sh: convert /proc/cpu/aligmnent, /proc/cpu/kernel_alignment to seq_file
[safe/jmp/linux-2.6] / include / linux / ring_buffer.h
index 9e6052b..5fcc31e 100644 (file)
@@ -1,6 +1,7 @@
 #ifndef _LINUX_RING_BUFFER_H
 #define _LINUX_RING_BUFFER_H
 
+#include <linux/kmemcheck.h>
 #include <linux/mm.h>
 #include <linux/seq_file.h>
 
@@ -11,17 +12,24 @@ struct ring_buffer_iter;
  * Don't refer to this struct directly, use functions below.
  */
 struct ring_buffer_event {
-       u32             type:2, len:3, time_delta:27;
+       kmemcheck_bitfield_begin(bitfield);
+       u32             type_len:5, time_delta:27;
+       kmemcheck_bitfield_end(bitfield);
+
        u32             array[];
 };
 
 /**
  * enum ring_buffer_type - internal ring buffer types
  *
- * @RINGBUF_TYPE_PADDING:      Left over page padding
- *                              array is ignored
- *                              size is variable depending on how much
+ * @RINGBUF_TYPE_PADDING:      Left over page padding or discarded event
+ *                              If time_delta is 0:
+ *                               array is ignored
+ *                               size is variable depending on how much
  *                               padding is needed
+ *                              If time_delta is non zero:
+ *                               array[0] holds the actual length
+ *                               size = 4 + length (bytes)
  *
  * @RINGBUF_TYPE_TIME_EXTEND:  Extend the time delta
  *                              array[0] = time delta (28 .. 59)
@@ -32,22 +40,23 @@ struct ring_buffer_event {
  *                              array[1..2] = tv_sec
  *                              size = 16 bytes
  *
- * @RINGBUF_TYPE_DATA:         Data record
- *                              If len is zero:
+ * <= @RINGBUF_TYPE_DATA_TYPE_LEN_MAX:
+ *                             Data record
+ *                              If type_len is zero:
  *                               array[0] holds the actual length
  *                               array[1..(length+3)/4] holds data
- *                               size = 4 + 4 + length (bytes)
+ *                               size = 4 + length (bytes)
  *                              else
- *                               length = len << 2
+ *                               length = type_len << 2
  *                               array[0..(length+3)/4-1] holds data
  *                               size = 4 + length (bytes)
  */
 enum ring_buffer_type {
+       RINGBUF_TYPE_DATA_TYPE_LEN_MAX = 28,
        RINGBUF_TYPE_PADDING,
        RINGBUF_TYPE_TIME_EXTEND,
        /* FIXME: RINGBUF_TYPE_TIME_STAMP not implemented */
        RINGBUF_TYPE_TIME_STAMP,
-       RINGBUF_TYPE_DATA,
 };
 
 unsigned ring_buffer_event_length(struct ring_buffer_event *event);
@@ -66,10 +75,39 @@ ring_buffer_event_time_delta(struct ring_buffer_event *event)
 }
 
 /*
+ * ring_buffer_discard_commit will remove an event that has not
+ *   ben committed yet. If this is used, then ring_buffer_unlock_commit
+ *   must not be called on the discarded event. This function
+ *   will try to remove the event from the ring buffer completely
+ *   if another event has not been written after it.
+ *
+ * Example use:
+ *
+ *  if (some_condition)
+ *    ring_buffer_discard_commit(buffer, event);
+ *  else
+ *    ring_buffer_unlock_commit(buffer, event);
+ */
+void ring_buffer_discard_commit(struct ring_buffer *buffer,
+                               struct ring_buffer_event *event);
+
+/*
  * size is in bytes for each per CPU buffer.
  */
 struct ring_buffer *
-ring_buffer_alloc(unsigned long size, unsigned flags);
+__ring_buffer_alloc(unsigned long size, unsigned flags, struct lock_class_key *key);
+
+/*
+ * Because the ring buffer is generic, if other users of the ring buffer get
+ * traced by ftrace, it can produce lockdep warnings. We need to keep each
+ * ring buffer's lock class separate.
+ */
+#define ring_buffer_alloc(size, flags)                 \
+({                                                     \
+       static struct lock_class_key __key;             \
+       __ring_buffer_alloc((size), (flags), &__key);   \
+})
+
 void ring_buffer_free(struct ring_buffer *buffer);
 
 int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size);
@@ -102,8 +140,17 @@ unsigned long ring_buffer_size(struct ring_buffer *buffer);
 void ring_buffer_reset_cpu(struct ring_buffer *buffer, int cpu);
 void ring_buffer_reset(struct ring_buffer *buffer);
 
+#ifdef CONFIG_RING_BUFFER_ALLOW_SWAP
 int ring_buffer_swap_cpu(struct ring_buffer *buffer_a,
                         struct ring_buffer *buffer_b, int cpu);
+#else
+static inline int
+ring_buffer_swap_cpu(struct ring_buffer *buffer_a,
+                    struct ring_buffer *buffer_b, int cpu)
+{
+       return -ENODEV;
+}
+#endif
 
 int ring_buffer_empty(struct ring_buffer *buffer);
 int ring_buffer_empty_cpu(struct ring_buffer *buffer, int cpu);
@@ -117,6 +164,7 @@ unsigned long ring_buffer_entries(struct ring_buffer *buffer);
 unsigned long ring_buffer_overruns(struct ring_buffer *buffer);
 unsigned long ring_buffer_entries_cpu(struct ring_buffer *buffer, int cpu);
 unsigned long ring_buffer_overrun_cpu(struct ring_buffer *buffer, int cpu);
+unsigned long ring_buffer_commit_overrun_cpu(struct ring_buffer *buffer, int cpu);
 
 u64 ring_buffer_time_stamp(struct ring_buffer *buffer, int cpu);
 void ring_buffer_normalize_time_stamp(struct ring_buffer *buffer,
@@ -132,6 +180,11 @@ void ring_buffer_free_read_page(struct ring_buffer *buffer, void *data);
 int ring_buffer_read_page(struct ring_buffer *buffer, void **data_page,
                          size_t len, int cpu, int full);
 
+struct trace_seq;
+
+int ring_buffer_print_entry_header(struct trace_seq *s);
+int ring_buffer_print_page_header(struct trace_seq *s);
+
 enum ring_buffer_flags {
        RB_FL_OVERWRITE         = 1 << 0,
 };