path: root/docs/devel/atomics.txt

CPUs perform independent memory operations effectively in random order.
but this can be a problem for CPU-CPU interaction (including interactions
between QEMU and the guest).  Multi-threaded programs use various tools
to instruct the compiler and the CPU to restrict the order to something
that is consistent with the expectations of the programmer.

The most basic tool is locking.  Mutexes, condition variables and
semaphores are used in QEMU, and should be the default approach to
synchronization.  Anything else is considerably harder, but it's
also justified more often than one would like.  The two tools that
are provided by qemu/atomic.h are memory barriers and atomic operations.

Macros defined by qemu/atomic.h fall in three camps:

- compiler barriers: barrier();

- weak atomic access and manual memory barriers: atomic_read(),
  atomic_set(), smp_rmb(), smp_wmb(), smp_mb(), smp_mb_acquire(),
  smp_mb_release(), smp_read_barrier_depends();

- sequentially consistent atomic access: everything else.


COMPILER MEMORY BARRIER
=======================

barrier() prevents the compiler from moving the memory accesses either
side of it to the other side.  The compiler barrier has no direct effect
on the CPU, which may then reorder things however it wishes.

barrier() is mostly used within qemu/atomic.h itself.  On some
architectures, CPU guarantees are strong enough that blocking compiler
optimizations already ensures the correct order of execution.  In this
case, qemu/atomic.h will reduce stronger memory barriers to simple
compiler barriers.

Still, barrier() can be useful when writing code that can be interrupted
by signal handlers.


SEQUENTIALLY CONSISTENT ATOMIC ACCESS
=====================================

Most of the operations in the qemu/atomic.h header ensure *sequential
consistency*, where "the result of any execution is the same as if the
operations of all the processors were executed in some sequential order,
and the operations of each individual processor appear in this sequence
in the order specified by its program".

qemu/atomic.h provides the following set of atomic read-modify-write
operations:

    void atomic_inc(ptr)
    void atomic_dec(ptr)
    void atomic_add(ptr, val)
    void atomic_sub(ptr, val)
    void atomic_and(ptr, val)
    void atomic_or(ptr, val)

    typeof(*ptr) atomic_fetch_inc(ptr)
    typeof(*ptr) atomic_fetch_dec(ptr)
    typeof(*ptr) atomic_fetch_add(ptr, val)
    typeof(*ptr) atomic_fetch_sub(ptr, val)
    typeof(*ptr) atomic_fetch_and(ptr, val)
    typeof(*ptr) atomic_fetch_or(ptr, val)
    typeof(*ptr) atomic_fetch_xor(ptr, val)
    typeof(*ptr) atomic_fetch_inc_nonzero(ptr)
    typeof(*ptr) atomic_xchg(ptr, val)
    typeof(*ptr) atomic_cmpxchg(ptr, old, new)

all of which return the old value of *ptr.  These operations are
polymorphic; they operate on any type that is as wide as a pointer.

Similar operations return the new value of *ptr:

    typeof(*ptr) atomic_inc_fetch(ptr)
    typeof(*ptr) atomic_dec_fetch(ptr)
    typeof(*ptr) atomic_add_fetch(ptr, val)
    typeof(*ptr) atomic_sub_fetch(ptr, val)
    typeof(*ptr) atomic_and_fetch(ptr, val)
    typeof(*ptr) atomic_or_fetch(ptr, val)
    typeof(*ptr) atomic_xor_fetch(ptr, val)

Sequentially consistent loads and stores can be done using:

    atomic_fetch_add(ptr, 0) for loads
    atomic_xchg(ptr, val) for stores

However, they are quite expensive on some platforms, notably POWER and
ARM.  Therefore, qemu/atomic.h provides two primitives with slightly
weaker constraints:

    typeof(*ptr) atomic_mb_read(ptr)
    void         atomic_mb_set(ptr, val)

The semantics of these primitives map to Java volatile variables,
and are strongly related to memory barriers as used in the Linux
kernel (see below).

As long as you use atomic_mb_read and atomic_mb_set, accesses cannot
be reordered with each other, and it is also not possible to reorder
"normal" accesses around them.

However, and this is the important difference between
atomic_mb_read/atomic_mb_set and sequential consistency, it is important
for both threads to access the same volatile variable.  It is not the
case that everything visible to thread A when it writes volatile field f
becomes visible to thread B after it reads volatile field g. The store
and load have to "match" (i.e., be performed on the same volatile
field) to achieve the right semantics.


These operations operate on any type that is as wide as an int or smaller.


WEAK ATOMIC ACCESS AND MANUAL MEMORY BARRIERS
=============================================

Compared to sequentially consistent atomic access, programming with
weaker consistency models can be considerably more complicated.
In general, if the algorithm you are writing includes both writes
and reads on the same side, it is generally simpler to use sequentially
consistent primitives.

