffer *buffer)
-{
- return buffer->nr_pages << (PAGE_SHIFT + page_order(buffer));
-}
-
static int perf_mmap_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
{
struct perf_event *event = vma->vm_file->private_data;
- struct perf_buffer *buffer;
+ struct ring_buffer *rb;
int ret = VM_FAULT_SIGBUS;
if (vmf->flags & FAULT_FLAG_MKWRITE) {
@@ -3623,14 +3430,14 @@ static int perf_mmap_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
}
rcu_read_lock();
- buffer = rcu_dereference(event->buffer);
- if (!buffer)
+ rb = rcu_dereference(event->rb);
+ if (!rb)
goto unlock;
if (vmf->pgoff && (vmf->flags & FAULT_FLAG_WRITE))
goto unlock;
- vmf->page = perf_mmap_to_page(buffer, vmf->pgoff);
+ vmf->page = perf_mmap_to_page(rb, vmf->pgoff);
if (!vmf->page)
goto unlock;
@@ -3645,35 +3452,35 @@ unlock:
return ret;
}
-static void perf_buffer_free_rcu(struct rcu_head *rcu_head)
+static void rb_free_rcu(struct rcu_head *rcu_head)
{
- struct perf_buffer *buffer;
+ struct ring_buffer *rb;
- buffer = container_of(rcu_head, struct perf_buffer, rcu_head);
- perf_buffer_free(buffer);
+ rb = container_of(rcu_head, struct ring_buffer, rcu_head);
+ rb_free(rb);
}
-static struct perf_buffer *perf_buffer_get(struct perf_event *event)
+static struct ring_buffer *ring_buffer_get(struct perf_event *event)
{
- struct perf_buffer *buffer;
+ struct ring_buffer *rb;
rcu_read_lock();
- buffer = rcu_dereference(event->buffer);
- if (buffer) {
- if (!atomic_inc_not_zero(&buffer->refcount))
- buffer = NULL;
+ rb = rcu_dereference(event->rb);
+ if (rb) {
+ if (!atomic_inc_not_zero(&rb->refcount))
+ rb = NULL;
}
rcu_read_unlock();
- return buffer;
+ return rb;
}
-static void perf_buffer_put(struct perf_buffer *buffer)
+static void ring_buffer_put(struct ring_buffer *rb)
{
- if (!atomic_dec_and_test(&buffer->refcount))
+ if (!atomic_dec_and_test(&rb->refcount))
return;
- call_rcu(&buffer->rcu_head, perf_buffer_free_rcu);
+ call_rcu(&rb->rcu_head, rb_free_rcu);
}
static void perf_mmap_open(struct vm_area_struct *vma)
@@ -3688,16 +3495,16 @@ static void perf_mmap_close(struct vm_area_struct *vma)
struct perf_event *event = vma->vm_file->private_data;
if (atomic_dec_and_mutex_lock(&event->mmap_count, &event->mmap_mutex)) {
- unsigned long size = perf_data_size(event->buffer);
+ unsigned long size = perf_data_size(event->rb);
struct user_struct *user = event->mmap_user;
- struct perf_buffer *buffer = event->buffer;
+ struct ring_buffer *rb = event->rb;
atomic_long_sub((size >> PAGE_SHIFT) + 1, &user->locked_vm);
vma->vm_mm->locked_vm -= event->mmap_locked;
- rcu_assign_pointer(event->buffer, NULL);
+ rcu_assign_pointer(event->rb, NULL);
mutex_unlock(&event->mmap_mutex);
- perf_buffer_put(buffer);
+ ring_buffer_put(rb);
free_uid(user);
}
}
@@ -3715,7 +3522,7 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma)
unsigned long user_locked, user_lock_limit;
struct user_struct *user = current_user();
unsigned long locked, lock_limit;
- struct perf_buffer *buffer;
+ struct ring_buffer *rb;
unsigned long vma_size;
unsigned long nr_pages;
long user_extra, extra;
@@ -3724,7 +3531,7 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma)
/*
* Don't allow mmap() of inherited per-task counters. This would
* create a performance issue due to all children writing to the
- * same buffer.
+ * same rb.
*/
if (event->cpu == -1 && event->attr.inherit)
return -EINVAL;
@@ -3736,7 +3543,7 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma)
nr_pages = (vma_size / PAGE_SIZE) - 1;
/*
- * If we have buffer pages ensure they're a power-of-two number, so we
+ * If we have rb pages ensure they're a power-of-two number, so we
* can do bitmasks instead of modulo.
*/
if (nr_pages != 0 && !is_power_of_2(nr_pages))
@@ -3750,9 +3557,9 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma)
WARN_ON_ONCE(event->ctx->parent_ctx);
mutex_lock(&event->mmap_mutex);
- if (event->buffer) {
- if (event->buffer->nr_pages == nr_pages)
- atomic_inc(&event->buffer->refcount);
+ if (event->rb) {
+ if (event->rb->nr_pages == nr_pages)
+ atomic_inc(&event->rb->refcount);
else
ret = -EINVAL;
goto unlock;
@@ -3782,18 +3589,20 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma)
goto unlock;
}
- WARN_ON(event->buffer);
+ WARN_ON(event->rb);
if (vma->vm_flags & VM_WRITE)
- flags |= PERF_BUFFER_WRITABLE;
+ flags |= RING_BUFFER_WRITABLE;
- buffer = perf_buffer_alloc(nr_pages, event->attr.wakeup_watermark,
- event->cpu, flags);
- if (!buffer) {
+ rb = rb_alloc(nr_pages,
+ event->attr.watermark ? event->attr.wakeup_watermark : 0,
+ event->cpu, flags);
+
+ if (!rb) {
ret = -ENOMEM;
goto unlock;
}
- rcu_assign_pointer(event->buffer, buffer);
+ rcu_assign_pointer(event->rb, rb);
atomic_long_add(user_extra, &user->locked_vm);
event->mmap_locked = extra;
@@ -3892,117 +3701,6 @@ int perf_unregister_guest_info_callbacks(struct perf_guest_info_callbacks *cbs)
}
EXPORT_SYMBOL_GPL(perf_unregister_guest_info_callbacks);
-/*
- * Output
- */
-static bool perf_output_space(struct perf_buffer *buffer, unsigned long tail,
- unsigned long offset, unsigned long head)
-{
- unsigned long mask;
-
- if (!buffer->writable)
- return true;
-
- mask = perf_data_size(buffer) - 1;
-
- offset = (offset - tail) & mask;
- head = (head - tail) & mask;
-
- if ((int)(head - offset) < 0)
- return false;
-
- return true;
-}
-
-static void perf_output_wakeup(struct perf_output_handle *handle)
-{
- atomic_set(&handle->buffer->poll, POLL_IN);
-
- if (handle->nmi) {
- handle->event->pending_wakeup = 1;
- irq_work_queue(&handle->event->pending);
- } else
- perf_event_wakeup(handle->event);
-}
-
-/*
- * We need to ensure a later event_id doesn't publish a head when a former
- * event isn't done writing. However since we need to deal with NMIs we
- * cannot fully serialize things.
- *
- * We only publish the head (and generate a wakeup) when the outer-most
- * event completes.
- */
-static void perf_output_get_handle(struct perf_output_handle *handle)
-{
- struct perf_buffer *buffer = handle->buffer;
-
- preempt_disable();
- local_inc(&buffer->nest);
- handle->wakeup = local_read(&buffer->wakeup);
-}
-
-static void perf_output_put_handle(struct perf_output_handle *handle)
-{
- struct perf_buffer *buffer = handle->buffer;
- unsigned long head;
-
-again:
- head = local_read(&buffer->head);
-
- /*
- * IRQ/NMI can happen here, which means we can miss a head update.
- */
-
- if (!local_dec_and_test(&buffer->nest))
- goto out;
-
- /*
- * Publish the known good head. Rely on the full barrier implied
- * by atomic_dec_and_test() order the buffer->head read and this
- * write.
- */
- buffer->user_page->data_head = head;
-
- /*
- * Now check if we missed an update, rely on the (compiler)
- * barrier in atomic_dec_and_test() to re-read buffer->head.
- */
- if (unlikely(head != local_read(&buffer->head))) {
- local_inc(&buffer->nest);
- goto again;
- }
-
- if (handle->wakeup != local_read(&buffer->wakeup))
- perf_output_wakeup(handle);
-
-out:
- preempt_enable();
-}
-
-__always_inline void perf_output_copy(struct perf_output_handle *handle,
- const void *buf, unsigned int len)
-{
- do {
- unsigned long size = min_t(unsigned long, handle->size, len);
-
- memcpy(handle->addr, buf, size);
-
- len -= size;
- handle->addr += size;
- buf += size;
- handle->size -= size;
- if (!handle->size) {
- struct perf_buffer *buffer = handle->buffer;
-
- handle->page++;
- handle->page &= buffer->nr_pages - 1;
- handle->addr = buffer->data_pages[handle->page];
- handle->size = PAGE_SIZE << page_order(buffer);
- }
- } while (len);
-}
-
static void __perf_event_header__init_id(struct perf_event_header *header,
struct perf_sample_data *data,
struct perf_event *event)
@@ -4033,9 +3731,9 @@ static void __perf_event_header__init_id(struct perf_event_header *header,
}
}
-static void perf_event_header__init_id(struct perf_event_header *header,
- struct perf_sample_data *data,
- struct perf_event *event)
+void perf_event_header__init_id(struct perf_event_header *header,
+ struct perf_sample_data *data,
+ struct perf_event *event)
{
if (event->attr.sample_id_all)
__perf_event_header__init_id(header, data, event);
@@ -4062,121 +3760,14 @@ static void __perf_event__output_id_sample(struct perf_output_handle *handle,
perf_output_put(handle, data->cpu_entry);
}
-static void perf_event__output_id_sample(struct perf_event *event,
- struct perf_output_handle *handle,
- struct perf_sample_data *sample)
+void perf_event__output_id_sample(struct perf_event *event,
+ struct perf_output_handle *handle,
+ struct perf_sample_data *sample)
{
if (event->attr.sample_id_all)
__perf_event__output_id_sample(handle, sample);
}
-int perf_output_begin(struct perf_output_handle *handle,
- struct perf_event *event, unsigned int size,
- int nmi, int sample)
-{
- struct perf_buffer *buffer;
- unsigned long tail, offset, head;
- int have_lost;
- struct perf_sample_data sample_data;
- struct {
- struct perf_event_header header;
- u64 id;
- u64 lost;
- } lost_event;
-
- rcu_read_lock();
- /*
- * For inherited events we send all the output towards the parent.
- */
- if (event->parent)
- event = event->parent;
-
- buffer = rcu_dereference(event->buffer);
- if (!buffer)
- goto out;
-
- handle->buffer = buffer;
- handle->event = event;
- handle->nmi = nmi;
- handle->sample = sample;
-
- if (!buffer->nr_pages)
- goto out;
-
- have_lost = local_read(&buffer->lost);
- if (have_lost) {
- lost_event.header.size = sizeof(lost_event);
- perf_event_header__init_id(&lost_event.header, &sample_data,
- event);
- size += lost_event.header.size;
- }
-
- perf_output_get_handle(handle);
-
- do {
- /*
- * Userspace could choose to issue a mb() before updating the
- * tail pointer. So that all reads will be completed before the
- * write is issued.
