#define RB_EVNT_HDR_SIZE (offsetof(struct ring_buffer_event, array))
#define RB_ALIGNMENT 4U
#define RB_MAX_SMALL_DATA (RB_ALIGNMENT * RINGBUF_TYPE_DATA_TYPE_LEN_MAX)
+#define RB_EVNT_MIN_SIZE 8U /* two 32bit words */
/* define RINGBUF_TYPE_DATA for 'case RINGBUF_TYPE_DATA:' */
#define RINGBUF_TYPE_DATA 0 ... RINGBUF_TYPE_DATA_TYPE_LEN_MAX
#define BUF_PAGE_SIZE (PAGE_SIZE - BUF_PAGE_HDR_SIZE)
+/* Max payload is BUF_PAGE_SIZE - header (8bytes) */
+#define BUF_MAX_DATA_SIZE (BUF_PAGE_SIZE - (sizeof(u32) * 2))
+
+/* Max number of timestamps that can fit on a page */
+#define RB_TIMESTAMPS_PER_PAGE (BUF_PAGE_SIZE / RB_LEN_TIME_STAMP)
+
int ring_buffer_print_page_header(struct trace_seq *s)
{
struct buffer_data_page field;
unsigned long overrun;
unsigned long read;
local_t entries;
+ local_t committing;
+ local_t commits;
u64 write_stamp;
u64 read_stamp;
atomic_t record_disabled;
atomic_t record_disabled;
cpumask_var_t cpumask;
+ struct lock_class_key *reader_lock_key;
+
struct mutex mutex;
struct ring_buffer_per_cpu **buffers;
/* Up this if you want to test the TIME_EXTENTS and normalization */
#define DEBUG_SHIFT 0
+static inline u64 rb_time_stamp(struct ring_buffer *buffer, int cpu)
+{
+ /* shift to debug/test normalization and TIME_EXTENTS */
+ return buffer->clock() << DEBUG_SHIFT;
+}
+
u64 ring_buffer_time_stamp(struct ring_buffer *buffer, int cpu)
{
u64 time;
preempt_disable_notrace();
- /* shift to debug/test normalization and TIME_EXTENTS */
- time = buffer->clock() << DEBUG_SHIFT;
+ time = rb_time_stamp(buffer, cpu);
preempt_enable_no_resched_notrace();
return time;
cpu_buffer->cpu = cpu;
cpu_buffer->buffer = buffer;
spin_lock_init(&cpu_buffer->reader_lock);
+ lockdep_set_class(&cpu_buffer->reader_lock, buffer->reader_lock_key);
cpu_buffer->lock = (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED;
INIT_LIST_HEAD(&cpu_buffer->pages);
* when the buffer wraps. If this flag is not set, the buffer will
* drop data when the tail hits the head.
*/
-struct ring_buffer *ring_buffer_alloc(unsigned long size, unsigned flags)
+struct ring_buffer *__ring_buffer_alloc(unsigned long size, unsigned flags,
+ struct lock_class_key *key)
{
struct ring_buffer *buffer;
int bsize;
buffer->pages = DIV_ROUND_UP(size, BUF_PAGE_SIZE);
buffer->flags = flags;
buffer->clock = trace_clock_local;
+ buffer->reader_lock_key = key;
/* need at least two pages */
if (buffer->pages == 1)
kfree(buffer);
return NULL;
}
-EXPORT_SYMBOL_GPL(ring_buffer_alloc);
+EXPORT_SYMBOL_GPL(__ring_buffer_alloc);
/**
* ring_buffer_free - free a ring buffer.
