[sysprof] lib: Add full barrier before writing data_tail

[Ray Strode <halfline src gnome org>]

commit f313ca4daefa8aee7837f545a4b43795bd692789
Author: Ray Strode <rstrode redhat com>
Date:   Tue Aug 23 16:53:39 2016 -0400

    lib: Add full barrier before writing data_tail
    
    The kernel says here 
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/kernel/events/ring_buffer.c?id=7a1dcf6ad#n61
 :
    
         * Since the mmap() consumer (userspace) can run on a different CPU:
         *
         *   kernel                             user
         *
         *   if (LOAD ->data_tail) {            LOAD ->data_head
         *                      (A)             smp_rmb()       (C)
         *      STORE $data                     LOAD $data
         *      smp_wmb()       (B)             smp_mb()        (D)
         *      STORE ->data_head               STORE ->data_tail
         *   }
         *
         * Where A pairs with D, and B pairs with C.
         *
         * In our case (A) is a control dependency that separates the load of
         * the ->data_tail and the stores of $data. In case ->data_tail
         * indicates there is no room in the buffer to store $data we do not.
         *
         * D needs to be a full barrier since it separates the data READ
         * from the tail WRITE.
         *
         * For B a WMB is sufficient since it separates two WRITEs, and for C
         * an RMB is sufficient since it separates two READs.
         *
         * See perf_output_begin().
    
    So I'm pretty sure we need a full barrier before writing out data_tail.

 lib/sp-perf-counter.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)
---
diff --git a/lib/sp-perf-counter.c b/lib/sp-perf-counter.c
index bd7315b..d41bc9e 100644
--- a/lib/sp-perf-counter.c
+++ b/lib/sp-perf-counter.c
@@ -291,6 +291,8 @@ sp_perf_counter_flush (SpPerfCounter     *self,
     }
 
   info->tail = tail;
+
+  atomic_thread_fence (memory_order_seq_cst);
   info->map->data_tail = tail;
 }