- */
- tail = ACCESS_ONCE(buffer->user_page->data_tail);
- smp_rmb();
- offset = head = local_read(&buffer->head);
- head += size;
- if (unlikely(!perf_output_space(buffer, tail, offset, head)))
- goto fail;
- } while (local_cmpxchg(&buffer->head, offset, head) != offset);
-
- if (head - local_read(&buffer->wakeup) > buffer->watermark)
- local_add(buffer->watermark, &buffer->wakeup);
-
- handle->page = offset >> (PAGE_SHIFT + page_order(buffer));
- handle->page &= buffer->nr_pages - 1;
- handle->size = offset & ((PAGE_SIZE << page_order(buffer)) - 1);
- handle->addr = buffer->data_pages[handle->page];
- handle->addr += handle->size;
- handle->size = (PAGE_SIZE << page_order(buffer)) - handle->size;
-
- if (have_lost) {
- lost_event.header.type = PERF_RECORD_LOST;
- lost_event.header.misc = 0;
- lost_event.id = event->id;
- lost_event.lost = local_xchg(&buffer->lost, 0);
-
- perf_output_put(handle, lost_event);
- perf_event__output_id_sample(event, handle, &sample_data);
- }
-
- return 0;
-
-fail:
- local_inc(&buffer->lost);
- perf_output_put_handle(handle);
-out:
- rcu_read_unlock();
-
- return -ENOSPC;
-}
-
-void perf_output_end(struct perf_output_handle *handle)
-{
- struct perf_event *event = handle->event;
- struct perf_buffer *buffer = handle->buffer;
-
- int wakeup_events = event->attr.wakeup_events;
-
- if (handle->sample && wakeup_events) {
- int events = local_inc_return(&buffer->events);
- if (events >= wakeup_events) {
- local_sub(wakeup_events, &buffer->events);
- local_inc(&buffer->wakeup);
- }
- }
-
- perf_output_put_handle(handle);
- rcu_read_unlock();
-}
-
static void perf_output_read_one(struct perf_output_handle *handle,
struct perf_event *event,
u64 enabled, u64 running)
@@ -4197,7 +3788,7 @@ static void perf_output_read_one(struct perf_output_handle *handle,
if (read_format & PERF_FORMAT_ID)
values[n++] = primary_event_id(event);
- perf_output_copy(handle, values, n * sizeof(u64));
+ __output_copy(handle, values, n * sizeof(u64));
}
/*
@@ -4227,7 +3818,7 @@ static void perf_output_read_group(struct perf_output_handle *handle,
if (read_format & PERF_FORMAT_ID)
values[n++] = primary_event_id(leader);
- perf_output_copy(handle, values, n * sizeof(u64));
+ __output_copy(handle, values, n * sizeof(u64));
list_for_each_entry(sub, &leader->sibling_list, group_entry) {
n = 0;
@@ -4239,7 +3830,7 @@ static void perf_output_read_group(struct perf_output_handle *handle,
if (read_format & PERF_FORMAT_ID)
values[n++] = primary_event_id(sub);
- perf_output_copy(handle, values, n * sizeof(u64));
+ __output_copy(handle, values, n * sizeof(u64));
}
}
@@ -4249,7 +3840,7 @@ static void perf_output_read_group(struct perf_output_handle *handle,
static void perf_output_read(struct perf_output_handle *handle,
struct perf_event *event)
{
- u64 enabled = 0, running = 0, now, ctx_time;
+ u64 enabled = 0, running = 0;
u64 read_format = event->attr.read_format;
/*
@@ -4261,12 +3852,8 @@ static void perf_output_read(struct perf_output_handle *handle,
* because of locking issue as we are called in
* NMI context
*/
- if (read_format & PERF_FORMAT_TOTAL_TIMES) {
- now = perf_clock();
- ctx_time = event->shadow_ctx_time + now;
- enabled = ctx_time - event->tstamp_enabled;
- running = ctx_time - event->tstamp_running;
- }
+ if (read_format & PERF_FORMAT_TOTAL_TIMES)
+ calc_timer_values(event, &enabled, &running);
if (event->attr.read_format & PERF_FORMAT_GROUP)
perf_output_read_group(handle, event, enabled, running);
@@ -4319,7 +3906,7 @@ void perf_output_sample(struct perf_output_handle *handle,
size *= sizeof(u64);
- perf_output_copy(handle, data->callchain, size);
+ __output_copy(handle, data->callchain, size);
} else {
u64 nr = 0;
perf_output_put(handle, nr);
@@ -4329,8 +3916,8 @@ void perf_output_sample(struct perf_output_handle *handle,
if (sample_type & PERF_SAMPLE_RAW) {
if (data->raw) {
perf_output_put(handle, data->raw->size);
- perf_output_copy(handle, data->raw->data,
- data->raw->size);
+ __output_copy(handle, data->raw->data,
+ data->raw->size);
} else {
struct {
u32 size;
@@ -4342,6 +3929,20 @@ void perf_output_sample(struct perf_output_handle *handle,
perf_output_put(handle, raw);
}
}
+
+ if (!event->attr.watermark) {
+ int wakeup_events = event->attr.wakeup_events;
+
+ if (wakeup_events) {
+ struct ring_buffer *rb = handle->rb;
+ int events = local_inc_return(&rb->events);
+
+ if (events >= wakeup_events) {
+ local_sub(wakeup_events, &rb->events);
+ local_inc(&rb->wakeup);
+ }
+ }
+ }
}
void perf_prepare_sample(struct perf_event_header *header,
@@ -4386,7 +3987,7 @@ void perf_prepare_sample(struct perf_event_header *header,
}
}
-static void perf_event_output(struct perf_event *event, int nmi,
+static void perf_event_output(struct perf_event *event,
struct perf_sample_data *data,
struct pt_regs *regs)
{
@@ -4398,7 +3999,7 @@ static void perf_event_output(struct perf_event *event, int nmi,
perf_prepare_sample(&header, data, event, regs);
- if (perf_output_begin(&handle, event, header.size, nmi, 1))
+ if (perf_output_begin(&handle, event, header.size))
goto exit;
perf_output_sample(&handle, &header, data, event);
@@ -4438,7 +4039,7 @@ perf_event_read_event(struct perf_event *event,
int ret;
perf_event_header__init_id(&read_event.header, &sample, event);
- ret = perf_output_begin(&handle, event, read_event.header.size, 0, 0);
+ ret = perf_output_begin(&handle, event, read_event.header.size);
if (ret)
return;
@@ -4481,7 +4082,7 @@ static void perf_event_task_output(struct perf_event *event,
perf_event_header__init_id(&task_event->event_id.header, &sample, event);
ret = perf_output_begin(&handle, event,
- task_event->event_id.header.size, 0, 0);
+ task_event->event_id.header.size);
if (ret)
goto out;
@@ -4618,7 +4219,7 @@ static void perf_event_comm_output(struct perf_event *event,
perf_event_header__init_id(&comm_event->event_id.header, &sample, event);
ret = perf_output_begin(&handle, event,
- comm_event->event_id.header.size, 0, 0);
+ comm_event->event_id.header.size);
if (ret)
goto out;
@@ -4627,7 +4228,7 @@ static void perf_event_comm_output(struct perf_event *event,
comm_event->event_id.tid = perf_event_tid(event, comm_event->task);
perf_output_put(&handle, comm_event->event_id);
- perf_output_copy(&handle, comm_event->comm,
+ __output_copy(&handle, comm_event->comm,
comm_event->comm_size);
perf_event__output_id_sample(event, &handle, &sample);
@@ -4765,7 +4366,7 @@ static void perf_event_mmap_output(struct perf_event *event,
perf_event_header__init_id(&mmap_event->event_id.header, &sample, event);
ret = perf_output_begin(&handle, event,
- mmap_event->event_id.header.size, 0, 0);
+ mmap_event->event_id.header.size);
if (ret)
goto out;
@@ -4773,7 +4374,7 @@ static void perf_event_mmap_output(struct perf_event *event,
mmap_event->event_id.tid = perf_event_tid(event, current);
perf_output_put(&handle, mmap_event->event_id);
- perf_output_copy(&handle, mmap_event->file_name,
+ __output_copy(&handle, mmap_event->file_name,
mmap_event->file_size);
perf_event__output_id_sample(event, &handle, &sample);
@@ -4829,7 +4430,7 @@ static void perf_event_mmap_event(struct perf_mmap_event *mmap_event)
if (file) {
/*
- * d_path works from the end of the buffer backwards, so we
+ * d_path works from the end of the rb backwards, so we
* need to add enough zero bytes after the string to handle
* the 64bit alignment we do later.
*/
@@ -4960,7 +4561,7 @@ static void perf_log_throttle(struct perf_event *event, int enable)
perf_event_header__init_id(&throttle_event.header, &sample, event);
ret = perf_output_begin(&handle, event,
- throttle_event.header.size, 1, 0);
+ throttle_event.header.size);
if (ret)
return;
@@ -4973,7 +4574,7 @@ static void perf_log_throttle(struct perf_event *event, int enable)
* Generic event overflow handling, sampling.
*/
-static int __perf_event_overflow(struct perf_event *event, int nmi,
+static int __perf_event_overflow(struct perf_event *event,
int throttle, struct perf_sample_data *data,
struct pt_regs *regs)
{
@@ -5016,34 +4617,28 @@ static int __perf_event_overflow(struct perf_event *event, int nmi,
if (events && atomic_dec_and_test(&event->event_limit)) {
ret = 1;
event->pending_kill = POLL_HUP;
- if (nmi) {
- event->pending_disable = 1;
- irq_work_queue(&event->pending);
- } else
- perf_event_disable(event);
+ event->pending_disable = 1;
+ irq_work_queue(&event->pending);
}
if (event->overflow_handler)
- event->overflow_handler(event, nmi, data, regs);
+ event->overflow_handler(event, data, regs);
else
- perf_event_output(event, nmi, data, regs);
+ perf_event_output(event, data, regs);
if (event->fasync && event->pending_kill) {
- if (nmi) {
- event->pending_wakeup = 1;
- irq_work_queue(&event->pending);
- } else
- perf_event_wakeup(event);
+ event->pending_wakeup = 1;
+ irq_work_queue(&event->pending);
}
return ret;
}
-int perf_event_overflow(struct perf_event *event, int nmi,
+int perf_event_overflow(struct perf_event *event,
struct perf_sample_data *data,
struct pt_regs *regs)
{
- return __perf_event_overflow(event, nmi, 1, data, regs);
+ return __perf_event_overflow(event, 1, data, regs);
}
/*
@@ -5092,7 +4687,7 @@ again:
}
static void perf_swevent_overflow(struct perf_event *event, u64 overflow,
- int nmi, struct perf_sample_data *data,
+ struct perf_sample_data *data,
struct pt_regs *regs)
{
struct hw_perf_event *hwc = &event->hw;
@@ -5106,7 +4701,7 @@ static void perf_swevent_overflow(struct perf_event *event, u64 overflow,
return;
for (; overflow; overflow--) {
- if (__perf_event_overflow(event, nmi, throttle,
+ if (__perf_event_overflow(event, throttle,
data, regs)) {
/*
* We inhibit the overflow from happening when
@@ -5119,7 +4714,7 @@ static void perf_swevent_overflow(struct perf_event *event, u64 overflow,
}
static void perf_swevent_event(struct perf_event *event, u64 nr,
- int nmi, struct perf_sample_data *data,
+ struct perf_sample_data *data,
struct pt_regs *regs)
{
struct hw_perf_event *hwc = &event->hw;
@@ -5133,12 +4728,12 @@ static void perf_swevent_event(struct perf_event *event, u64 nr,
return;
if (nr == 1 && hwc->sample_period == 1 && !event->attr.freq)
- return perf_swevent_overflow(event, 1, nmi, data, regs);
+ return perf_swevent_overflow(event, 1, data, regs);
if (local64_add_negative(nr, &hwc->period_left))
return;
- perf_swevent_overflow(event, 0, nmi, data, regs);
+ perf_swevent_overflow(event, 0, data, regs);
}
static int perf_exclude_event(struct perf_event *event,
@@ -5226,7 +4821,7 @@ find_swevent_head(struct swevent_htable *swhash, struct perf_event *event)
}
static void do_perf_sw_event(enum perf_type_id type, u32 event_id,
- u64 nr, int nmi,
+ u64 nr,
struct perf_sample_data *data,
struct pt_regs *regs)
{
@@ -5242,7 +4837,7 @@ static void do_perf_sw_event(enum perf_type_id type, u32 event_id,
hlist_for_each_entry_rcu(event, node, head, hlist_entry) {
if (perf_swevent_match(event, type, event_id, data, regs))
- perf_swevent_event(event, nr, nmi, data, regs);
+ perf_swevent_event(event, nr, data, regs);
}
end:
rcu_read_unlock();
@@ -5263,8 +4858,7 @@ inline void perf_swevent_put_recursion_context(int rctx)
put_recursion_context(swhash->recursion, rctx);
}
-void __perf_sw_event(u32 event_id, u64 nr, int nmi,
- struct pt_regs *regs, u64 addr)
+void __perf_sw_event(u32 event_id, u64 nr, struct pt_regs *regs, u64 addr)
{
struct perf_sample_data data;
int rctx;
@@ -5276,7 +4870,7 @@ void __perf_sw_event(u32 event_id, u64 nr, int nmi,
perf_sample_data_init(&data, addr);
- do_perf_sw_event(PERF_TYPE_SOFTWARE, event_id, nr, nmi, &data, regs);
+ do_perf_sw_event(PERF_TYPE_SOFTWARE, event_id, nr, &data, regs);
perf_swevent_put_recursion_context(rctx);
preempt_enable_notrace();
@@ -5524,7 +5118,7 @@ void perf_tp_event(u64 addr, u64 count, void *record, int entry_size,
hlist_for_each_entry_rcu(event, node, head, hlist_entry) {
if (perf_tp_event_match(event, &data, regs))
- perf_swevent_event(event, count, 1, &data, regs);
+ perf_swevent_event(event, count, &data, regs);
}
perf_swevent_put_recursion_context(rctx);
@@ -5617,7 +5211,7 @@ void perf_bp_event(struct perf_event *bp, void *data)
perf_sample_data_init(&sample, bp->attr.bp_addr);
if (!bp->hw.state && !perf_exclude_event(bp, regs))
- perf_swevent_event(bp, 1, 1, &sample, regs);
+ perf_swevent_event(bp, 1, &sample, regs);
}
#endif
@@ -5646,7 +5240,7 @@ static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer)
if (regs && !perf_exclude_event(event, regs)) {
if (!(event->attr.exclude_idle && current->pid == 0))
- if (perf_event_overflow(event, 0, &data, regs))
+ if (perf_event_overflow(event, &data, regs))
ret = HRTIMER_NORESTART;
}
@@ -5986,6 +5580,7 @@ free_dev:
}
static struct lock_class_key cpuctx_mutex;
+static struct lock_class_key cpuctx_lock;
int perf_pmu_register(struct pmu *pmu, char *name, int type)
{
@@ -6036,6 +5631,7 @@ skip_type:
cpuctx = per_cpu_ptr(pmu->pmu_cpu_context, cpu);
__perf_event_init_context(&cpuctx->ctx);
lockdep_set_class(&cpuctx->ctx.mutex, &cpuctx_mutex);
+ lockdep_set_class(&cpuctx->ctx.lock, &cpuctx_lock);
cpuctx->ctx.type = cpu_context;
cpuctx->ctx.pmu = pmu;
cpuctx->jiffies_interval = 1;
@@ -6150,7 +5746,8 @@ perf_event_alloc(struct perf_event_attr *attr, int cpu,
struct task_struct *task,
struct perf_event *group_leader,
struct perf_event *parent_event,
- perf_overflow_handler_t overflow_handler)
+ perf_overflow_handler_t overflow_handler,
+ void *context)
{
struct pmu *pmu;
struct perf_event *event;
@@ -6208,10 +5805,13 @@ perf_event_alloc(struct perf_event_attr *attr, int cpu,
#endif
}
- if (!overflow_handler && parent_event)
+ if (!overflow_handler && parent_event) {
overflow_handler = parent_event->overflow_handler;
+ context = parent_event->overflow_handler_context;
+ }
event->overflow_handler = overflow_handler;
+ event->overflow_handler_context = context;
if (attr->disabled)
event->state = PERF_EVENT_STATE_OFF;
@@ -6354,7 +5954,7 @@ err_size:
static int
perf_event_set_output(struct perf_event *event, struct perf_event *output_event)
{
- struct perf_buffer *buffer = NULL, *old_buffer = NULL;
+ struct ring_buffer *rb = NULL, *old_rb = NULL;
int ret = -EINVAL;
if (!output_event)
@@ -6371,7 +5971,7 @@ perf_event_set_output(struct perf_event *event, struct perf_event *output_event)
goto out;
/*
- * If its not a per-cpu buffer, it must be the same task.
+ * If its not a per-cpu rb, it must be the same task.