{
unsigned long addr = (unsigned long)event;
- return (addr & ~PAGE_MASK) - (PAGE_SIZE - BUF_PAGE_SIZE);
+ return (addr & ~PAGE_MASK) - BUF_PAGE_HDR_SIZE;
}
-static int
-rb_is_commit(struct ring_buffer_per_cpu *cpu_buffer,
- struct ring_buffer_event *event)
+static inline int
+rb_event_is_commit(struct ring_buffer_per_cpu *cpu_buffer,
+ struct ring_buffer_event *event)
{
unsigned long addr = (unsigned long)event;
unsigned long index;
rb_commit_index(cpu_buffer) == index;
}
-static void
-rb_set_commit_event(struct ring_buffer_per_cpu *cpu_buffer,
- struct ring_buffer_event *event)
-{
- unsigned long addr = (unsigned long)event;
- unsigned long index;
-
- index = rb_event_index(event);
- addr &= PAGE_MASK;
-
- while (cpu_buffer->commit_page->page != (void *)addr) {
- if (RB_WARN_ON(cpu_buffer,
- cpu_buffer->commit_page == cpu_buffer->tail_page))
- return;
- cpu_buffer->commit_page->page->commit =
- cpu_buffer->commit_page->write;
- rb_inc_page(cpu_buffer, &cpu_buffer->commit_page);
- cpu_buffer->write_stamp =
- cpu_buffer->commit_page->page->time_stamp;
- }
-
- /* Now set the commit to the event's index */
- local_set(&cpu_buffer->commit_page->page->commit, index);
-}
-
static void
rb_set_commit_to_write(struct ring_buffer_per_cpu *cpu_buffer)
{
return length;
}
-static struct ring_buffer_event *
-__rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer,
- unsigned type, unsigned long length, u64 *ts)
+static inline void
+rb_reset_tail(struct ring_buffer_per_cpu *cpu_buffer,
+ struct buffer_page *tail_page,
+ unsigned long tail, unsigned long length)
{
- struct buffer_page *tail_page, *head_page, *reader_page, *commit_page;
- unsigned long tail, write;
- struct ring_buffer *buffer = cpu_buffer->buffer;
struct ring_buffer_event *event;
- unsigned long flags;
- bool lock_taken = false;
- commit_page = cpu_buffer->commit_page;
- /* we just need to protect against interrupts */
- barrier();
- tail_page = cpu_buffer->tail_page;
- write = local_add_return(length, &tail_page->write);
- tail = write - length;
+ /*
+ * Only the event that crossed the page boundary
+ * must fill the old tail_page with padding.
+ */
+ if (tail >= BUF_PAGE_SIZE) {
+ local_sub(length, &tail_page->write);
+ return;
+ }
- /* See if we shot pass the end of this buffer page */
- if (write > BUF_PAGE_SIZE) {
- struct buffer_page *next_page = tail_page;
+ event = __rb_page_index(tail_page, tail);
- local_irq_save(flags);
- /*
- * Since the write to the buffer is still not
- * fully lockless, we must be careful with NMIs.
- * The locks in the writers are taken when a write
- * crosses to a new page. The locks protect against
- * races with the readers (this will soon be fixed
- * with a lockless solution).
- *
- * Because we can not protect against NMIs, and we
- * want to keep traces reentrant, we need to manage
- * what happens when we are in an NMI.
- *
- * NMIs can happen after we take the lock.
- * If we are in an NMI, only take the lock
- * if it is not already taken. Otherwise
- * simply fail.
- */
- if (unlikely(in_nmi())) {
- if (!__raw_spin_trylock(&cpu_buffer->lock)) {
- cpu_buffer->nmi_dropped++;
- goto out_reset;
- }
- } else
- __raw_spin_lock(&cpu_buffer->lock);
+ /*
+ * If this event is bigger than the minimum size, then
+ * we need to be careful that we don't subtract the
+ * write counter enough to allow another writer to slip
+ * in on this page.
+ * We put in a discarded commit instead, to make sure
+ * that this space is not used again.
+ *
+ * If we are less than the minimum size, we don't need to
+ * worry about it.
