ring-buffer: move page indexes into page headers

Remove the global head and tail indexes and move them into the
page header. Each page will now keep track of where the last
write and read was made. We also rename the head and tail to read
and write for better clarification.

This patch is needed for future enhancements to move the ring buffer
to a lockless solution.

Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

1 files changed, 41 insertions, 34 deletions

diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
index 6b8dac02364f..09d4f0d879a7 100644
--- a/kernel/trace/ring_buffer.c
+++ b/kernel/trace/ring_buffer.c
@@ -117,6 +117,8 @@ void *ring_buffer_event_data(struct ring_buffer_event *event)
 struct buffer_page {
         u64              time_stamp;    /* page time stamp */
         unsigned         size;          /* size of page data */
+        unsigned         write;         /* index for next write */
+        unsigned         read;          /* index for next read */
         struct list_head list;          /* list of free pages */
         void *page;                     /* Actual data page */
 };
@@ -153,11 +155,8 @@ struct ring_buffer_per_cpu {
         spinlock_t                      lock;
         struct lock_class_key           lock_key;
         struct list_head                pages;
-        unsigned long                   head;   /* read from head */
-        unsigned long                   tail;   /* write to tail */
-        unsigned long                   reader;
-        struct buffer_page              *head_page;
-        struct buffer_page              *tail_page;
+        struct buffer_page              *head_page;     /* read from head */
+        struct buffer_page              *tail_page;     /* write to tail */
         struct buffer_page              *reader_page;
         unsigned long                   overrun;
         unsigned long                   entries;
@@ -566,10 +565,11 @@ int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size)
 
 static inline int rb_per_cpu_empty(struct ring_buffer_per_cpu *cpu_buffer)
 {
-        return (cpu_buffer->reader == cpu_buffer->reader_page->size &&
+        return cpu_buffer->reader_page->read == cpu_buffer->reader_page->size &&
                 (cpu_buffer->tail_page == cpu_buffer->reader_page ||
                  (cpu_buffer->tail_page == cpu_buffer->head_page &&
-                  cpu_buffer->head == cpu_buffer->tail)));
+                  cpu_buffer->head_page->read ==
+                  cpu_buffer->tail_page->write));
 }
 
 static inline int rb_null_event(struct ring_buffer_event *event)
@@ -577,7 +577,7 @@ static inline int rb_null_event(struct ring_buffer_event *event)
         return event->type == RINGBUF_TYPE_PADDING;
 }
 
-static inline void *rb_page_index(struct buffer_page *page, unsigned index)
+static inline void *__rb_page_index(struct buffer_page *page, unsigned index)
 {
         return page->page + index;
 }
@@ -585,15 +585,21 @@ static inline void *rb_page_index(struct buffer_page *page, unsigned index)
 static inline struct ring_buffer_event *
 rb_reader_event(struct ring_buffer_per_cpu *cpu_buffer)
 {
-        return rb_page_index(cpu_buffer->reader_page,
-                             cpu_buffer->reader);
+        return __rb_page_index(cpu_buffer->reader_page,
+                               cpu_buffer->reader_page->read);
+}
+
+static inline struct ring_buffer_event *
+rb_head_event(struct ring_buffer_per_cpu *cpu_buffer)
+{
+        return __rb_page_index(cpu_buffer->head_page,
+                               cpu_buffer->head_page->read);
 }
 
 static inline struct ring_buffer_event *
 rb_iter_head_event(struct ring_buffer_iter *iter)
 {
-        return rb_page_index(iter->head_page,
-                             iter->head);
+        return __rb_page_index(iter->head_page, iter->head);
 }
 
 /*
@@ -610,7 +616,7 @@ static void rb_update_overflow(struct ring_buffer_per_cpu *cpu_buffer)
         for (head = 0; head < rb_head_size(cpu_buffer);
              head += rb_event_length(event)) {
 
-                event = rb_page_index(cpu_buffer->head_page, head);
+                event = __rb_page_index(cpu_buffer->head_page, head);
                 BUG_ON(rb_null_event(event));
                 /* Only count data entries */
                 if (event->type != RINGBUF_TYPE_DATA)
@@ -640,13 +646,13 @@ rb_add_stamp(struct ring_buffer_per_cpu *cpu_buffer, u64 *ts)
 
 static void rb_reset_head_page(struct ring_buffer_per_cpu *cpu_buffer)
 {
-        cpu_buffer->head = 0;
+        cpu_buffer->head_page->read = 0;
 }
 
 static void rb_reset_reader_page(struct ring_buffer_per_cpu *cpu_buffer)
 {
         cpu_buffer->read_stamp = cpu_buffer->reader_page->time_stamp;
-        cpu_buffer->reader = 0;
+        cpu_buffer->reader_page->read = 0;
 }
 
 static inline void rb_inc_iter(struct ring_buffer_iter *iter)
@@ -743,9 +749,8 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer,
         struct ring_buffer *buffer = cpu_buffer->buffer;
         struct ring_buffer_event *event;
 
-        /* No locking needed for tail page */
         tail_page = cpu_buffer->tail_page;
-        tail = cpu_buffer->tail;
+        tail = cpu_buffer->tail_page->write;
 
         if (tail + length > BUF_PAGE_SIZE) {
                 struct buffer_page *next_page = tail_page;
@@ -774,7 +779,7 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer,
                 }
 
                 if (tail != BUF_PAGE_SIZE) {
-                        event = rb_page_index(tail_page, tail);
+                        event = __rb_page_index(tail_page, tail);
                         /* page padding */
                         event->type = RINGBUF_TYPE_PADDING;
                 }
@@ -784,14 +789,14 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer,
                 tail_page->size = 0;
                 tail = 0;
                 cpu_buffer->tail_page = tail_page;
-                cpu_buffer->tail = tail;
+                cpu_buffer->tail_page->write = tail;
                 rb_add_stamp(cpu_buffer, ts);
                 spin_unlock(&cpu_buffer->lock);
         }
 
         BUG_ON(tail + length > BUF_PAGE_SIZE);
 
-        event = rb_page_index(tail_page, tail);
+        event = __rb_page_index(tail_page, tail);
         rb_update_event(event, type, length);
 
         return event;
@@ -823,12 +828,12 @@ rb_add_time_stamp(struct ring_buffer_per_cpu *cpu_buffer,
                 return -1;
 
         /* check to see if we went to the next page */
-        if (cpu_buffer->tail) {
+        if (cpu_buffer->tail_page->write) {
                 /* Still on same page, update timestamp */
                 event->time_delta = *delta & TS_MASK;
                 event->array[0] = *delta >> TS_SHIFT;
                 /* commit the time event */
-                cpu_buffer->tail +=
+                cpu_buffer->tail_page->write +=
                         rb_event_length(event);
                 cpu_buffer->write_stamp = *ts;
                 *delta = 0;
@@ -846,7 +851,7 @@ rb_reserve_next_event(struct ring_buffer_per_cpu *cpu_buffer,