*/
if (output_event->cpu == -1 && output_event->ctx != event->ctx)
goto out;
@@ -6383,20 +5983,20 @@ set:
goto unlock;
if (output_event) {
- /* get the buffer we want to redirect to */
- buffer = perf_buffer_get(output_event);
- if (!buffer)
+ /* get the rb we want to redirect to */
+ rb = ring_buffer_get(output_event);
+ if (!rb)
goto unlock;
}
- old_buffer = event->buffer;
- rcu_assign_pointer(event->buffer, buffer);
+ old_rb = event->rb;
+ rcu_assign_pointer(event->rb, rb);
ret = 0;
unlock:
mutex_unlock(&event->mmap_mutex);
- if (old_buffer)
- perf_buffer_put(old_buffer);
+ if (old_rb)
+ ring_buffer_put(old_rb);
out:
return ret;
}
@@ -6478,7 +6078,8 @@ SYSCALL_DEFINE5(perf_event_open,
}
}
- event = perf_event_alloc(&attr, cpu, task, group_leader, NULL, NULL);
+ event = perf_event_alloc(&attr, cpu, task, group_leader, NULL,
+ NULL, NULL);
if (IS_ERR(event)) {
err = PTR_ERR(event);
goto err_task;
@@ -6663,7 +6264,8 @@ err_fd:
struct perf_event *
perf_event_create_kernel_counter(struct perf_event_attr *attr, int cpu,
struct task_struct *task,
- perf_overflow_handler_t overflow_handler)
+ perf_overflow_handler_t overflow_handler,
+ void *context)
{
struct perf_event_context *ctx;
struct perf_event *event;
@@ -6673,7 +6275,8 @@ perf_event_create_kernel_counter(struct perf_event_attr *attr, int cpu,
* Get the target context (task or percpu):
*/
- event = perf_event_alloc(attr, cpu, task, NULL, NULL, overflow_handler);
+ event = perf_event_alloc(attr, cpu, task, NULL, NULL,
+ overflow_handler, context);
if (IS_ERR(event)) {
err = PTR_ERR(event);
goto err;
@@ -6780,7 +6383,6 @@ static void perf_event_exit_task_context(struct task_struct *child, int ctxn)
* our context.
*/
child_ctx = rcu_dereference_raw(child->perf_event_ctxp[ctxn]);
- task_ctx_sched_out(child_ctx, EVENT_ALL);
/*
* Take the context lock here so that if find_get_context is
@@ -6788,6 +6390,7 @@ static void perf_event_exit_task_context(struct task_struct *child, int ctxn)
* incremented the context's refcount before we do put_ctx below.
*/
raw_spin_lock(&child_ctx->lock);
+ task_ctx_sched_out(child_ctx);
child->perf_event_ctxp[ctxn] = NULL;
/*
* If this context is a clone; unclone it so it can't get
@@ -6957,7 +6560,7 @@ inherit_event(struct perf_event *parent_event,
parent_event->cpu,
child,
group_leader, parent_event,
- NULL);
+ NULL, NULL);
if (IS_ERR(child_event))
return child_event;
get_ctx(child_ctx);
@@ -6984,6 +6587,8 @@ inherit_event(struct perf_event *parent_event,
child_event->ctx = child_ctx;
child_event->overflow_handler = parent_event->overflow_handler;
+ child_event->overflow_handler_context
+ = parent_event->overflow_handler_context;
/*
* Precalculate sample_data sizes
diff --git a/kernel/events/hw_breakpoint.c b/kernel/events/hw_breakpoint.c
index 086adf25a55..b7971d6f38b 100644
--- a/kernel/events/hw_breakpoint.c
+++ b/kernel/events/hw_breakpoint.c
@@ -431,9 +431,11 @@ int register_perf_hw_breakpoint(struct perf_event *bp)
struct perf_event *
register_user_hw_breakpoint(struct perf_event_attr *attr,
perf_overflow_handler_t triggered,
+ void *context,
struct task_struct *tsk)
{
- return perf_event_create_kernel_counter(attr, -1, tsk, triggered);
+ return perf_event_create_kernel_counter(attr, -1, tsk, triggered,
+ context);
}
EXPORT_SYMBOL_GPL(register_user_hw_breakpoint);
@@ -502,7 +504,8 @@ EXPORT_SYMBOL_GPL(unregister_hw_breakpoint);
*/
struct perf_event * __percpu *
register_wide_hw_breakpoint(struct perf_event_attr *attr,
- perf_overflow_handler_t triggered)
+ perf_overflow_handler_t triggered,
+ void *context)
{
struct perf_event * __percpu *cpu_events, **pevent, *bp;
long err;
@@ -515,7 +518,8 @@ register_wide_hw_breakpoint(struct perf_event_attr *attr,
get_online_cpus();
for_each_online_cpu(cpu) {
pevent = per_cpu_ptr(cpu_events, cpu);
- bp = perf_event_create_kernel_counter(attr, cpu, NULL, triggered);
+ bp = perf_event_create_kernel_counter(attr, cpu, NULL,
+ triggered, context);
*pevent = bp;
diff --git a/kernel/events/internal.h b/kernel/events/internal.h
new file mode 100644
index 00000000000..09097dd8116
--- /dev/null
+++ b/kernel/events/internal.h
@@ -0,0 +1,96 @@
+#ifndef _KERNEL_EVENTS_INTERNAL_H
+#define _KERNEL_EVENTS_INTERNAL_H
+
+#define RING_BUFFER_WRITABLE 0x01
+
+struct ring_buffer {
+ atomic_t refcount;
+ struct rcu_head rcu_head;
+#ifdef CONFIG_PERF_USE_VMALLOC
+ struct work_struct work;
+ int page_order; /* allocation order */
+#endif
+ int nr_pages; /* nr of data pages */
+ int writable; /* are we writable */
+
+ atomic_t poll; /* POLL_ for wakeups */
+
+ local_t head; /* write position */
+ local_t nest; /* nested writers */
+ local_t events; /* event limit */
+ local_t wakeup; /* wakeup stamp */
+ local_t lost; /* nr records lost */
+
+ long watermark; /* wakeup watermark */
+
+ struct perf_event_mmap_page *user_page;
+ void *data_pages[0];
+};
+
+extern void rb_free(struct ring_buffer *rb);
+extern struct ring_buffer *
+rb_alloc(int nr_pages, long watermark, int cpu, int flags);
+extern void perf_event_wakeup(struct perf_event *event);
+
+extern void
+perf_event_header__init_id(struct perf_event_header *header,
+ struct perf_sample_data *data,
+ struct perf_event *event);
+extern void
+perf_event__output_id_sample(struct perf_event *event,
+ struct perf_output_handle *handle,
+ struct perf_sample_data *sample);
+
+extern struct page *
+perf_mmap_to_page(struct ring_buffer *rb, unsigned long pgoff);
+
+#ifdef CONFIG_PERF_USE_VMALLOC
+/*
+ * Back perf_mmap() with vmalloc memory.
+ *
+ * Required for architectures that have d-cache aliasing issues.
+ */
+
+static inline int page_order(struct ring_buffer *rb)
+{
+ return rb->page_order;
+}
+
+#else
+
+static inline int page_order(struct ring_buffer *rb)
+{
+ return 0;
+}
+#endif
+
+static unsigned long perf_data_size(struct ring_buffer *rb)
+{
+ return rb->nr_pages << (PAGE_SHIFT + page_order(rb));
+}
+
+static inline void
+__output_copy(struct perf_output_handle *handle,
+ const void *buf, unsigned int len)
+{
+ do {
+ unsigned long size = min_t(unsigned long, handle->size, len);
+
+ memcpy(handle->addr, buf, size);
+
+ len -= size;
+ handle->addr += size;
+ buf += size;
+ handle->size -= size;
+ if (!handle->size) {
+ struct ring_buffer *rb = handle->rb;
+
+ handle->page++;
+ handle->page &= rb->nr_pages - 1;
+ handle->addr = rb->data_pages[handle->page];
+ handle->size = PAGE_SIZE << page_order(rb);
+ }
+ } while (len);
+}
+
+#endif /* _KERNEL_EVENTS_INTERNAL_H */
diff --git a/kernel/events/ring_buffer.c b/kernel/events/ring_buffer.c
new file mode 100644
index 00000000000..a2a29205cc0
--- /dev/null
+++ b/kernel/events/ring_buffer.c
@@ -0,0 +1,380 @@
+/*
+ * Performance events ring-buffer code:
+ *
+ * Copyright (C) 2008 Thomas Gleixner <tglx@linutronix.de>
+ * Copyright (C) 2008-2011 Red Hat, Inc., Ingo Molnar
+ * Copyright (C) 2008-2011 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com>
+ * Copyright © 2009 Paul Mackerras, IBM Corp. <paulus@au1.ibm.com>
+ *
+ * For licensing details see kernel-base/COPYING
+ */
+
+#include <linux/perf_event.h>
+#include <linux/vmalloc.h>
+#include <linux/slab.h>
+
+#include "internal.h"
+
+static bool perf_output_space(struct ring_buffer *rb, unsigned long tail,
+ unsigned long offset, unsigned long head)
+{
+ unsigned long mask;
+
+ if (!rb->writable)
+ return true;
+
+ mask = perf_data_size(rb) - 1;
+
+ offset = (offset - tail) & mask;
+ head = (head - tail) & mask;
+
+ if ((int)(head - offset) < 0)
+ return false;
+
+ return true;
+}
+
+static void perf_output_wakeup(struct perf_output_handle *handle)
+{
+ atomic_set(&handle->rb->poll, POLL_IN);
+
+ handle->event->pending_wakeup = 1;
+ irq_work_queue(&handle->event->pending);
+}
+
+/*
+ * We need to ensure a later event_id doesn't publish a head when a former
+ * event isn't done writing. However since we need to deal with NMIs we
+ * cannot fully serialize things.
+ *
+ * We only publish the head (and generate a wakeup) when the outer-most
+ * event completes.
+ */
+static void perf_output_get_handle(struct perf_output_handle *handle)
+{
+ struct ring_buffer *rb = handle->rb;
+
+ preempt_disable();
+ local_inc(&rb->nest);
+ handle->wakeup = local_read(&rb->wakeup);
+}
+
+static void perf_output_put_handle(struct perf_output_handle *handle)
+{
+ struct ring_buffer *rb = handle->rb;
+ unsigned long head;
+
+again:
+ head = local_read(&rb->head);
+
+ /*
+ * IRQ/NMI can happen here, which means we can miss a head update.
+ */
+
+ if (!local_dec_and_test(&rb->nest))
+ goto out;
+
+ /*
+ * Publish the known good head. Rely on the full barrier implied
+ * by atomic_dec_and_test() order the rb->head read and this
+ * write.
+ */
+ rb->user_page->data_head = head;
+
+ /*
+ * Now check if we missed an update, rely on the (compiler)
+ * barrier in atomic_dec_and_test() to re-read rb->head.
+ */
+ if (unlikely(head != local_read(&rb->head))) {
+ local_inc(&rb->nest);
+ goto again;
+ }
+
+ if (handle->wakeup != local_read(&rb->wakeup))
+ perf_output_wakeup(handle);
+
+out:
+ preempt_enable();
+}
+
+int perf_output_begin(struct perf_output_handle *handle,
+ struct perf_event *event, unsigned int size)
+{
+ struct ring_buffer *rb;
+ unsigned long tail, offset, head;
+ int have_lost;
+ struct perf_sample_data sample_data;
+ struct {
+ struct perf_event_header header;
+ u64 id;
+ u64 lost;
+ } lost_event;
+
+ rcu_read_lock();
+ /*
+ * For inherited events we send all the output towards the parent.
+ */
+ if (event->parent)
+ event = event->parent;
+
+ rb = rcu_dereference(event->rb);
+ if (!rb)
+ goto out;
+
+ handle->rb = rb;
+ handle->event = event;
+
+ if (!rb->nr_pages)
+ goto out;
+
+ have_lost = local_read(&rb->lost);
+ if (have_lost) {
+ lost_event.header.size = sizeof(lost_event);
+ perf_event_header__init_id(&lost_event.header, &sample_data,
+ event);
+ size += lost_event.header.size;
+ }
+
+ perf_output_get_handle(handle);
+
+ do {
+ /*
+ * Userspace could choose to issue a mb() before updating the
+ * tail pointer. So that all reads will be completed before the
+ * write is issued.
+ */
+ tail = ACCESS_ONCE(rb->user_page->data_tail);
+ smp_rmb();
+ offset = head = local_read(&rb->head);
+ head += size;
+ if (unlikely(!perf_output_space(rb, tail, offset, head)))
+ goto fail;
+ } while (local_cmpxchg(&rb->head, offset, head) != offset);
+
+ if (head - local_read(&rb->wakeup) > rb->watermark)
+ local_add(rb->watermark, &rb->wakeup);
+
+ handle->page = offset >> (PAGE_SHIFT + page_order(rb));
+ handle->page &= rb->nr_pages - 1;
+ handle->size = offset & ((PAGE_SIZE << page_order(rb)) - 1);
+ handle->addr = rb->data_pages[handle->page];
+ handle->addr += handle->size;
+ handle->size = (PAGE_SIZE << page_order(rb)) - handle->size;
+
+ if (have_lost) {
+ lost_event.header.type = PERF_RECORD_LOST;
+ lost_event.header.misc = 0;
+ lost_event.id = event->id;
+ lost_event.lost = local_xchg(&rb->lost, 0);
+
+ perf_output_put(handle, lost_event);
+ perf_event__output_id_sample(event, handle, &sample_data);
+ }
+
+ return 0;
+
+fail:
+ local_inc(&rb->lost);
+ perf_output_put_handle(handle);
+out:
+ rcu_read_unlock();
+
+ return -ENOSPC;
+}
+
+void perf_output_copy(struct perf_output_handle *handle,
+ const void *buf, unsigned int len)
+{
+ __output_copy(handle, buf, len);
+}
+
+void perf_output_end(struct perf_output_handle *handle)
+{
+ perf_output_put_handle(handle);
+ rcu_read_unlock();
+}
+
+static void
+ring_buffer_init(struct ring_buffer *rb, long watermark, int flags)
+{
+ long max_size = perf_data_size(rb);
+
+ if (watermark)
+ rb->watermark = min(max_size, watermark);
+
+ if (!rb->watermark)
+ rb->watermark = max_size / 2;
+
+ if (flags & RING_BUFFER_WRITABLE)
+ rb->writable = 1;
+
+ atomic_set(&rb->refcount, 1);
+}
+
+#ifndef CONFIG_PERF_USE_VMALLOC
+
+/*
+ * Back perf_mmap() with regular GFP_KERNEL-0 pages.