+ */
+ if (tail > (BUF_PAGE_SIZE - RB_EVNT_MIN_SIZE)) {
+ /* No room for any events */
- lock_taken = true;
+ /* Mark the rest of the page with padding */
+ rb_event_set_padding(event);
- rb_inc_page(cpu_buffer, &next_page);
+ /* Set the write back to the previous setting */
+ local_sub(length, &tail_page->write);
+ return;
+ }
- head_page = cpu_buffer->head_page;
- reader_page = cpu_buffer->reader_page;
+ /* Put in a discarded event */
+ event->array[0] = (BUF_PAGE_SIZE - tail) - RB_EVNT_HDR_SIZE;
+ event->type_len = RINGBUF_TYPE_PADDING;
+ /* time delta must be non zero */
+ event->time_delta = 1;
+ /* Account for this as an entry */
+ local_inc(&tail_page->entries);
+ local_inc(&cpu_buffer->entries);
- /* we grabbed the lock before incrementing */
- if (RB_WARN_ON(cpu_buffer, next_page == reader_page))
- goto out_reset;
+ /* Set write to end of buffer */
+ length = (tail + length) - BUF_PAGE_SIZE;
+ local_sub(length, &tail_page->write);
+}
- /*
- * If for some reason, we had an interrupt storm that made
- * it all the way around the buffer, bail, and warn
- * about it.
- */
- if (unlikely(next_page == commit_page)) {
- cpu_buffer->commit_overrun++;
+static struct ring_buffer_event *
+rb_move_tail(struct ring_buffer_per_cpu *cpu_buffer,
+ unsigned long length, unsigned long tail,
+ struct buffer_page *commit_page,
+ struct buffer_page *tail_page, u64 *ts)
+{
+ struct buffer_page *next_page, *head_page, *reader_page;
+ struct ring_buffer *buffer = cpu_buffer->buffer;
+ bool lock_taken = false;
+ unsigned long flags;
+
+ next_page = tail_page;
+
+ local_irq_save(flags);
+ /*
+ * Since the write to the buffer is still not
+ * fully lockless, we must be careful with NMIs.
+ * The locks in the writers are taken when a write
+ * crosses to a new page. The locks protect against
+ * races with the readers (this will soon be fixed
+ * with a lockless solution).
+ *
+ * Because we can not protect against NMIs, and we
+ * want to keep traces reentrant, we need to manage
+ * what happens when we are in an NMI.
+ *
+ * NMIs can happen after we take the lock.
+ * If we are in an NMI, only take the lock
+ * if it is not already taken. Otherwise
+ * simply fail.
+ */
+ if (unlikely(in_nmi())) {
+ if (!__raw_spin_trylock(&cpu_buffer->lock)) {
+ cpu_buffer->nmi_dropped++;
goto out_reset;
}
+ } else
+ __raw_spin_lock(&cpu_buffer->lock);
- if (next_page == head_page) {
- if (!(buffer->flags & RB_FL_OVERWRITE))
- goto out_reset;
+ lock_taken = true;
- /* tail_page has not moved yet? */
- if (tail_page == cpu_buffer->tail_page) {
- /* count overflows */
- cpu_buffer->overrun +=
- local_read(&head_page->entries);
+ rb_inc_page(cpu_buffer, &next_page);
- rb_inc_page(cpu_buffer, &head_page);
- cpu_buffer->head_page = head_page;
- cpu_buffer->head_page->read = 0;
- }
- }
+ head_page = cpu_buffer->head_page;
+ reader_page = cpu_buffer->reader_page;
- /*
- * If the tail page is still the same as what we think
- * it is, then it is up to us to update the tail
- * pointer.
- */
+ /* we grabbed the lock before incrementing */
+ if (RB_WARN_ON(cpu_buffer, next_page == reader_page))
+ goto out_reset;
+
+ /*
+ * If for some reason, we had an interrupt storm that made
+ * it all the way around the buffer, bail, and warn
+ * about it.