When using this model, variables are accessed with atomic_read() and
atomic_set(), and restrictions to the ordering of accesses is enforced
using the memory barrier macros: smp_rmb(), smp_wmb(), smp_mb(),
smp_mb_acquire(), smp_mb_release(), smp_read_barrier_depends().

atomic_read() and atomic_set() prevents the compiler from using
optimizations that might otherwise optimize accesses out of existence
on the one hand, or that might create unsolicited accesses on the other.
In general this should not have any effect, because the same compiler
barriers are already implied by memory barriers.  However, it is useful
to do so, because it tells readers which variables are shared with
other threads, and which are local to the current thread or protected
by other, more mundane means.

Memory barriers control the order of references to shared memory.
They come in six kinds:

- smp_rmb() guarantees that all the LOAD operations specified before
  the barrier will appear to happen before all the LOAD operations
  specified after the barrier with respect to the other components of
  the system.

  In other words, smp_rmb() puts a partial ordering on loads, but is not
  required to have any effect on stores.

- smp_wmb() guarantees that all the STORE operations specified before
  the barrier will appear to happen before all the STORE operations
  specified after the barrier with respect to the other components of
  the system.

  In other words, smp_wmb() puts a partial ordering on stores, but is not
  required to have any effect on loads.

- smp_mb_acquire() guarantees that all the LOAD operations specified before
  the barrier will appear to happen before all the LOAD or STORE operations
  specified after the barrier with respect to the other components of
  the system.

- smp_mb_release() guarantees that all the STORE operations specified *after*
  the barrier will appear to happen after all the LOAD or STORE operations
  specified *before* the barrier with respect to the other components of
  the system.

- smp_mb() guarantees that all the LOAD and STORE operations specified
  before the barrier will appear to happen before all the LOAD and
  STORE operations specified after the barrier with respect to the other
  components of the system.

  smp_mb() puts a partial ordering on both loads and stores.  It is
  stronger than both a read and a write memory barrier; it implies both
  smp_mb_acquire() and smp_mb_release(), but it also prevents STOREs
  coming before the barrier from overtaking LOADs coming after the
  barrier and vice versa.

- smp_read_barrier_depends() is a weaker kind of read barrier.  On
  most processors, whenever two loads are performed such that the
  second depends on the result of the first (e.g., the first load
  retrieves the address to which the second load will be directed),
  the processor will guarantee that the first LOAD will appear to happen
  before the second with respect to the other components of the system.
  However, this is not always true---for example, it was not true on
  Alpha processors.  Whenever this kind of access happens to shared
  memory (that is not protected by a lock), a read barrier is needed,
  and smp_read_barrier_depends() can be used instead of smp_rmb().

  Note that the first load really has to have a _data_ dependency and not
  a control dependency.  If the address for the second load is dependent
  on the first load, but the dependency is through a conditional rather
  than actually loading the address itself, then it's a _control_
  dependency and a full read barrier or better is required.


This is the set of barriers that is required *between* two atomic_read()
and atomic_set() operations to achieve sequential consistency:

                    |               2nd operation                   |
                    |-----------------------------------------------|
     1st operation  | (after last)   | atomic_read | atomic_set     |
     ---------------+----------------+-------------+----------------|
     (before first) |                | none        | smp_mb_release |
     ---------------+----------------+-------------+----------------|
     atomic_read    | smp_mb_acquire | smp_rmb     | **             |
     ---------------+----------------+-------------+----------------|
     atomic_set     | none           | smp_mb()*** | smp_wmb()      |
     ---------------+----------------+-------------+----------------|

       * Or smp_read_barrier_depends().

      ** This requires a load-store barrier.  This is achieved by
         either smp_mb_acquire() or smp_mb_release().

     *** This requires a store-load barrier.  On most machines, the only
         way to achieve this is a full barrier.


You can see that the two possible definitions of atomic_mb_read()
and atomic_mb_set() are the following:

    1) atomic_mb_read(p)   = atomic_read(p); smp_mb_acquire()
       atomic_mb_set(p, v) = smp_mb_release(); atomic_set(p, v); smp_mb()

    2) atomic_mb_read(p)   = smp_mb() atomic_read(p); smp_mb_acquire()
       atomic_mb_set(p, v) = smp_mb_release(); atomic_set(p, v);

Usually the former is used, because smp_mb() is expensive and a program
normally has more reads than writes.  Therefore it makes more sense to
make atomic_mb_set() the more expensive operation.

There are two common cases in which atomic_mb_read and atomic_mb_set
generate too many memory barriers, and thus it can be useful to manually
place barriers instead:

- when a data structure has one thread that is always a writer
  and one thread that is always a reader, manual placement of
  memory barriers makes the write side faster.  Furthermore,
  correctness is easy to check for in this case using the "pairing"
  trick that is explained below:

     thread 1                                thread 1
     -------------------------               ------------------------
     (other writes)
                                             smp_mb_release()
     atomic_mb_set(&a, x)                    atomic_set(&a, x)
                                             smp_wmb()
     atomic_mb_set(&b, y)                    atomic_set(&b, y)

                                       =>
     thread 2                                thread 2
     -------------------------               ------------------------
     y = atomic_mb_read(&b)                  y = atomic_read(&b)
                                             smp_rmb()
     x = atomic_mb_read(&a)                  x = atomic_read(&a)
                                             smp_mb_acquire()

  Note that the barrier between the stores in thread 1, and between
  the loads in thread 2, has been optimized here to a write or a
  read memory barrier respectively.  On some architectures, notably
  ARMv7, smp_mb_acquire and smp_mb_release are just as expensive as
  smp_mb, but smp_rmb and/or smp_wmb are more efficient.