+ */
+
+struct page *
+perf_mmap_to_page(struct ring_buffer *rb, unsigned long pgoff)
+{
+ if (pgoff > rb->nr_pages)
+ return NULL;
+
+ if (pgoff == 0)
+ return virt_to_page(rb->user_page);
+
+ return virt_to_page(rb->data_pages[pgoff - 1]);
+}
+
+static void *perf_mmap_alloc_page(int cpu)
+{
+ struct page *page;
+ int node;
+
+ node = (cpu == -1) ? cpu : cpu_to_node(cpu);
+ page = alloc_pages_node(node, GFP_KERNEL | __GFP_ZERO, 0);
+ if (!page)
+ return NULL;
+
+ return page_address(page);
+}
+
+struct ring_buffer *rb_alloc(int nr_pages, long watermark, int cpu, int flags)
+{
+ struct ring_buffer *rb;
+ unsigned long size;
+ int i;
+
+ size = sizeof(struct ring_buffer);
+ size += nr_pages * sizeof(void *);
+
+ rb = kzalloc(size, GFP_KERNEL);
+ if (!rb)
+ goto fail;
+
+ rb->user_page = perf_mmap_alloc_page(cpu);
+ if (!rb->user_page)
+ goto fail_user_page;
+
+ for (i = 0; i < nr_pages; i++) {
+ rb->data_pages[i] = perf_mmap_alloc_page(cpu);
+ if (!rb->data_pages[i])
+ goto fail_data_pages;
+ }
+
+ rb->nr_pages = nr_pages;
+
+ ring_buffer_init(rb, watermark, flags);
+
+ return rb;
+
+fail_data_pages:
+ for (i--; i >= 0; i--)
+ free_page((unsigned long)rb->data_pages[i]);
+
+ free_page((unsigned long)rb->user_page);
+
+fail_user_page:
+ kfree(rb);
+
+fail:
+ return NULL;
+}
+
+static void perf_mmap_free_page(unsigned long addr)
+{
+ struct page *page = virt_to_page((void *)addr);
+
+ page->mapping = NULL;
+ __free_page(page);
+}
+
+void rb_free(struct ring_buffer *rb)
+{
+ int i;
+
+ perf_mmap_free_page((unsigned long)rb->user_page);
+ for (i = 0; i < rb->nr_pages; i++)
+ perf_mmap_free_page((unsigned long)rb->data_pages[i]);
+ kfree(rb);
+}
+
+#else
+
+struct page *
+perf_mmap_to_page(struct ring_buffer *rb, unsigned long pgoff)
+{
+ if (pgoff > (1UL << page_order(rb)))
+ return NULL;
+
+ return vmalloc_to_page((void *)rb->user_page + pgoff * PAGE_SIZE);
+}
+
+static void perf_mmap_unmark_page(void *addr)
+{
+ struct page *page = vmalloc_to_page(addr);
+
+ page->mapping = NULL;
+}
+
+static void rb_free_work(struct work_struct *work)
+{
+ struct ring_buffer *rb;
+ void *base;
+ int i, nr;
+
+ rb = container_of(work, struct ring_buffer, work);
+ nr = 1 << page_order(rb);
+
+ base = rb->user_page;
+ for (i = 0; i < nr + 1; i++)
+ perf_mmap_unmark_page(base + (i * PAGE_SIZE));
+
+ vfree(base);
+ kfree(rb);
+}
+
+void rb_free(struct ring_buffer *rb)
+{
+ schedule_work(&rb->work);
+}
+
+struct ring_buffer *rb_alloc(int nr_pages, long watermark, int cpu, int flags)
+{
+ struct ring_buffer *rb;
+ unsigned long size;
+ void *all_buf;
+
+ size = sizeof(struct ring_buffer);
+ size += sizeof(void *);
+
+ rb = kzalloc(size, GFP_KERNEL);
+ if (!rb)
+ goto fail;
+
+ INIT_WORK(&rb->work, rb_free_work);
+
+ all_buf = vmalloc_user((nr_pages + 1) * PAGE_SIZE);
+ if (!all_buf)
+ goto fail_all_buf;
+
+ rb->user_page = all_buf;
+ rb->data_pages[0] = all_buf + PAGE_SIZE;
+ rb->page_order = ilog2(nr_pages);
+ rb->nr_pages = 1;
+
+ ring_buffer_init(rb, watermark, flags);
+
+ return rb;
+
+fail_all_buf:
+ kfree(rb);
+
+fail:
+ return NULL;
+}
+
+#endif
diff --git a/kernel/exit.c b/kernel/exit.c
index 20a40647152..f2b321bae44 100644
--- a/kernel/exit.c
+++ b/kernel/exit.c
@@ -561,29 +561,28 @@ void exit_files(struct task_struct *tsk)
#ifdef CONFIG_MM_OWNER
/*
- * Task p is exiting and it owned mm, lets find a new owner for it
+ * A task is exiting. If it owned this mm, find a new owner for the mm.
*/
-static inline int
-mm_need_new_owner(struct mm_struct *mm, struct task_struct *p)
-{
- /*
- * If there are other users of the mm and the owner (us) is exiting
- * we need to find a new owner to take on the responsibility.
- */
- if (atomic_read(&mm->mm_users) <= 1)
- return 0;
- if (mm->owner != p)
- return 0;
- return 1;
-}
-
void mm_update_next_owner(struct mm_struct *mm)
{
struct task_struct *c, *g, *p = current;
retry:
- if (!mm_need_new_owner(mm, p))
+ /*
+ * If the exiting or execing task is not the owner, it's
+ * someone else's problem.
+ */
+ if (mm->owner != p)
return;
+ /*
+ * The current owner is exiting/execing and there are no other
+ * candidates. Do not leave the mm pointing to a possibly
+ * freed task structure.
+ */
+ if (atomic_read(&mm->mm_users) <= 1) {
+ mm->owner = NULL;
+ return;
+ }
read_lock(&tasklist_lock);
/*
diff --git a/kernel/gcov/Kconfig b/kernel/gcov/Kconfig
index b8cadf70b1f..5bf924d80b5 100644
--- a/kernel/gcov/Kconfig
+++ b/kernel/gcov/Kconfig
@@ -2,7 +2,8 @@ menu "GCOV-based kernel profiling"
config GCOV_KERNEL
bool "Enable gcov-based kernel profiling"
- depends on DEBUG_FS && CONSTRUCTORS
+ depends on DEBUG_FS
+ select CONSTRUCTORS
default n
---help---
This option enables gcov-based code profiling (e.g. for code coverage
diff --git a/kernel/kmod.c b/kernel/kmod.c
index ad6a81c58b4..47613dfb7b2 100644
--- a/kernel/kmod.c
+++ b/kernel/kmod.c
@@ -156,12 +156,6 @@ static int ____call_usermodehelper(void *data)
*/
set_user_nice(current, 0);
- if (sub_info->init) {
- retval = sub_info->init(sub_info);
- if (retval)
- goto fail;
- }
-
retval = -ENOMEM;
new = prepare_kernel_cred(current);
if (!new)
@@ -173,6 +167,14 @@ static int ____call_usermodehelper(void *data)
new->cap_inheritable);
spin_unlock(&umh_sysctl_lock);
+ if (sub_info->init) {
+ retval = sub_info->init(sub_info, new);
+ if (retval) {
+ abort_creds(new);
+ goto fail;
+ }
+ }
+
commit_creds(new);
retval = kernel_execve(sub_info->path,
@@ -388,7 +390,7 @@ EXPORT_SYMBOL(call_usermodehelper_setup);
* context in which call_usermodehelper_exec is called.
*/
void call_usermodehelper_setfns(struct subprocess_info *info,
- int (*init)(struct subprocess_info *info),
+ int (*init)(struct subprocess_info *info, struct cred *new),
void (*cleanup)(struct subprocess_info *info),
void *data)
{
diff --git a/kernel/power/user.c b/kernel/power/user.c
index 7d02d33be69..42ddbc6f0de 100644
--- a/kernel/power/user.c
+++ b/kernel/power/user.c
@@ -113,8 +113,10 @@ static int snapshot_open(struct inode *inode, struct file *filp)
if (error)
pm_notifier_call_chain(PM_POST_RESTORE);
}
- if (error)
+ if (error) {
+ free_basic_memory_bitmaps();
atomic_inc(&snapshot_device_available);
+ }
data->frozen = 0;
data->ready = 0;
data->platform_support = 0;
diff --git a/kernel/rcutree.c b/kernel/rcutree.c
index 89419ff92e9..7e59ffb3d0b 100644
--- a/kernel/rcutree.c
+++ b/kernel/rcutree.c
@@ -87,6 +87,8 @@ static struct rcu_state *rcu_state;
int rcu_scheduler_active __read_mostly;
EXPORT_SYMBOL_GPL(rcu_scheduler_active);
+#ifdef CONFIG_RCU_BOOST
+
/*
* Control variables for per-CPU and per-rcu_node kthreads. These
* handle all flavors of RCU.
@@ -98,8 +100,11 @@ DEFINE_PER_CPU(unsigned int, rcu_cpu_kthread_loops);
DEFINE_PER_CPU(char, rcu_cpu_has_work);
static char rcu_kthreads_spawnable;
+#endif /* #ifdef CONFIG_RCU_BOOST */
+
static void rcu_node_kthread_setaffinity(struct rcu_node *rnp, int outgoingcpu);
-static void invoke_rcu_cpu_kthread(void);
+static void invoke_rcu_core(void);
+static void invoke_rcu_callbacks(struct rcu_state *rsp, struct rcu_data *rdp);
#define RCU_KTHREAD_PRIO 1 /* RT priority for per-CPU kthreads. */
@@ -1088,14 +1093,8 @@ static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp)
int need_report = 0;
struct rcu_data *rdp = per_cpu_ptr(rsp->rda, cpu);
struct rcu_node *rnp;
- struct task_struct *t;
- /* Stop the CPU's kthread. */
- t = per_cpu(rcu_cpu_kthread_task, cpu);
- if (t != NULL) {
- per_cpu(rcu_cpu_kthread_task, cpu) = NULL;
- kthread_stop(t);
- }
+ rcu_stop_cpu_kthread(cpu);
/* Exclude any attempts to start a new grace period. */
raw_spin_lock_irqsave(&rsp->onofflock, flags);
@@ -1231,7 +1230,7 @@ static void rcu_do_batch(struct rcu_state *rsp, struct rcu_data *rdp)
/* Re-raise the RCU softirq if there are callbacks remaining. */
if (cpu_has_callbacks_ready_to_invoke(rdp))
- invoke_rcu_cpu_kthread();
+ invoke_rcu_core();
}
/*
@@ -1277,7 +1276,7 @@ void rcu_check_callbacks(int cpu, int user)
}
rcu_preempt_check_callbacks(cpu);
if (rcu_pending(cpu))
- invoke_rcu_cpu_kthread();
+ invoke_rcu_core();
}
#ifdef CONFIG_SMP
@@ -1442,13 +1441,14 @@ __rcu_process_callbacks(struct rcu_state *rsp, struct rcu_data *rdp)
}
/* If there are callbacks ready, invoke them. */
- rcu_do_batch(rsp, rdp);
+ if (cpu_has_callbacks_ready_to_invoke(rdp))
+ invoke_rcu_callbacks(rsp, rdp);
}
/*
* Do softirq processing for the current CPU.
*/
-static void rcu_process_callbacks(void)
+static void rcu_process_callbacks(struct softirq_action *unused)
{
__rcu_process_callbacks(&rcu_sched_state,
&__get_cpu_var(rcu_sched_data));
@@ -1465,342 +1465,20 @@ static void rcu_process_callbacks(void)
* the current CPU with interrupts disabled, the rcu_cpu_kthread_task
* cannot disappear out from under us.
*/
-static void invoke_rcu_cpu_kthread(void)
-{
- unsigned long flags;
-
- local_irq_save(flags);
- __this_cpu_write(rcu_cpu_has_work, 1);
- if (__this_cpu_read(rcu_cpu_kthread_task) == NULL) {
- local_irq_restore(flags);
- return;
- }
- wake_up_process(__this_cpu_read(rcu_cpu_kthread_task));
- local_irq_restore(flags);
-}
-
-/*
- * Wake up the specified per-rcu_node-structure kthread.
- * Because the per-rcu_node kthreads are immortal, we don't need
- * to do anything to keep them alive.
- */
-static void invoke_rcu_node_kthread(struct rcu_node *rnp)
-{
- struct task_struct *t;
-
- t = rnp->node_kthread_task;
- if (t != NULL)
- wake_up_process(t);
-}
-
-/*
- * Set the specified CPU's kthread to run RT or not, as specified by
- * the to_rt argument. The CPU-hotplug locks are held, so the task
- * is not going away.
- */
-static void rcu_cpu_kthread_setrt(int cpu, int to_rt)
-{
- int policy;
- struct sched_param sp;
- struct task_struct *t;
-
- t = per_cpu(rcu_cpu_kthread_task, cpu);
- if (t == NULL)
- return;
- if (to_rt) {
- policy = SCHED_FIFO;
- sp.sched_priority = RCU_KTHREAD_PRIO;
- } else {
- policy = SCHED_NORMAL;
- sp.sched_priority = 0;
- }
- sched_setscheduler_nocheck(t, policy, &sp);
-}
-
-/*
- * Timer handler to initiate the waking up of per-CPU kthreads that
- * have yielded the CPU due to excess numbers of RCU callbacks.
- * We wake up the per-rcu_node kthread, which in turn will wake up
- * the booster kthread.
- */
-static void rcu_cpu_kthread_timer(unsigned long arg)
-{
- struct rcu_data *rdp = per_cpu_ptr(rcu_state->rda, arg);
- struct rcu_node *rnp = rdp->mynode;
-
- atomic_or(rdp->grpmask, &rnp->wakemask);
- invoke_rcu_node_kthread(rnp);
-}
-
-/*
- * Drop to non-real-time priority and yield, but only after posting a
- * timer that will cause us to regain our real-time priority if we
- * remain preempted. Either way, we restore our real-time priority
- * before returning.
- */
-static void rcu_yield(void (*f)(unsigned long), unsigned long arg)
-{
- struct sched_param sp;
- struct timer_list yield_timer;
-
- setup_timer_on_stack(&yield_timer, f, arg);
- mod_timer(&yield_timer, jiffies + 2);
- sp.sched_priority = 0;
- sched_setscheduler_nocheck(current, SCHED_NORMAL, &sp);
- set_user_nice(current, 19);
- schedule();
- sp.sched_priority = RCU_KTHREAD_PRIO;
- sched_setscheduler_nocheck(current, SCHED_FIFO, &sp);
- del_timer(&yield_timer);
-}
-
-/*
- * Handle cases where the rcu_cpu_kthread() ends up on the wrong CPU.
- * This can happen while the corresponding CPU is either coming online
- * or going offline. We cannot wait until the CPU is fully online
- * before starting the kthread, because the various notifier functions
- * can wait for RCU grace periods. So we park rcu_cpu_kthread() until
- * the corresponding CPU is online.
- *
- * Return 1 if the kthread needs to stop, 0 otherwise.
- *
- * Caller must disable bh. This function can momentarily enable it.