+ */
+ if (unlikely(next_page == commit_page)) {
+ cpu_buffer->commit_overrun++;
+ goto out_reset;
+ }
+
+ if (next_page == head_page) {
+ if (!(buffer->flags & RB_FL_OVERWRITE))
+ goto out_reset;
+
+ /* tail_page has not moved yet? */
if (tail_page == cpu_buffer->tail_page) {
- local_set(&next_page->write, 0);
- local_set(&next_page->entries, 0);
- local_set(&next_page->page->commit, 0);
- cpu_buffer->tail_page = next_page;
-
- /* reread the time stamp */
- *ts = ring_buffer_time_stamp(buffer, cpu_buffer->cpu);
- cpu_buffer->tail_page->page->time_stamp = *ts;
- }
+ /* count overflows */
+ cpu_buffer->overrun +=
+ local_read(&head_page->entries);
- /*
- * The actual tail page has moved forward.
- */
- if (tail < BUF_PAGE_SIZE) {
- /* Mark the rest of the page with padding */
- event = __rb_page_index(tail_page, tail);
- rb_event_set_padding(event);
+ rb_inc_page(cpu_buffer, &head_page);
+ cpu_buffer->head_page = head_page;
+ cpu_buffer->head_page->read = 0;
}
+ }
+
+ /*
+ * If the tail page is still the same as what we think
+ * it is, then it is up to us to update the tail
+ * pointer.
+ */
+ if (tail_page == cpu_buffer->tail_page) {
+ local_set(&next_page->write, 0);
+ local_set(&next_page->entries, 0);
+ local_set(&next_page->page->commit, 0);
+ cpu_buffer->tail_page = next_page;
+
+ /* reread the time stamp */
+ *ts = rb_time_stamp(buffer, cpu_buffer->cpu);
+ cpu_buffer->tail_page->page->time_stamp = *ts;
+ }
- if (tail <= BUF_PAGE_SIZE)
- /* Set the write back to the previous setting */
- local_set(&tail_page->write, tail);
+ rb_reset_tail(cpu_buffer, tail_page, tail, length);
- /*
- * If this was a commit entry that failed,
- * increment that too
- */
- if (tail_page == cpu_buffer->commit_page &&
- tail == rb_commit_index(cpu_buffer)) {
- rb_set_commit_to_write(cpu_buffer);
- }
+ __raw_spin_unlock(&cpu_buffer->lock);
+ local_irq_restore(flags);
+ /* fail and let the caller try again */
+ return ERR_PTR(-EAGAIN);
+
+ out_reset:
+ /* reset write */
+ rb_reset_tail(cpu_buffer, tail_page, tail, length);
+
+ if (likely(lock_taken))
__raw_spin_unlock(&cpu_buffer->lock);
- local_irq_restore(flags);
+ local_irq_restore(flags);
+ return NULL;
+}
- /* fail and let the caller try again */
- return ERR_PTR(-EAGAIN);
- }
+static struct ring_buffer_event *
+__rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer,
+ unsigned type, unsigned long length, u64 *ts)
+{
+ struct buffer_page *tail_page, *commit_page;
+ struct ring_buffer_event *event;
+ unsigned long tail, write;
+
+ commit_page = cpu_buffer->commit_page;
+ /* we just need to protect against interrupts */
+ barrier();
+ tail_page = cpu_buffer->tail_page;
+ write = local_add_return(length, &tail_page->write);
+ tail = write - length;
+
+ /* See if we shot pass the end of this buffer page */
+ if (write > BUF_PAGE_SIZE)
+ return rb_move_tail(cpu_buffer, length, tail,
+ commit_page, tail_page, ts);
/* We reserved something on the buffer */
local_inc(&tail_page->entries);
/*
- * If this is a commit and the tail is zero, then update
- * this page's time stamp.
+ * If this is the first commit on the page, then update
+ * its timestamp.
