docs/atomics: update atomic_read/set comparison with Linux

Recently Linux did a mass conversion of its atomic_read/set calls
so that they at least are READ/WRITE_ONCE. See Linux's commit
62e8a325 ("atomic, arch: Audit atomic_{read,set}()"). It seems though
that their documentation hasn't been updated to reflect this.

The appended updates our documentation to reflect the change, which
means there is effectively no difference between our atomic_read/set
and the current Linux implementation.

While at it, fix the statement that a barrier is implied by
atomic_read/set, which is incorrect. Volatile/atomic semantics prevent
transformations pertaining the variable they apply to; this, however,
has no effect on surrounding statements like barriers do. For more
details on this, see:
  https://gcc.gnu.org/onlinedocs/gcc/Volatiles.html

Signed-off-by: Emilio G. Cota <cota@braap.org>
Message-Id: <1464120374-8950-2-git-send-email-cota@braap.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

1 files changed, 13 insertions, 3 deletions

diff --git a/docs/atomics.txt b/docs/atomics.txt
index bba771e..67a27ad 100644
--- a/docs/atomics.txt
+++ b/docs/atomics.txt
@@ -326,9 +326,19 @@ and memory barriers, and the equivalents in QEMU:
   use a boxed atomic_t type; atomic operations in QEMU are polymorphic
   and use normal C types.
 
-- atomic_read and atomic_set in Linux give no guarantee at all;
-  atomic_read and atomic_set in QEMU include a compiler barrier
-  (similar to the READ_ONCE/WRITE_ONCE macros in Linux).
+- Originally, atomic_read and atomic_set in Linux gave no guarantee
+  at all. Linux 4.1 updated them to implement volatile
+  semantics via ACCESS_ONCE (or the more recent READ/WRITE_ONCE).
+
+  QEMU's atomic_read/set implement, if the compiler supports it, C11
+  atomic relaxed semantics, and volatile semantics otherwise.
+  Both semantics prevent the compiler from doing certain transformations;
+  the difference is that atomic accesses are guaranteed to be atomic,
+  while volatile accesses aren't. Thus, in the volatile case we just cross
+  our fingers hoping that the compiler will generate atomic accesses,
+  since we assume the variables passed are machine-word sized and
+  properly aligned.
+  No barriers are implied by atomic_read/set in either Linux or QEMU.
 
 - most atomic read-modify-write operations in Linux return void;
   in QEMU, all of them return the old value of the variable.