- */
-static int rcu_cpu_kthread_should_stop(int cpu)
-{
- while (cpu_is_offline(cpu) ||
- !cpumask_equal(&current->cpus_allowed, cpumask_of(cpu)) ||
- smp_processor_id() != cpu) {
- if (kthread_should_stop())
- return 1;
- per_cpu(rcu_cpu_kthread_status, cpu) = RCU_KTHREAD_OFFCPU;
- per_cpu(rcu_cpu_kthread_cpu, cpu) = raw_smp_processor_id();
- local_bh_enable();
- schedule_timeout_uninterruptible(1);
- if (!cpumask_equal(&current->cpus_allowed, cpumask_of(cpu)))
- set_cpus_allowed_ptr(current, cpumask_of(cpu));
- local_bh_disable();
- }
- per_cpu(rcu_cpu_kthread_cpu, cpu) = cpu;
- return 0;
-}
-
-/*
- * Per-CPU kernel thread that invokes RCU callbacks. This replaces the
- * earlier RCU softirq.
- */
-static int rcu_cpu_kthread(void *arg)
-{
- int cpu = (int)(long)arg;
- unsigned long flags;
- int spincnt = 0;
- unsigned int *statusp = &per_cpu(rcu_cpu_kthread_status, cpu);
- char work;
- char *workp = &per_cpu(rcu_cpu_has_work, cpu);
-
- for (;;) {
- *statusp = RCU_KTHREAD_WAITING;
- rcu_wait(*workp != 0 || kthread_should_stop());
- local_bh_disable();
- if (rcu_cpu_kthread_should_stop(cpu)) {
- local_bh_enable();
- break;
- }
- *statusp = RCU_KTHREAD_RUNNING;
- per_cpu(rcu_cpu_kthread_loops, cpu)++;
- local_irq_save(flags);
- work = *workp;
- *workp = 0;
- local_irq_restore(flags);
- if (work)
- rcu_process_callbacks();
- local_bh_enable();
- if (*workp != 0)
- spincnt++;
- else
- spincnt = 0;
- if (spincnt > 10) {
- *statusp = RCU_KTHREAD_YIELDING;
- rcu_yield(rcu_cpu_kthread_timer, (unsigned long)cpu);
- spincnt = 0;
- }
- }
- *statusp = RCU_KTHREAD_STOPPED;
- return 0;
-}
-
-/*
- * Spawn a per-CPU kthread, setting up affinity and priority.
- * Because the CPU hotplug lock is held, no other CPU will be attempting
- * to manipulate rcu_cpu_kthread_task. There might be another CPU
- * attempting to access it during boot, but the locking in kthread_bind()
- * will enforce sufficient ordering.
- */
-static int __cpuinit rcu_spawn_one_cpu_kthread(int cpu)
+static void invoke_rcu_callbacks(struct rcu_state *rsp, struct rcu_data *rdp)
{
- struct sched_param sp;
- struct task_struct *t;
-
- if (!rcu_kthreads_spawnable ||
- per_cpu(rcu_cpu_kthread_task, cpu) != NULL)
- return 0;
- t = kthread_create(rcu_cpu_kthread, (void *)(long)cpu, "rcuc%d", cpu);
- if (IS_ERR(t))
- return PTR_ERR(t);
- kthread_bind(t, cpu);
- per_cpu(rcu_cpu_kthread_cpu, cpu) = cpu;
- WARN_ON_ONCE(per_cpu(rcu_cpu_kthread_task, cpu) != NULL);
- per_cpu(rcu_cpu_kthread_task, cpu) = t;
- sp.sched_priority = RCU_KTHREAD_PRIO;
- sched_setscheduler_nocheck(t, SCHED_FIFO, &sp);
- return 0;
-}
-
-/*
- * Per-rcu_node kthread, which is in charge of waking up the per-CPU
- * kthreads when needed. We ignore requests to wake up kthreads
- * for offline CPUs, which is OK because force_quiescent_state()
- * takes care of this case.
- */
-static int rcu_node_kthread(void *arg)
-{
- int cpu;
- unsigned long flags;
- unsigned long mask;
- struct rcu_node *rnp = (struct rcu_node *)arg;
- struct sched_param sp;
- struct task_struct *t;
-
- for (;;) {
- rnp->node_kthread_status = RCU_KTHREAD_WAITING;
- rcu_wait(atomic_read(&rnp->wakemask) != 0);
- rnp->node_kthread_status = RCU_KTHREAD_RUNNING;
- raw_spin_lock_irqsave(&rnp->lock, flags);
- mask = atomic_xchg(&rnp->wakemask, 0);
- rcu_initiate_boost(rnp, flags); /* releases rnp->lock. */
- for (cpu = rnp->grplo; cpu <= rnp->grphi; cpu++, mask >>= 1) {
- if ((mask & 0x1) == 0)
- continue;
- preempt_disable();
- t = per_cpu(rcu_cpu_kthread_task, cpu);
- if (!cpu_online(cpu) || t == NULL) {
- preempt_enable();
- continue;
- }
- per_cpu(rcu_cpu_has_work, cpu) = 1;
- sp.sched_priority = RCU_KTHREAD_PRIO;
- sched_setscheduler_nocheck(t, SCHED_FIFO, &sp);
- preempt_enable();
- }
- }
- /* NOTREACHED */
- rnp->node_kthread_status = RCU_KTHREAD_STOPPED;
- return 0;
-}
-
-/*
- * Set the per-rcu_node kthread's affinity to cover all CPUs that are
- * served by the rcu_node in question. The CPU hotplug lock is still
- * held, so the value of rnp->qsmaskinit will be stable.
- *
- * We don't include outgoingcpu in the affinity set, use -1 if there is
- * no outgoing CPU. If there are no CPUs left in the affinity set,
- * this function allows the kthread to execute on any CPU.
- */
-static void rcu_node_kthread_setaffinity(struct rcu_node *rnp, int outgoingcpu)
-{
- cpumask_var_t cm;
- int cpu;
- unsigned long mask = rnp->qsmaskinit;
-
- if (rnp->node_kthread_task == NULL)
- return;
- if (!alloc_cpumask_var(&cm, GFP_KERNEL))
+ if (likely(!rsp->boost)) {
+ rcu_do_batch(rsp, rdp);
return;
- cpumask_clear(cm);
- for (cpu = rnp->grplo; cpu <= rnp->grphi; cpu++, mask >>= 1)
- if ((mask & 0x1) && cpu != outgoingcpu)
- cpumask_set_cpu(cpu, cm);
- if (cpumask_weight(cm) == 0) {
- cpumask_setall(cm);
- for (cpu = rnp->grplo; cpu <= rnp->grphi; cpu++)
- cpumask_clear_cpu(cpu, cm);
- WARN_ON_ONCE(cpumask_weight(cm) == 0);
}
- set_cpus_allowed_ptr(rnp->node_kthread_task, cm);
- rcu_boost_kthread_setaffinity(rnp, cm);
- free_cpumask_var(cm);
+ invoke_rcu_callbacks_kthread();
}
-/*
- * Spawn a per-rcu_node kthread, setting priority and affinity.
- * Called during boot before online/offline can happen, or, if
- * during runtime, with the main CPU-hotplug locks held. So only
- * one of these can be executing at a time.
- */
-static int __cpuinit rcu_spawn_one_node_kthread(struct rcu_state *rsp,
- struct rcu_node *rnp)
+static void invoke_rcu_core(void)
{
- unsigned long flags;
- int rnp_index = rnp - &rsp->node[0];
- struct sched_param sp;
- struct task_struct *t;
-
- if (!rcu_kthreads_spawnable ||
- rnp->qsmaskinit == 0)
- return 0;
- if (rnp->node_kthread_task == NULL) {
- t = kthread_create(rcu_node_kthread, (void *)rnp,
- "rcun%d", rnp_index);
- if (IS_ERR(t))
- return PTR_ERR(t);
- raw_spin_lock_irqsave(&rnp->lock, flags);
- rnp->node_kthread_task = t;
- raw_spin_unlock_irqrestore(&rnp->lock, flags);
- sp.sched_priority = 99;
- sched_setscheduler_nocheck(t, SCHED_FIFO, &sp);
- }
- return rcu_spawn_one_boost_kthread(rsp, rnp, rnp_index);
+ raise_softirq(RCU_SOFTIRQ);
}
-static void rcu_wake_one_boost_kthread(struct rcu_node *rnp);
-
-/*
- * Spawn all kthreads -- called as soon as the scheduler is running.
- */
-static int __init rcu_spawn_kthreads(void)
-{
- int cpu;
- struct rcu_node *rnp;
- struct task_struct *t;
-
- rcu_kthreads_spawnable = 1;
- for_each_possible_cpu(cpu) {
- per_cpu(rcu_cpu_has_work, cpu) = 0;
- if (cpu_online(cpu)) {
- (void)rcu_spawn_one_cpu_kthread(cpu);
- t = per_cpu(rcu_cpu_kthread_task, cpu);
- if (t)
- wake_up_process(t);
- }
- }
- rnp = rcu_get_root(rcu_state);
- (void)rcu_spawn_one_node_kthread(rcu_state, rnp);
- if (rnp->node_kthread_task)
- wake_up_process(rnp->node_kthread_task);
- if (NUM_RCU_NODES > 1) {
- rcu_for_each_leaf_node(rcu_state, rnp) {
- (void)rcu_spawn_one_node_kthread(rcu_state, rnp);
- t = rnp->node_kthread_task;
- if (t)
- wake_up_process(t);
- rcu_wake_one_boost_kthread(rnp);
- }
- }
- return 0;
-}
-early_initcall(rcu_spawn_kthreads);
-
static void
__call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu),
struct rcu_state *rsp)
@@ -2207,44 +1885,6 @@ static void __cpuinit rcu_prepare_cpu(int cpu)
rcu_preempt_init_percpu_data(cpu);
}
-static void __cpuinit rcu_prepare_kthreads(int cpu)
-{
- struct rcu_data *rdp = per_cpu_ptr(rcu_state->rda, cpu);
- struct rcu_node *rnp = rdp->mynode;
-
- /* Fire up the incoming CPU's kthread and leaf rcu_node kthread. */
- if (rcu_kthreads_spawnable) {
- (void)rcu_spawn_one_cpu_kthread(cpu);
- if (rnp->node_kthread_task == NULL)
- (void)rcu_spawn_one_node_kthread(rcu_state, rnp);
- }
-}
-
-/*
- * kthread_create() creates threads in TASK_UNINTERRUPTIBLE state,
- * but the RCU threads are woken on demand, and if demand is low this
- * could be a while triggering the hung task watchdog.
- *
- * In order to avoid this, poke all tasks once the CPU is fully
- * up and running.
- */
-static void __cpuinit rcu_online_kthreads(int cpu)
-{
- struct rcu_data *rdp = per_cpu_ptr(rcu_state->rda, cpu);
- struct rcu_node *rnp = rdp->mynode;
- struct task_struct *t;
-
- t = per_cpu(rcu_cpu_kthread_task, cpu);
- if (t)
- wake_up_process(t);
-
- t = rnp->node_kthread_task;
- if (t)
- wake_up_process(t);
-
- rcu_wake_one_boost_kthread(rnp);
-}
-
/*
* Handle CPU online/offline notification events.
*/
@@ -2262,7 +1902,6 @@ static int __cpuinit rcu_cpu_notify(struct notifier_block *self,
rcu_prepare_kthreads(cpu);
break;
case CPU_ONLINE:
- rcu_online_kthreads(cpu);
case CPU_DOWN_FAILED:
rcu_node_kthread_setaffinity(rnp, -1);
rcu_cpu_kthread_setrt(cpu, 1);
@@ -2410,6 +2049,7 @@ void __init rcu_init(void)
rcu_init_one(&rcu_sched_state, &rcu_sched_data);
rcu_init_one(&rcu_bh_state, &rcu_bh_data);
__rcu_init_preempt();
+ open_softirq(RCU_SOFTIRQ, rcu_process_callbacks);
/*
* We don't need protection against CPU-hotplug here because
diff --git a/kernel/rcutree.h b/kernel/rcutree.h
index 7b9a08b4aae..01b2ccda26f 100644
--- a/kernel/rcutree.h
+++ b/kernel/rcutree.h
@@ -369,6 +369,7 @@ struct rcu_state {
/* period because */
/* force_quiescent_state() */
/* was running. */
+ u8 boost; /* Subject to priority boost. */
unsigned long gpnum; /* Current gp number. */
unsigned long completed; /* # of last completed gp. */
@@ -426,6 +427,7 @@ static int rcu_preempt_blocked_readers_cgp(struct rcu_node *rnp);
#ifdef CONFIG_HOTPLUG_CPU
static void rcu_report_unblock_qs_rnp(struct rcu_node *rnp,
unsigned long flags);
+static void rcu_stop_cpu_kthread(int cpu);
#endif /* #ifdef CONFIG_HOTPLUG_CPU */
static void rcu_print_detail_task_stall(struct rcu_state *rsp);
static void rcu_print_task_stall(struct rcu_node *rnp);
@@ -450,11 +452,19 @@ static void rcu_preempt_send_cbs_to_online(void);
static void __init __rcu_init_preempt(void);
static void rcu_needs_cpu_flush(void);
static void rcu_initiate_boost(struct rcu_node *rnp, unsigned long flags);
+static void rcu_preempt_boost_start_gp(struct rcu_node *rnp);
+static void invoke_rcu_callbacks_kthread(void);
+#ifdef CONFIG_RCU_BOOST
+static void rcu_preempt_do_callbacks(void);
static void rcu_boost_kthread_setaffinity(struct rcu_node *rnp,
cpumask_var_t cm);
-static void rcu_preempt_boost_start_gp(struct rcu_node *rnp);
static int __cpuinit rcu_spawn_one_boost_kthread(struct rcu_state *rsp,
struct rcu_node *rnp,
int rnp_index);
+static void invoke_rcu_node_kthread(struct rcu_node *rnp);
+static void rcu_yield(void (*f)(unsigned long), unsigned long arg);
+#endif /* #ifdef CONFIG_RCU_BOOST */
+static void rcu_cpu_kthread_setrt(int cpu, int to_rt);
+static void __cpuinit rcu_prepare_kthreads(int cpu);
#endif /* #ifndef RCU_TREE_NONCORE */
diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h
index c8bff3099a8..14dc7dd0090 100644
--- a/kernel/rcutree_plugin.h
+++ b/kernel/rcutree_plugin.h
@@ -602,6 +602,15 @@ static void rcu_preempt_process_callbacks(void)
&__get_cpu_var(rcu_preempt_data));
}
+#ifdef CONFIG_RCU_BOOST
+
+static void rcu_preempt_do_callbacks(void)
+{
+ rcu_do_batch(&rcu_preempt_state, &__get_cpu_var(rcu_preempt_data));
+}
+
+#endif /* #ifdef CONFIG_RCU_BOOST */
+
/*
* Queue a preemptible-RCU callback for invocation after a grace period.