*/
- if (!tail && rb_is_commit(cpu_buffer, event))
- cpu_buffer->commit_page->page->time_stamp = *ts;
+ if (!tail)
+ tail_page->page->time_stamp = *ts;
return event;
+}
- out_reset:
- /* reset write */
- if (tail <= BUF_PAGE_SIZE)
- local_set(&tail_page->write, tail);
+static inline int
+rb_try_to_discard(struct ring_buffer_per_cpu *cpu_buffer,
+ struct ring_buffer_event *event)
+{
+ unsigned long new_index, old_index;
+ struct buffer_page *bpage;
+ unsigned long index;
+ unsigned long addr;
- if (likely(lock_taken))
- __raw_spin_unlock(&cpu_buffer->lock);
- local_irq_restore(flags);
- return NULL;
+ new_index = rb_event_index(event);
+ old_index = new_index + rb_event_length(event);
+ addr = (unsigned long)event;
+ addr &= PAGE_MASK;
+
+ bpage = cpu_buffer->tail_page;
+
+ if (bpage->page == (void *)addr && rb_page_write(bpage) == old_index) {
+ /*
+ * This is on the tail page. It is possible that
+ * a write could come in and move the tail page
+ * and write to the next page. That is fine
+ * because we just shorten what is on this page.
+ */
+ index = local_cmpxchg(&bpage->write, old_index, new_index);
+ if (index == old_index)
+ return 1;
+ }
+
+ /* could not discard */
+ return 0;
}
static int
return -EAGAIN;
/* Only a commited time event can update the write stamp */
- if (rb_is_commit(cpu_buffer, event)) {
+ if (rb_event_is_commit(cpu_buffer, event)) {
/*
- * If this is the first on the page, then we need to
- * update the page itself, and just put in a zero.
+ * If this is the first on the page, then it was
+ * updated with the page itself. Try to discard it
+ * and if we can't just make it zero.
*/
if (rb_event_index(event)) {
event->time_delta = *delta & TS_MASK;
event->array[0] = *delta >> TS_SHIFT;
} else {
- cpu_buffer->commit_page->page->time_stamp = *ts;
- event->time_delta = 0;
- event->array[0] = 0;
+ /* try to discard, since we do not need this */
+ if (!rb_try_to_discard(cpu_buffer, event)) {
+ /* nope, just zero it */
+ event->time_delta = 0;
+ event->array[0] = 0;
+ }
}
cpu_buffer->write_stamp = *ts;
/* let the caller know this was the commit */
ret = 1;
} else {
- /* Darn, this is just wasted space */
- event->time_delta = 0;
- event->array[0] = 0;
+ /* Try to discard the event */
+ if (!rb_try_to_discard(cpu_buffer, event)) {
+ /* Darn, this is just wasted space */
+ event->time_delta = 0;
+ event->array[0] = 0;
+ }
ret = 0;
}
return ret;
}
+static void rb_start_commit(struct ring_buffer_per_cpu *cpu_buffer)
+{
+ local_inc(&cpu_buffer->committing);
+ local_inc(&cpu_buffer->commits);
+}
+
+static void rb_end_commit(struct ring_buffer_per_cpu *cpu_buffer)
+{
+ unsigned long commits;
+
+ if (RB_WARN_ON(cpu_buffer,
+ !local_read(&cpu_buffer->committing)))
+ return;
+
+ again:
+ commits = local_read(&cpu_buffer->commits);
+ /* synchronize with interrupts */
+ barrier();
+ if (local_read(&cpu_buffer->committing) == 1)
+ rb_set_commit_to_write(cpu_buffer);
+
+ local_dec(&cpu_buffer->committing);
+
+ /* synchronize with interrupts */
+ barrier();
+
+ /*
+ * Need to account for interrupts coming in between the
+ * updating of the commit page and the clearing of the
+ * committing counter.