- sometimes, a thread is accessing many variables that are otherwise
  unrelated to each other (for example because, apart from the current
  thread, exactly one other thread will read or write each of these
  variables).  In this case, it is possible to "hoist" the implicit
  barriers provided by atomic_mb_read() and atomic_mb_set() outside
  a loop.  For example, the above definition atomic_mb_read() gives
  the following transformation:

     n = 0;                                  n = 0;
     for (i = 0; i < 10; i++)          =>    for (i = 0; i < 10; i++)
       n += atomic_mb_read(&a[i]);             n += atomic_read(&a[i]);
                                             smp_mb_acquire();

  Similarly, atomic_mb_set() can be transformed as follows:
  smp_mb():

                                             smp_mb_release();
     for (i = 0; i < 10; i++)          =>    for (i = 0; i < 10; i++)
       atomic_mb_set(&a[i], false);            atomic_set(&a[i], false);
                                             smp_mb();


The two tricks can be combined.  In this case, splitting a loop in
two lets you hoist the barriers out of the loops _and_ eliminate the
expensive smp_mb():

                                             smp_mb_release();
     for (i = 0; i < 10; i++) {        =>    for (i = 0; i < 10; i++)
       atomic_mb_set(&a[i], false);            atomic_set(&a[i], false);
       atomic_mb_set(&b[i], false);          smb_wmb();
     }                                       for (i = 0; i < 10; i++)
                                               atomic_set(&a[i], false);
                                             smp_mb();

  The other thread can still use atomic_mb_read()/atomic_mb_set()


Memory barrier pairing
----------------------

A useful rule of thumb is that memory barriers should always, or almost
always, be paired with another barrier.  In the case of QEMU, however,
note that the other barrier may actually be in a driver that runs in
the guest!

For the purposes of pairing, smp_read_barrier_depends() and smp_rmb()
both count as read barriers.  A read barrier shall pair with a write
barrier or a full barrier; a write barrier shall pair with a read
barrier or a full barrier.  A full barrier can pair with anything.
For example:

        thread 1             thread 2
        ===============      ===============
        a = 1;
        smp_wmb();
        b = 2;               x = b;
                             smp_rmb();
                             y = a;

Note that the "writing" thread is accessing the variables in the
opposite order as the "reading" thread.  This is expected: stores
before the write barrier will normally match the loads after the
read barrier, and vice versa.  The same is true for more than 2
access and for data dependency barriers:

        thread 1             thread 2
        ===============      ===============
        b[2] = 1;
        smp_wmb();
        x->i = 2;
        smp_wmb();
        a = x;               x = a;
                             smp_read_barrier_depends();
                             y = x->i;
                             smp_read_barrier_depends();
                             z = b[y];

smp_wmb() also pairs with atomic_mb_read() and smp_mb_acquire().
and smp_rmb() also pairs with atomic_mb_set() and smp_mb_release().


COMPARISON WITH LINUX KERNEL MEMORY BARRIERS
============================================

Here is a list of differences between Linux kernel atomic operations
and memory barriers, and the equivalents in QEMU:

- atomic operations in Linux are always on a 32-bit int type and
  use a boxed atomic_t type; atomic operations in QEMU are polymorphic
  and use normal C types.

- 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.

- atomic read-modify-write operations in Linux are of three kinds:

         atomic_OP          returns void
         atomic_OP_return   returns new value of the variable
         atomic_fetch_OP    returns the old value of the variable
         atomic_cmpxchg     returns the old value of the variable

  In QEMU, the second kind does not exist.  Currently Linux has
  atomic_fetch_or only.  QEMU provides and, or, inc, dec, add, sub.

- different atomic read-modify-write operations in Linux imply
  a different set of memory barriers; in QEMU, all of them enforce
  sequential consistency, which means they imply full memory barriers
  before and after the operation.

- Linux does not have an equivalent of atomic_mb_set().  In particular,
  note that smp_store_mb() is a little weaker than atomic_mb_set().
  atomic_mb_read() compiles to the same instructions as Linux's
  smp_load_acquire(), but this should be treated as an implementation
  detail.  QEMU does have atomic_load_acquire() and atomic_store_release()
  macros, but for now they are only used within atomic.h.  This may
  change in the future.


SOURCES
=======

* Documentation/memory-barriers.txt from the Linux kernel

* "The JSR-133 Cookbook for Compiler Writers", available at
  http://g.oswego.edu/dl/jmm/cookbook.html