*/
@@ -1249,6 +1258,23 @@ static void rcu_initiate_boost(struct rcu_node *rnp, unsigned long flags)
}
/*
+ * Wake up the per-CPU kthread to invoke RCU callbacks.
+ */
+static void invoke_rcu_callbacks_kthread(void)
+{
+ unsigned long flags;
+
+ local_irq_save(flags);
+ __this_cpu_write(rcu_cpu_has_work, 1);
+ if (__this_cpu_read(rcu_cpu_kthread_task) == NULL) {
+ local_irq_restore(flags);
+ return;
+ }
+ wake_up_process(__this_cpu_read(rcu_cpu_kthread_task));
+ local_irq_restore(flags);
+}
+
+/*
* Set the affinity of the boost kthread. The CPU-hotplug locks are
* held, so no one should be messing with the existence of the boost
* kthread.
@@ -1288,6 +1314,7 @@ static int __cpuinit rcu_spawn_one_boost_kthread(struct rcu_state *rsp,
if (&rcu_preempt_state != rsp)
return 0;
+ rsp->boost = 1;
if (rnp->boost_kthread_task != NULL)
return 0;
t = kthread_create(rcu_boost_kthread, (void *)rnp,
@@ -1299,13 +1326,372 @@ static int __cpuinit rcu_spawn_one_boost_kthread(struct rcu_state *rsp,
raw_spin_unlock_irqrestore(&rnp->lock, flags);
sp.sched_priority = RCU_KTHREAD_PRIO;
sched_setscheduler_nocheck(t, SCHED_FIFO, &sp);
+ wake_up_process(t); /* get to TASK_INTERRUPTIBLE quickly. */
return 0;
}
-static void __cpuinit rcu_wake_one_boost_kthread(struct rcu_node *rnp)
+#ifdef CONFIG_HOTPLUG_CPU
+
+/*
+ * Stop the RCU's per-CPU kthread when its CPU goes offline,.
+ */
+static void rcu_stop_cpu_kthread(int cpu)
{
- if (rnp->boost_kthread_task)
- wake_up_process(rnp->boost_kthread_task);
+ struct task_struct *t;
+
+ /* Stop the CPU's kthread. */
+ t = per_cpu(rcu_cpu_kthread_task, cpu);
+ if (t != NULL) {
+ per_cpu(rcu_cpu_kthread_task, cpu) = NULL;
+ kthread_stop(t);
+ }
+}
+
+#endif /* #ifdef CONFIG_HOTPLUG_CPU */
+
+static void rcu_kthread_do_work(void)
+{
+ rcu_do_batch(&rcu_sched_state, &__get_cpu_var(rcu_sched_data));
+ rcu_do_batch(&rcu_bh_state, &__get_cpu_var(rcu_bh_data));
+ rcu_preempt_do_callbacks();
+}
+
+/*
+ * Wake up the specified per-rcu_node-structure kthread.
+ * Because the per-rcu_node kthreads are immortal, we don't need
+ * to do anything to keep them alive.
+ */
+static void invoke_rcu_node_kthread(struct rcu_node *rnp)
+{
+ struct task_struct *t;
+
+ t = rnp->node_kthread_task;
+ if (t != NULL)
+ wake_up_process(t);
+}
+
+/*
+ * Set the specified CPU's kthread to run RT or not, as specified by
+ * the to_rt argument. The CPU-hotplug locks are held, so the task
+ * is not going away.
+ */
+static void rcu_cpu_kthread_setrt(int cpu, int to_rt)
+{
+ int policy;
+ struct sched_param sp;
+ struct task_struct *t;
+
+ t = per_cpu(rcu_cpu_kthread_task, cpu);
+ if (t == NULL)
+ return;
+ if (to_rt) {
+ policy = SCHED_FIFO;
+ sp.sched_priority = RCU_KTHREAD_PRIO;
+ } else {
+ policy = SCHED_NORMAL;
+ sp.sched_priority = 0;
+ }
+ sched_setscheduler_nocheck(t, policy, &sp);
+}
+
+/*
+ * Timer handler to initiate the waking up of per-CPU kthreads that
+ * have yielded the CPU due to excess numbers of RCU callbacks.
+ * We wake up the per-rcu_node kthread, which in turn will wake up
+ * the booster kthread.
+ */
+static void rcu_cpu_kthread_timer(unsigned long arg)
+{
+ struct rcu_data *rdp = per_cpu_ptr(rcu_state->rda, arg);
+ struct rcu_node *rnp = rdp->mynode;
+
+ atomic_or(rdp->grpmask, &rnp->wakemask);
+ invoke_rcu_node_kthread(rnp);
+}
+
+/*
+ * Drop to non-real-time priority and yield, but only after posting a
+ * timer that will cause us to regain our real-time priority if we
+ * remain preempted. Either way, we restore our real-time priority
+ * before returning.
+ */
+static void rcu_yield(void (*f)(unsigned long), unsigned long arg)
+{
+ struct sched_param sp;
+ struct timer_list yield_timer;
+
+ setup_timer_on_stack(&yield_timer, f, arg);
+ mod_timer(&yield_timer, jiffies + 2);
+ sp.sched_priority = 0;
+ sched_setscheduler_nocheck(current, SCHED_NORMAL, &sp);
+ set_user_nice(current, 19);
+ schedule();
+ sp.sched_priority = RCU_KTHREAD_PRIO;
+ sched_setscheduler_nocheck(current, SCHED_FIFO, &sp);
+ del_timer(&yield_timer);
+}
+
+/*
+ * Handle cases where the rcu_cpu_kthread() ends up on the wrong CPU.
+ * This can happen while the corresponding CPU is either coming online
+ * or going offline. We cannot wait until the CPU is fully online
+ * before starting the kthread, because the various notifier functions
+ * can wait for RCU grace periods. So we park rcu_cpu_kthread() until
+ * the corresponding CPU is online.
+ *
+ * Return 1 if the kthread needs to stop, 0 otherwise.
+ *
+ * Caller must disable bh. This function can momentarily enable it.
+ */
+static int rcu_cpu_kthread_should_stop(int cpu)
+{
+ while (cpu_is_offline(cpu) ||
+ !cpumask_equal(&current->cpus_allowed, cpumask_of(cpu)) ||
+ smp_processor_id() != cpu) {
+ if (kthread_should_stop())
+ return 1;
+ per_cpu(rcu_cpu_kthread_status, cpu) = RCU_KTHREAD_OFFCPU;
+ per_cpu(rcu_cpu_kthread_cpu, cpu) = raw_smp_processor_id();
+ local_bh_enable();
+ schedule_timeout_uninterruptible(1);
+ if (!cpumask_equal(&current->cpus_allowed, cpumask_of(cpu)))
+ set_cpus_allowed_ptr(current, cpumask_of(cpu));
+ local_bh_disable();
+ }
+ per_cpu(rcu_cpu_kthread_cpu, cpu) = cpu;
+ return 0;
+}
+
+/*
+ * Per-CPU kernel thread that invokes RCU callbacks. This replaces the
+ * earlier RCU softirq.
+ */
+static int rcu_cpu_kthread(void *arg)
+{
+ int cpu = (int)(long)arg;
+ unsigned long flags;
+ int spincnt = 0;
+ unsigned int *statusp = &per_cpu(rcu_cpu_kthread_status, cpu);
+ char work;
+ char *workp = &per_cpu(rcu_cpu_has_work, cpu);
+
+ for (;;) {
+ *statusp = RCU_KTHREAD_WAITING;
+ rcu_wait(*workp != 0 || kthread_should_stop());
+ local_bh_disable();
+ if (rcu_cpu_kthread_should_stop(cpu)) {
+ local_bh_enable();
+ break;
+ }
+ *statusp = RCU_KTHREAD_RUNNING;
+ per_cpu(rcu_cpu_kthread_loops, cpu)++;
+ local_irq_save(flags);
+ work = *workp;
+ *workp = 0;
+ local_irq_restore(flags);
+ if (work)
+ rcu_kthread_do_work();
+ local_bh_enable();
+ if (*workp != 0)
+ spincnt++;
+ else
+ spincnt = 0;
+ if (spincnt > 10) {
+ *statusp = RCU_KTHREAD_YIELDING;
+ rcu_yield(rcu_cpu_kthread_timer, (unsigned long)cpu);
+ spincnt = 0;
+ }
+ }
+ *statusp = RCU_KTHREAD_STOPPED;
+ return 0;
+}
+
+/*
+ * Spawn a per-CPU kthread, setting up affinity and priority.
+ * Because the CPU hotplug lock is held, no other CPU will be attempting
+ * to manipulate rcu_cpu_kthread_task. There might be another CPU
+ * attempting to access it during boot, but the locking in kthread_bind()
+ * will enforce sufficient ordering.
+ *
+ * Please note that we cannot simply refuse to wake up the per-CPU
+ * kthread because kthreads are created in TASK_UNINTERRUPTIBLE state,
+ * which can result in softlockup complaints if the task ends up being
+ * idle for more than a couple of minutes.
+ *
+ * However, please note also that we cannot bind the per-CPU kthread to its
+ * CPU until that CPU is fully online. We also cannot wait until the
+ * CPU is fully online before we create its per-CPU kthread, as this would
+ * deadlock the system when CPU notifiers tried waiting for grace
+ * periods. So we bind the per-CPU kthread to its CPU only if the CPU
+ * is online. If its CPU is not yet fully online, then the code in
+ * rcu_cpu_kthread() will wait until it is fully online, and then do
+ * the binding.
+ */
+static int __cpuinit rcu_spawn_one_cpu_kthread(int cpu)
+{
+ struct sched_param sp;
+ struct task_struct *t;
+
+ if (!rcu_kthreads_spawnable ||
+ per_cpu(rcu_cpu_kthread_task, cpu) != NULL)
+ return 0;
+ t = kthread_create(rcu_cpu_kthread, (void *)(long)cpu, "rcuc%d", cpu);
+ if (IS_ERR(t))
+ return PTR_ERR(t);
+ if (cpu_online(cpu))
+ kthread_bind(t, cpu);
+ per_cpu(rcu_cpu_kthread_cpu, cpu) = cpu;
+ WARN_ON_ONCE(per_cpu(rcu_cpu_kthread_task, cpu) != NULL);
+ sp.sched_priority = RCU_KTHREAD_PRIO;
+ sched_setscheduler_nocheck(t, SCHED_FIFO, &sp);
+ per_cpu(rcu_cpu_kthread_task, cpu) = t;
+ wake_up_process(t); /* Get to TASK_INTERRUPTIBLE quickly. */
+ return 0;
+}
+
+/*
+ * Per-rcu_node kthread, which is in charge of waking up the per-CPU
+ * kthreads when needed. We ignore requests to wake up kthreads
+ * for offline CPUs, which is OK because force_quiescent_state()
+ * takes care of this case.
+ */
+static int rcu_node_kthread(void *arg)
+{
+ int cpu;
+ unsigned long flags;
+ unsigned long mask;
+ struct rcu_node *rnp = (struct rcu_node *)arg;
+ struct sched_param sp;
+ struct task_struct *t;
+
+ for (;;) {
+ rnp->node_kthread_status = RCU_KTHREAD_WAITING;
+ rcu_wait(atomic_read(&rnp->wakemask) != 0);
+ rnp->node_kthread_status = RCU_KTHREAD_RUNNING;
+ raw_spin_lock_irqsave(&rnp->lock, flags);
+ mask = atomic_xchg(&rnp->wakemask, 0);
+ rcu_initiate_boost(rnp, flags); /* releases rnp->lock. */
+ for (cpu = rnp->grplo; cpu <= rnp->grphi; cpu++, mask >>= 1) {
+ if ((mask & 0x1) == 0)
+ continue;
+ preempt_disable();
+ t = per_cpu(rcu_cpu_kthread_task, cpu);
+ if (!cpu_online(cpu) || t == NULL) {
+ preempt_enable();
+ continue;
+ }
+ per_cpu(rcu_cpu_has_work, cpu) = 1;
+ sp.sched_priority = RCU_KTHREAD_PRIO;
+ sched_setscheduler_nocheck(t, SCHED_FIFO, &sp);
+ preempt_enable();
+ }
+ }
+ /* NOTREACHED */
+ rnp->node_kthread_status = RCU_KTHREAD_STOPPED;
+ return 0;
+}
+
+/*
+ * Set the per-rcu_node kthread's affinity to cover all CPUs that are
+ * served by the rcu_node in question. The CPU hotplug lock is still
+ * held, so the value of rnp->qsmaskinit will be stable.
+ *
+ * We don't include outgoingcpu in the affinity set, use -1 if there is
+ * no outgoing CPU. If there are no CPUs left in the affinity set,
+ * this function allows the kthread to execute on any CPU.
+ */
+static void rcu_node_kthread_setaffinity(struct rcu_node *rnp, int outgoingcpu)
+{
+ cpumask_var_t cm;
+ int cpu;
+ unsigned long mask = rnp->qsmaskinit;
+
+ if (rnp->node_kthread_task == NULL)
+ return;
+ if (!alloc_cpumask_var(&cm, GFP_KERNEL))
+ return;
+ cpumask_clear(cm);
+ for (cpu = rnp->grplo; cpu <= rnp->grphi; cpu++, mask >>= 1)
+ if ((mask & 0x1) && cpu != outgoingcpu)
+ cpumask_set_cpu(cpu, cm);
+ if (cpumask_weight(cm) == 0) {
+ cpumask_setall(cm);
+ for (cpu = rnp->grplo; cpu <= rnp->grphi; cpu++)
+ cpumask_clear_cpu(cpu, cm);
+ WARN_ON_ONCE(cpumask_weight(cm) == 0);
+ }
+ set_cpus_allowed_ptr(rnp->node_kthread_task, cm);
+ rcu_boost_kthread_setaffinity(rnp, cm);
+ free_cpumask_var(cm);
+}
+
+/*
+ * Spawn a per-rcu_node kthread, setting priority and affinity.
+ * Called during boot before online/offline can happen, or, if
+ * during runtime, with the main CPU-hotplug locks held. So only
+ * one of these can be executing at a time.