+ */
+ if (unlikely(local_read(&cpu_buffer->commits) != commits) &&
+ !local_read(&cpu_buffer->committing)) {
+ local_inc(&cpu_buffer->committing);
+ goto again;
+ }
+}
+
static struct ring_buffer_event *
rb_reserve_next_event(struct ring_buffer_per_cpu *cpu_buffer,
- unsigned type, unsigned long length)
+ unsigned long length)
{
struct ring_buffer_event *event;
- u64 ts, delta;
+ u64 ts, delta = 0;
int commit = 0;
int nr_loops = 0;
+ rb_start_commit(cpu_buffer);
+
+ length = rb_calculate_event_length(length);
again:
/*
* We allow for interrupts to reenter here and do a trace.
* Bail!
*/
if (RB_WARN_ON(cpu_buffer, ++nr_loops > 1000))
- return NULL;
+ goto out_fail;
- ts = ring_buffer_time_stamp(cpu_buffer->buffer, cpu_buffer->cpu);
+ ts = rb_time_stamp(cpu_buffer->buffer, cpu_buffer->cpu);
/*
* Only the first commit can update the timestamp.
* also be made. But only the entry that did the actual
* commit will be something other than zero.
*/
- if (cpu_buffer->tail_page == cpu_buffer->commit_page &&
- rb_page_write(cpu_buffer->tail_page) ==
- rb_commit_index(cpu_buffer)) {
+ if (likely(cpu_buffer->tail_page == cpu_buffer->commit_page &&
+ rb_page_write(cpu_buffer->tail_page) ==
+ rb_commit_index(cpu_buffer))) {
+ u64 diff;
- delta = ts - cpu_buffer->write_stamp;
+ diff = ts - cpu_buffer->write_stamp;
- /* make sure this delta is calculated here */
+ /* make sure this diff is calculated here */
barrier();
/* Did the write stamp get updated already? */
if (unlikely(ts < cpu_buffer->write_stamp))
- delta = 0;
+ goto get_event;
- if (test_time_stamp(delta)) {
+ delta = diff;
+ if (unlikely(test_time_stamp(delta))) {
commit = rb_add_time_stamp(cpu_buffer, &ts, &delta);
-
if (commit == -EBUSY)
- return NULL;
+ goto out_fail;
if (commit == -EAGAIN)
goto again;
RB_WARN_ON(cpu_buffer, commit < 0);
}
- } else
- /* Non commits have zero deltas */
- delta = 0;
+ }
- event = __rb_reserve_next(cpu_buffer, type, length, &ts);
- if (PTR_ERR(event) == -EAGAIN)
+ get_event:
+ event = __rb_reserve_next(cpu_buffer, 0, length, &ts);
+ if (unlikely(PTR_ERR(event) == -EAGAIN))
goto again;
- if (!event) {
- if (unlikely(commit))
- /*
- * Ouch! We needed a timestamp and it was commited. But
- * we didn't get our event reserved.
- */
- rb_set_commit_to_write(cpu_buffer);
- return NULL;
- }
+ if (!event)
+ goto out_fail;
- /*
- * If the timestamp was commited, make the commit our entry
- * now so that we will update it when needed.
- */
- if (commit)
- rb_set_commit_event(cpu_buffer, event);
- else if (!rb_is_commit(cpu_buffer, event))
+ if (!rb_event_is_commit(cpu_buffer, event))
delta = 0;
event->time_delta = delta;
return event;
+
+ out_fail:
+ rb_end_commit(cpu_buffer);
+ return NULL;
}
#define TRACE_RECURSIVE_DEPTH 16
if (atomic_read(&cpu_buffer->record_disabled))
goto out;
- length = rb_calculate_event_length(length);
- if (length > BUF_PAGE_SIZE)
+ if (length > BUF_MAX_DATA_SIZE)
goto out;
- event = rb_reserve_next_event(cpu_buffer, 0, length);
+ event = rb_reserve_next_event(cpu_buffer, length);
if (!event)
goto out;
{
local_inc(&cpu_buffer->entries);
- /* Only process further if we own the commit */
- if (!rb_is_commit(cpu_buffer, event))
- return;
-
- cpu_buffer->write_stamp += event->time_delta;
+ /*
+ * The event first in the commit queue updates the
+ * time stamp.
+ */
+ if (rb_event_is_commit(cpu_buffer, event))
+ cpu_buffer->write_stamp += event->time_delta;
- rb_set_commit_to_write(cpu_buffer);
+ rb_end_commit(cpu_buffer);
}
/**
struct ring_buffer_event *event)
{
struct ring_buffer_per_cpu *cpu_buffer;
- unsigned long new_index, old_index;
- struct buffer_page *bpage;
- unsigned long index;
- unsigned long addr;
int cpu;
/* The event is discarded regardless */
rb_event_discard(event);
+ cpu = smp_processor_id();
+ cpu_buffer = buffer->buffers[cpu];
+
/*
* This must only be called if the event has not been
* committed yet. Thus we can assume that preemption
* is still disabled.
*/
- RB_WARN_ON(buffer, !preempt_count());
-
- cpu = smp_processor_id();
- cpu_buffer = buffer->buffers[cpu];
-
- new_index = rb_event_index(event);
- old_index = new_index + rb_event_length(event);
- addr = (unsigned long)event;
- addr &= PAGE_MASK;
-
- bpage = cpu_buffer->tail_page;
+ RB_WARN_ON(buffer, !local_read(&cpu_buffer->committing));
- if (bpage == (void *)addr && rb_page_write(bpage) == old_index) {
- /*
- * This is on the tail page. It is possible that
- * a write could come in and move the tail page
- * and write to the next page. That is fine
- * because we just shorten what is on this page.
- */
- index = local_cmpxchg(&bpage->write, old_index, new_index);
- if (index == old_index)
- goto out;
- }
+ if (!rb_try_to_discard(cpu_buffer, event))
+ goto out;
/*
* The commit is still visible by the reader, so we
*/
local_inc(&cpu_buffer->entries);
out:
- /*
- * If a write came in and pushed the tail page
- * we still need to update the commit pointer
- * if we were the commit.
- */
- if (rb_is_commit(cpu_buffer, event))
- rb_set_commit_to_write(cpu_buffer);
+ rb_end_commit(cpu_buffer);
trace_recursive_unlock();
{
struct ring_buffer_per_cpu *cpu_buffer;
struct ring_buffer_event *event;
- unsigned long event_length;
void *body;
int ret = -EBUSY;
int cpu, resched;
if (atomic_read(&cpu_buffer->record_disabled))
goto out;
- event_length = rb_calculate_event_length(length);
- event = rb_reserve_next_event(cpu_buffer, 0, event_length);
+ if (length > BUF_MAX_DATA_SIZE)
+ goto out;
+
+ event = rb_reserve_next_event(cpu_buffer, length);
if (!event)
goto out;
* Check if we are at the end of the buffer.
*/
if (iter->head >= rb_page_size(iter->head_page)) {
- if (RB_WARN_ON(buffer,
- iter->head_page == cpu_buffer->commit_page))
+ /* discarded commits can make the page empty */
+ if (iter->head_page == cpu_buffer->commit_page)
return;
rb_inc_iter(iter);
return;
/*
* We repeat when a timestamp is encountered. It is possible
* to get multiple timestamps from an interrupt entering just
- * as one timestamp is about to be written. The max times
- * that this can happen is the number of nested interrupts we
- * can have. Nesting 10 deep of interrupts is clearly
- * an anomaly.
+ * as one timestamp is about to be written, or from discarded
+ * commits. The most that we can have is the number on a single page.