+ */
+static int __cpuinit rcu_spawn_one_node_kthread(struct rcu_state *rsp,
+ struct rcu_node *rnp)
+{
+ unsigned long flags;
+ int rnp_index = rnp - &rsp->node[0];
+ struct sched_param sp;
+ struct task_struct *t;
+
+ if (!rcu_kthreads_spawnable ||
+ rnp->qsmaskinit == 0)
+ return 0;
+ if (rnp->node_kthread_task == NULL) {
+ t = kthread_create(rcu_node_kthread, (void *)rnp,
+ "rcun%d", rnp_index);
+ if (IS_ERR(t))
+ return PTR_ERR(t);
+ raw_spin_lock_irqsave(&rnp->lock, flags);
+ rnp->node_kthread_task = t;
+ raw_spin_unlock_irqrestore(&rnp->lock, flags);
+ sp.sched_priority = 99;
+ sched_setscheduler_nocheck(t, SCHED_FIFO, &sp);
+ wake_up_process(t); /* get to TASK_INTERRUPTIBLE quickly. */
+ }
+ return rcu_spawn_one_boost_kthread(rsp, rnp, rnp_index);
+}
+
+/*
+ * Spawn all kthreads -- called as soon as the scheduler is running.
+ */
+static int __init rcu_spawn_kthreads(void)
+{
+ int cpu;
+ struct rcu_node *rnp;
+
+ rcu_kthreads_spawnable = 1;
+ for_each_possible_cpu(cpu) {
+ per_cpu(rcu_cpu_has_work, cpu) = 0;
+ if (cpu_online(cpu))
+ (void)rcu_spawn_one_cpu_kthread(cpu);
+ }
+ rnp = rcu_get_root(rcu_state);
+ (void)rcu_spawn_one_node_kthread(rcu_state, rnp);
+ if (NUM_RCU_NODES > 1) {
+ rcu_for_each_leaf_node(rcu_state, rnp)
+ (void)rcu_spawn_one_node_kthread(rcu_state, rnp);
+ }
+ return 0;
+}
+early_initcall(rcu_spawn_kthreads);
+
+static void __cpuinit rcu_prepare_kthreads(int cpu)
+{
+ struct rcu_data *rdp = per_cpu_ptr(rcu_state->rda, cpu);
+ struct rcu_node *rnp = rdp->mynode;
+
+ /* Fire up the incoming CPU's kthread and leaf rcu_node kthread. */
+ if (rcu_kthreads_spawnable) {
+ (void)rcu_spawn_one_cpu_kthread(cpu);
+ if (rnp->node_kthread_task == NULL)
+ (void)rcu_spawn_one_node_kthread(rcu_state, rnp);
+ }
}
#else /* #ifdef CONFIG_RCU_BOOST */
@@ -1315,23 +1701,32 @@ static void rcu_initiate_boost(struct rcu_node *rnp, unsigned long flags)
raw_spin_unlock_irqrestore(&rnp->lock, flags);
}
-static void rcu_boost_kthread_setaffinity(struct rcu_node *rnp,
- cpumask_var_t cm)
+static void invoke_rcu_callbacks_kthread(void)
{
+ WARN_ON_ONCE(1);
}
static void rcu_preempt_boost_start_gp(struct rcu_node *rnp)
{
}
-static int __cpuinit rcu_spawn_one_boost_kthread(struct rcu_state *rsp,
- struct rcu_node *rnp,
- int rnp_index)
+#ifdef CONFIG_HOTPLUG_CPU
+
+static void rcu_stop_cpu_kthread(int cpu)
+{
+}
+
+#endif /* #ifdef CONFIG_HOTPLUG_CPU */
+
+static void rcu_node_kthread_setaffinity(struct rcu_node *rnp, int outgoingcpu)
+{
+}
+
+static void rcu_cpu_kthread_setrt(int cpu, int to_rt)
{
- return 0;
}
-static void __cpuinit rcu_wake_one_boost_kthread(struct rcu_node *rnp)
+static void __cpuinit rcu_prepare_kthreads(int cpu)
{
}
@@ -1509,7 +1904,7 @@ static DEFINE_PER_CPU(unsigned long, rcu_dyntick_holdoff);
*
* Because it is not legal to invoke rcu_process_callbacks() with irqs
* disabled, we do one pass of force_quiescent_state(), then do a
- * invoke_rcu_cpu_kthread() to cause rcu_process_callbacks() to be invoked
+ * invoke_rcu_core() to cause rcu_process_callbacks() to be invoked
* later. The per-cpu rcu_dyntick_drain variable controls the sequencing.
*/
int rcu_needs_cpu(int cpu)
@@ -1560,7 +1955,7 @@ int rcu_needs_cpu(int cpu)
/* If RCU callbacks are still pending, RCU still needs this CPU. */
if (c)
- invoke_rcu_cpu_kthread();
+ invoke_rcu_core();
return c;
}
diff --git a/kernel/rcutree_trace.c b/kernel/rcutree_trace.c
index 9678cc3650f..4e144876dc6 100644
--- a/kernel/rcutree_trace.c
+++ b/kernel/rcutree_trace.c
@@ -46,6 +46,8 @@
#define RCU_TREE_NONCORE
#include "rcutree.h"
+#ifdef CONFIG_RCU_BOOST
+
DECLARE_PER_CPU(unsigned int, rcu_cpu_kthread_status);
DECLARE_PER_CPU(unsigned int, rcu_cpu_kthread_cpu);
DECLARE_PER_CPU(unsigned int, rcu_cpu_kthread_loops);
@@ -58,6 +60,8 @@ static char convert_kthread_status(unsigned int kthread_status)
return "SRWOY"[kthread_status];
}
+#endif /* #ifdef CONFIG_RCU_BOOST */
+
static void print_one_rcu_data(struct seq_file *m, struct rcu_data *rdp)
{
if (!rdp->beenonline)
@@ -76,7 +80,7 @@ static void print_one_rcu_data(struct seq_file *m, struct rcu_data *rdp)
rdp->dynticks_fqs);
#endif /* #ifdef CONFIG_NO_HZ */
seq_printf(m, " of=%lu ri=%lu", rdp->offline_fqs, rdp->resched_ipi);
- seq_printf(m, " ql=%ld qs=%c%c%c%c kt=%d/%c/%d ktl=%x b=%ld",
+ seq_printf(m, " ql=%ld qs=%c%c%c%c",
rdp->qlen,
".N"[rdp->nxttail[RCU_NEXT_READY_TAIL] !=
rdp->nxttail[RCU_NEXT_TAIL]],
@@ -84,13 +88,16 @@ static void print_one_rcu_data(struct seq_file *m, struct rcu_data *rdp)
rdp->nxttail[RCU_NEXT_READY_TAIL]],
".W"[rdp->nxttail[RCU_DONE_TAIL] !=
rdp->nxttail[RCU_WAIT_TAIL]],
- ".D"[&rdp->nxtlist != rdp->nxttail[RCU_DONE_TAIL]],
+ ".D"[&rdp->nxtlist != rdp->nxttail[RCU_DONE_TAIL]]);
+#ifdef CONFIG_RCU_BOOST
+ seq_printf(m, " kt=%d/%c/%d ktl=%x",
per_cpu(rcu_cpu_has_work, rdp->cpu),
convert_kthread_status(per_cpu(rcu_cpu_kthread_status,
rdp->cpu)),
per_cpu(rcu_cpu_kthread_cpu, rdp->cpu),
- per_cpu(rcu_cpu_kthread_loops, rdp->cpu) & 0xffff,
- rdp->blimit);
+ per_cpu(rcu_cpu_kthread_loops, rdp->cpu) & 0xffff);
+#endif /* #ifdef CONFIG_RCU_BOOST */
+ seq_printf(m, " b=%ld", rdp->blimit);
seq_printf(m, " ci=%lu co=%lu ca=%lu\n",
rdp->n_cbs_invoked, rdp->n_cbs_orphaned, rdp->n_cbs_adopted);
}
@@ -147,18 +154,21 @@ static void print_one_rcu_data_csv(struct seq_file *m, struct rcu_data *rdp)
rdp->dynticks_fqs);
#endif /* #ifdef CONFIG_NO_HZ */
seq_printf(m, ",%lu,%lu", rdp->offline_fqs, rdp->resched_ipi);
- seq_printf(m, ",%ld,\"%c%c%c%c\",%d,\"%c\",%ld", rdp->qlen,
+ seq_printf(m, ",%ld,\"%c%c%c%c\"", rdp->qlen,
".N"[rdp->nxttail[RCU_NEXT_READY_TAIL] !=
rdp->nxttail[RCU_NEXT_TAIL]],
".R"[rdp->nxttail[RCU_WAIT_TAIL] !=
rdp->nxttail[RCU_NEXT_READY_TAIL]],
".W"[rdp->nxttail[RCU_DONE_TAIL] !=
rdp->nxttail[RCU_WAIT_TAIL]],
- ".D"[&rdp->nxtlist != rdp->nxttail[RCU_DONE_TAIL]],
+ ".D"[&rdp->nxtlist != rdp->nxttail[RCU_DONE_TAIL]]);
+#ifdef CONFIG_RCU_BOOST
+ seq_printf(m, ",%d,\"%c\"",
per_cpu(rcu_cpu_has_work, rdp->cpu),
convert_kthread_status(per_cpu(rcu_cpu_kthread_status,
- rdp->cpu)),
- rdp->blimit);
+ rdp->cpu)));
+#endif /* #ifdef CONFIG_RCU_BOOST */
+ seq_printf(m, ",%ld", rdp->blimit);
seq_printf(m, ",%lu,%lu,%lu\n",
rdp->n_cbs_invoked, rdp->n_cbs_orphaned, rdp->n_cbs_adopted);
}
@@ -169,7 +179,11 @@ static int show_rcudata_csv(struct seq_file *m, void *unused)
#ifdef CONFIG_NO_HZ
seq_puts(m, "\"dt\",\"dt nesting\",\"dt NMI nesting\",\"df\",");
#endif /* #ifdef CONFIG_NO_HZ */
- seq_puts(m, "\"of\",\"ri\",\"ql\",\"b\",\"ci\",\"co\",\"ca\"\n");
+ seq_puts(m, "\"of\",\"ri\",\"ql\",\"qs\"");
+#ifdef CONFIG_RCU_BOOST
+ seq_puts(m, "\"kt\",\"ktl\"");
+#endif /* #ifdef CONFIG_RCU_BOOST */
+ seq_puts(m, ",\"b\",\"ci\",\"co\",\"ca\"\n");
#ifdef CONFIG_TREE_PREEMPT_RCU
seq_puts(m, "\"rcu_preempt:\"\n");
PRINT_RCU_DATA(rcu_preempt_data, print_one_rcu_data_csv, m);
diff --git a/kernel/sched.c b/kernel/sched.c
index 3f2e502d609..d08d110b897 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -2220,7 +2220,7 @@ void set_task_cpu(struct task_struct *p, unsigned int new_cpu)
if (task_cpu(p) != new_cpu) {
p->se.nr_migrations++;
- perf_sw_event(PERF_COUNT_SW_CPU_MIGRATIONS, 1, 1, NULL, 0);
+ perf_sw_event(PERF_COUNT_SW_CPU_MIGRATIONS, 1, NULL, 0);
}
__set_task_cpu(p, new_cpu);
diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c
index 88725c939e0..10d018212ba 100644
--- a/kernel/sched_rt.c
+++ b/kernel/sched_rt.c
@@ -1096,7 +1096,7 @@ static void check_preempt_curr_rt(struct rq *rq, struct task_struct *p, int flag
* to move current somewhere else, making room for our non-migratable
* task.
*/
- if (p->prio == rq->curr->prio && !need_resched())
+ if (p->prio == rq->curr->prio && !test_tsk_need_resched(rq->curr))
check_preempt_equal_prio(rq, p);
#endif
}
@@ -1239,6 +1239,10 @@ static int find_lowest_rq(struct task_struct *task)
int this_cpu = smp_processor_id();
int cpu = task_cpu(task);
+ /* Make sure the mask is initialized first */
+ if (unlikely(!lowest_mask))
+ return -1;
+
if (task->rt.nr_cpus_allowed == 1)
return -1; /* No other targets possible */
diff --git a/kernel/signal.c b/kernel/signal.c
index 86c32b884f8..ff767860332 100644
--- a/kernel/signal.c
+++ b/kernel/signal.c
@@ -2365,7 +2365,7 @@ int sigprocmask(int how, sigset_t *set, sigset_t *oldset)
/**
* sys_rt_sigprocmask - change the list of currently blocked signals
* @how: whether to add, remove, or set signals
- * @set: stores pending signals
+ * @nset: stores pending signals
* @oset: previous value of signal mask if non-null
* @sigsetsize: size of sigset_t type
*/
diff --git a/kernel/smp.c b/kernel/smp.c
index 73a19519355..fb67dfa8394 100644
--- a/kernel/smp.c
+++ b/kernel/smp.c
@@ -74,7 +74,7 @@ static struct notifier_block __cpuinitdata hotplug_cfd_notifier = {
.notifier_call = hotplug_cfd,
};
-static int __cpuinit init_call_single_data(void)
+void __init call_function_init(void)
{
void *cpu = (void *)(long)smp_processor_id();
int i;
@@ -88,10 +88,7 @@ static int __cpuinit init_call_single_data(void)
hotplug_cfd(&hotplug_cfd_notifier, CPU_UP_PREPARE, cpu);
register_cpu_notifier(&hotplug_cfd_notifier);
-
- return 0;
}
-early_initcall(init_call_single_data);
/*
* csd_lock/csd_unlock used to serialize access to per-cpu csd resources
diff --git a/kernel/softirq.c b/kernel/softirq.c
index 13960170cad..40cf63ddd4b 100644
--- a/kernel/softirq.c
+++ b/kernel/softirq.c
@@ -58,7 +58,7 @@ DEFINE_PER_CPU(struct task_struct *, ksoftirqd);
char *softirq_to_name[NR_SOFTIRQS] = {
"HI", "TIMER", "NET_TX", "NET_RX", "BLOCK", "BLOCK_IOPOLL",
- "TASKLET", "SCHED", "HRTIMER"
+ "TASKLET", "SCHED", "HRTIMER", "RCU"
};
/*
diff --git a/kernel/taskstats.c b/kernel/taskstats.c
index 9ffea360a77..fc0f2200541 100644
--- a/kernel/taskstats.c
+++ b/kernel/taskstats.c
@@ -285,16 +285,18 @@ ret:
static int add_del_listener(pid_t pid, const struct cpumask *mask, int isadd)
{
struct listener_list *listeners;
- struct listener *s, *tmp;
+ struct listener *s, *tmp, *s2;
unsigned int cpu;
if (!cpumask_subset(mask, cpu_possible_mask))
return -EINVAL;
+ s = NULL;
if (isadd == REGISTER) {
for_each_cpu(cpu, mask) {
- s = kmalloc_node(sizeof(struct listener), GFP_KERNEL,
- cpu_to_node(cpu));
+ if (!s)
+ s = kmalloc_node(sizeof(struct listener),
+ GFP_KERNEL, cpu_to_node(cpu));
if (!s)
goto cleanup;
s->pid = pid;
@@ -303,9 +305,16 @@ static int add_del_listener(pid_t pid, const struct cpumask *mask, int isadd)
listeners = &per_cpu(listener_array, cpu);
down_write(&listeners->sem);
+ list_for_each_entry_safe(s2, tmp, &listeners->list, list) {
+ if (s2->pid == pid)
+ goto next_cpu;
+ }
list_add(&s->list, &listeners->list);
+ s = NULL;
+next_cpu:
up_write(&listeners->sem);
}
+ kfree(s);
return 0;
}
diff --git a/kernel/time/alarmtimer.c b/kernel/time/alarmtimer.c
index 2d966244ea6..59f369f98a0 100644
--- a/kernel/time/alarmtimer.c
+++ b/kernel/time/alarmtimer.c
@@ -42,15 +42,75 @@ static struct alarm_base {
clockid_t base_clockid;
} alarm_bases[ALARM_NUMTYPE];
+/* freezer delta & lock used to handle clock_nanosleep triggered wakeups */
+static ktime_t freezer_delta;
+static DEFINE_SPINLOCK(freezer_delta_lock);
+
#ifdef CONFIG_RTC_CLASS
/* rtc timer and device for setting alarm wakeups at suspend */
static struct rtc_timer rtctimer;
static struct rtc_device *rtcdev;
-#endif
+static DEFINE_SPINLOCK(rtcdev_lock);
-/* freezer delta & lock used to handle clock_nanosleep triggered wakeups */
-static ktime_t freezer_delta;
-static DEFINE_SPINLOCK(freezer_delta_lock);
+/**
+ * has_wakealarm - check rtc device has wakealarm ability
+ * @dev: current device
+ * @name_ptr: name to be returned
+ *
+ * This helper function checks to see if the rtc device can wake
+ * from suspend.
+ */
+static int has_wakealarm(struct device *dev, void *name_ptr)
+{
+ struct rtc_device *candidate = to_rtc_device(dev);
+
+ if (!candidate->ops->set_alarm)
+ return 0;
+ if (!device_may_wakeup(candidate->dev.parent))
+ return 0;
+
+ *(const char **)name_ptr = dev_name(dev);
+ return 1;
+}
+
+/**
+ * alarmtimer_get_rtcdev - Return selected rtcdevice
+ *
+ * This function returns the rtc device to use for wakealarms.
+ * If one has not already been chosen, it checks to see if a
+ * functional rtc device is available.
+ */
+static struct rtc_device *alarmtimer_get_rtcdev(void)
+{
+ struct device *dev;
+ char *str;
+ unsigned long flags;
+ struct rtc_device *ret;
+
+ spin_lock_irqsave(&rtcdev_lock, flags);
+ if (!rtcdev) {
+ /* Find an rtc device and init the rtc_timer */
+ dev = class_find_device(rtc_class, NULL, &str, has_wakealarm);
+ /* If we have a device then str is valid. See has_wakealarm() */
+ if (dev) {
+ rtcdev = rtc_class_open(str);
+ /*
+ * Drop the reference we got in class_find_device,
+ * rtc_open takes its own.
+ */
+ put_device(dev);
+ rtc_timer_init(&rtctimer, NULL, NULL);
+ }
+ }
+ ret = rtcdev;
+ spin_unlock_irqrestore(&rtcdev_lock, flags);
+
+ return ret;
+}
+#else
+#define alarmtimer_get_rtcdev() (0)
+#define rtcdev (0)
+#endif
/**
@@ -166,6 +226,7 @@ static int alarmtimer_suspend(struct device *dev)
struct rtc_time tm;
ktime_t min, now;
unsigned long flags;
+ struct rtc_device *rtc;
int i;
spin_lock_irqsave(&freezer_delta_lock, flags);
@@ -173,8 +234,9 @@ static int alarmtimer_suspend(struct device *dev)
freezer_delta = ktime_set(0, 0);
spin_unlock_irqrestore(&freezer_delta_lock, flags);
+ rtc = rtcdev;
/* If we have no rtcdev, just return */
- if (!rtcdev)
+ if (!rtc)
return 0;
/* Find the soonest timer to expire*/
@@ -199,12 +261,12 @@ static int alarmtimer_suspend(struct device *dev)
WARN_ON(min.tv64 < NSEC_PER_SEC);
/* Setup an rtc timer to fire that far in the future */
- rtc_timer_cancel(rtcdev, &rtctimer);
- rtc_read_time(rtcdev, &tm);
+ rtc_timer_cancel(rtc, &rtctimer);
+ rtc_read_time(rtc, &tm);
now = rtc_tm_to_ktime(tm);
now = ktime_add(now, min);
- rtc_timer_start(rtcdev, &rtctimer, now, ktime_set(0, 0));
+ rtc_timer_start(rtc, &rtctimer, now, ktime_set(0, 0));
return 0;
}
@@ -322,6 +384,9 @@ static int alarm_clock_getres(const clockid_t which_clock, struct timespec *tp)
{
clockid_t baseid = alarm_bases[clock2alarm(which_clock)].base_clockid;
+ if (!alarmtimer_get_rtcdev())
+ return -ENOTSUPP;
+
return hrtimer_get_res(baseid, tp);
}
@@ -336,6 +401,9 @@ static int alarm_clock_get(clockid_t which_clock, struct timespec *tp)
{
struct alarm_base *base = &alarm_bases[clock2alarm(which_clock)];
+ if (!alarmtimer_get_rtcdev())
+ return -ENOTSUPP;
+
*tp = ktime_to_timespec(base->gettime());
return 0;
}
@@ -351,6 +419,9 @@ static int alarm_timer_create(struct k_itimer *new_timer)
enum alarmtimer_type type;
struct alarm_base *base;
+ if (!alarmtimer_get_rtcdev())
+ return -ENOTSUPP;
+
if (!capable(CAP_WAKE_ALARM))
return -EPERM;
@@ -385,6 +456,9 @@ static void alarm_timer_get(struct k_itimer *timr,
*/
static int alarm_timer_del(struct k_itimer *timr)
{
+ if (!rtcdev)
+ return -ENOTSUPP;
+
alarm_cancel(&timr->it.alarmtimer);
return 0;
}
@@ -402,6 +476,9 @@ static int alarm_timer_set(struct k_itimer *timr, int flags,
struct itimerspec *new_setting,
struct itimerspec *old_setting)
{
+ if (!rtcdev)
+ return -ENOTSUPP;
+
/* Save old values */
old_setting->it_interval =
ktime_to_timespec(timr->it.alarmtimer.period);
@@ -541,6 +618,9 @@ static int alarm_timer_nsleep(const clockid_t which_clock, int flags,
int ret = 0;
struct restart_block *restart;
+ if (!alarmtimer_get_rtcdev())
+ return -ENOTSUPP;
+
if (!capable(CAP_WAKE_ALARM))
return -EPERM;
@@ -638,65 +718,3 @@ static int __init alarmtimer_init(void)
}
device_initcall(alarmtimer_init);
-#ifdef CONFIG_RTC_CLASS
-/**
- * has_wakealarm - check rtc device has wakealarm ability
- * @dev: current device
- * @name_ptr: name to be returned
- *
- * This helper function checks to see if the rtc device can wake
- * from suspend.
- */
-static int __init has_wakealarm(struct device *dev, void *name_ptr)
-{
- struct rtc_device *candidate = to_rtc_device(dev);
-
- if (!candidate->ops->set_alarm)
- return 0;
- if (!device_may_wakeup(candidate->dev.parent))
- return 0;
-
- *(const char **)name_ptr = dev_name(dev);
- return 1;
-}
-
-/**
- * alarmtimer_init_late - Late initializing of alarmtimer code
- *
- * This function locates a rtc device to use for wakealarms.
- * Run as late_initcall to make sure rtc devices have been
- * registered.
- */
-static int __init alarmtimer_init_late(void)
-{
- struct device *dev;
- char *str;
-
- /* Find an rtc device and init the rtc_timer */
- dev = class_find_device(rtc_class, NULL, &str, has_wakealarm);
- /* If we have a device then str is valid. See has_wakealarm() */
- if (dev) {
- rtcdev = rtc_class_open(str);
- /*
- * Drop the reference we got in class_find_device,
- * rtc_open takes its own.
- */
- put_device(dev);
- }
- if (!rtcdev) {
- printk(KERN_WARNING "No RTC device found, ALARM timers will"
- " not wake from suspend");
- }
- rtc_timer_init(&rtctimer, NULL, NULL);
-
- return 0;
-}
-#else
-static int __init alarmtimer_init_late(void)
-{
- printk(KERN_WARNING "Kernel not built with RTC support, ALARM timers"
- " will not wake from suspend");
- return 0;
-}
-#endif
-late_initcall(alarmtimer_init_late);
diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c
index 1c95fd67732..e0980f0d9a0 100644
--- a/kernel/time/clocksource.c
+++ b/kernel/time/clocksource.c
@@ -185,7 +185,6 @@ static struct clocksource *watchdog;
static struct timer_list watchdog_timer;
static DECLARE_WORK(watchdog_work, clocksource_watchdog_work);
static DEFINE_SPINLOCK(watchdog_lock);
-static cycle_t watchdog_last;
static int watchdog_running;
static int clocksource_watchdog_kthread(void *data);
@@ -254,11 +253,6 @@ static void clocksource_watchdog(unsigned long data)
if (!watchdog_running)
goto out;
- wdnow = watchdog->read(watchdog);
- wd_nsec = clocksource_cyc2ns((wdnow - watchdog_last) & watchdog->mask,
- watchdog->mult, watchdog->shift);
- watchdog_last = wdnow;
-
list_for_each_entry(cs, &watchdog_list, wd_list) {
/* Clocksource already marked unstable? */
@@ -268,19 +262,28 @@ static void clocksource_watchdog(unsigned long data)
continue;
}
+ local_irq_disable();
csnow = cs->read(cs);
+ wdnow = watchdog->read(watchdog);
+ local_irq_enable();
/* Clocksource initialized ? */
if (!(cs->flags & CLOCK_SOURCE_WATCHDOG)) {
cs->flags |= CLOCK_SOURCE_WATCHDOG;
- cs->wd_last = csnow;
+ cs->wd_last = wdnow;
+ cs->cs_last = csnow;
continue;
}
- /* Check the deviation from the watchdog clocksource. */
- cs_nsec = clocksource_cyc2ns((csnow - cs->wd_last) &
+ wd_nsec = clocksource_cyc2ns((wdnow - cs->wd_last) & watchdog->mask,
+ watchdog->mult, watchdog->shift);
+
+ cs_nsec = clocksource_cyc2ns((csnow - cs->cs_last) &
cs->mask, cs->mult, cs->shift);
- cs->wd_last = csnow;
+ cs->cs_last = csnow;
+ cs->wd_last = wdnow;
+
+ /* Check the deviation from the watchdog clocksource. */
if (abs(cs_nsec - wd_nsec) > WATCHDOG_THRESHOLD) {
clocksource_unstable(cs, cs_nsec - wd_nsec);
continue;
@@ -318,7 +321,6 @@ static inline void clocksource_start_watchdog(void)
return;
init_timer(&watchdog_timer);
watchdog_timer.function = clocksource_watchdog;
- watchdog_last = watchdog->read(watchdog);
watchdog_timer.expires = jiffies + WATCHDOG_INTERVAL;
add_timer_on(&watchdog_timer, cpumask_first(cpu_online_mask));
watchdog_running = 1;
diff --git a/kernel/trace/trace_printk.c b/kernel/trace/trace_printk.c
index dff763b7baf..1f06468a10d 100644
--- a/kernel/trace/trace_printk.c
+++ b/kernel/trace/trace_printk.c
@@ -240,13 +240,10 @@ static const char **find_next(void *v, loff_t *pos)
const char **fmt = v;
int start_index;
- if (!fmt)
- fmt = __start___trace_bprintk_fmt + *pos;
-
start_index = __stop___trace_bprintk_fmt - __start___trace_bprintk_fmt;
if (*pos < start_index)
- return fmt;
+ return __start___trace_bprintk_fmt + *pos;
return find_next_mod_format(start_index, v, fmt, pos);
}
diff --git a/kernel/watchdog.c b/kernel/watchdog.c
index 3d0c56ad479..a933e3a0398 100644
--- a/kernel/watchdog.c
+++ b/kernel/watchdog.c
@@ -200,6 +200,8 @@ static int is_softlockup(unsigned long touch_ts)
}
#ifdef CONFIG_HARDLOCKUP_DETECTOR
+void __weak hw_nmi_watchdog_set_attr(struct perf_event_attr *wd_attr) { }
+
static struct perf_event_attr wd_hw_attr = {
.type = PERF_TYPE_HARDWARE,
.config = PERF_COUNT_HW_CPU_CYCLES,
@@ -209,7 +211,7 @@ static struct perf_event_attr wd_hw_attr = {
};
/* Callback function for perf event subsystem */
-static void watchdog_overflow_callback(struct perf_event *event, int nmi,
+static void watchdog_overflow_callback(struct perf_event *event,
struct perf_sample_data *data,
struct pt_regs *regs)
{
@@ -368,10 +370,12 @@ static int watchdog_nmi_enable(int cpu)
if (event != NULL)
goto out_enable;
- /* Try to register using hardware perf events */
wd_attr = &wd_hw_attr;
wd_attr->sample_period = hw_nmi_get_sample_period(watchdog_thresh);
- event = perf_event_create_kernel_counter(wd_attr, cpu, NULL, watchdog_overflow_callback);
+ hw_nmi_watchdog_set_attr(wd_attr);
+
+ /* Try to register using hardware perf events */
+ event = perf_event_create_kernel_counter(wd_attr, cpu, NULL, watchdog_overflow_callback, NULL);
if (!IS_ERR(event)) {
printk(KERN_INFO "NMI watchdog enabled, takes one hw-pmu counter.\n");
goto out_save;
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2025-04-12 10:16:24 +0000