*/
- if (RB_WARN_ON(cpu_buffer, ++nr_loops > 10))
+ if (RB_WARN_ON(cpu_buffer, ++nr_loops > RB_TIMESTAMPS_PER_PAGE))
return NULL;
reader = rb_get_reader_page(cpu_buffer);
again:
/*
- * We repeat when a timestamp is encountered. It is possible
- * to get multiple timestamps from an interrupt entering just
- * as one timestamp is about to be written. The max times
- * that this can happen is the number of nested interrupts we
- * can have. Nesting 10 deep of interrupts is clearly
- * an anomaly.
+ * We repeat when a timestamp is encountered.
+ * We can get multiple timestamps by nested interrupts or also
+ * if filtering is on (discarding commits). Since discarding
+ * commits can be frequent we can get a lot of timestamps.
+ * But we limit them by not adding timestamps if they begin
+ * at the start of a page.
*/
- if (RB_WARN_ON(cpu_buffer, ++nr_loops > 10))
+ if (RB_WARN_ON(cpu_buffer, ++nr_loops > RB_TIMESTAMPS_PER_PAGE))
return NULL;
if (rb_per_cpu_empty(cpu_buffer))
cpu_buffer->overrun = 0;
cpu_buffer->read = 0;
local_set(&cpu_buffer->entries, 0);
+ local_set(&cpu_buffer->committing, 0);
+ local_set(&cpu_buffer->commits, 0);
cpu_buffer->write_stamp = 0;
cpu_buffer->read_stamp = 0;
if (!cpumask_test_cpu(cpu, buffer->cpumask))
return;
+ atomic_inc(&cpu_buffer->record_disabled);
+
spin_lock_irqsave(&cpu_buffer->reader_lock, flags);
__raw_spin_lock(&cpu_buffer->lock);
__raw_spin_unlock(&cpu_buffer->lock);
spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags);
+
+ atomic_dec(&cpu_buffer->record_disabled);
}
EXPORT_SYMBOL_GPL(ring_buffer_reset_cpu);
}
EXPORT_SYMBOL_GPL(ring_buffer_swap_cpu);
-static void rb_remove_entries(struct ring_buffer_per_cpu *cpu_buffer,
- struct buffer_data_page *bpage,
- unsigned int offset)
-{
- struct ring_buffer_event *event;
- unsigned long head;
-
- __raw_spin_lock(&cpu_buffer->lock);
- for (head = offset; head < local_read(&bpage->commit);
- head += rb_event_length(event)) {
-
- event = __rb_data_page_index(bpage, head);
- if (RB_WARN_ON(cpu_buffer, rb_null_event(event)))
- return;
- /* Only count data entries */
- if (event->type_len > RINGBUF_TYPE_DATA_TYPE_LEN_MAX)
- continue;
- cpu_buffer->read++;
- }
- __raw_spin_unlock(&cpu_buffer->lock);
-}
-
/**
* ring_buffer_alloc_read_page - allocate a page to read from buffer
* @buffer: the buffer to allocate for.
/* we copied everything to the beginning */
read = 0;
} else {
+ /* update the entry counter */
+ cpu_buffer->read += local_read(&reader->entries);
+
/* swap the pages */
rb_init_page(bpage);
bpage = reader->page;
local_set(&reader->entries, 0);
reader->read = 0;
*data_page = bpage;
-
- /* update the entry counter */
- rb_remove_entries(cpu_buffer, bpage, read);
}
ret = read;
switch (action) {
case CPU_UP_PREPARE:
case CPU_UP_PREPARE_FROZEN:
- if (cpu_isset(cpu, *buffer->cpumask))
+ if (cpumask_test_cpu(cpu, buffer->cpumask))
return NOTIFY_OK;
buffer->buffers[cpu] =
return NOTIFY_OK;
}
smp_wmb();
- cpu_set(cpu, *buffer->cpumask);
+ cpumask_set_cpu(cpu, buffer->cpumask);
break;
case CPU_DOWN_PREPARE:
case CPU_DOWN_PREPARE_FROZEN: