[LAD] Atomic Operations

From: Gabriel M. Beddingfield <gabriel@email-addr-hidden>
Date: Sat Dec 12 2009 - 19:47:19 EET

Recently on this list, Paul referred to an "atomic integer
swap" and an "atomic pointer swap." This was a new concept
to me (and possibly others), and this e-mail shares what
I've learned.

If you access a variable from multiple threads -- even a
built-in variable like 'int', it is important to control
access to the variable with a mutex, semaphore, or an atomic
operation.

Alexander Sandler, on his blog, wrote a couple of good
articles on the subject:

     "Do you need a mutex to protect an int?"
     http://www.alexonlinux.com/do-you-need-mutex-to-protect-int

     "Multithreaded simple data type access and atomic
     variables"
     http://www.alexonlinux.com/multithreaded-simple-data-type-access-and-atomic-variables

The first article contains code that calculates a wrong
answer on multiprocessor machines. I've attached a similar
example that will even fail on a single-processor machine.

There is a wealth of reading material on using Mutexes and
Semaphores. However, information on atomic operations
appears to be sparse and hard-to-follow. So, here's what
I've found:

     + At the moment, there is no built-in support in
       C/C++ for atomic operations. You will need to use
       a library, compiler extension, or write your own
       in assembly code.

     + The GCC compiler has the built-in __sync_*()
       functions[1] that provide atomic operations.
       Note that the attached example is using this.

     + glib provides the g_atomic_*() functions[2].

     + Qt 4 has the q_atomic_*() functions.[3] While
       they are accessible, they are /not/ a part of
       their stable, public API.

     + The next version of ISO C++ (code name c++0x)
       is expected to have support for atomic operations
       (E.g. the std::atomic<T> template) and memory
       barriers. It may even require that all built-in
       types be atomic.

     + In the x86 instruction set, these are usually
       implemented using the 'LOCK' instruction prefix.[5]

When using atomic operations, perhaps the best advice I
found is near the end of Sandler's second article:

     "When using atomic variables, some extra
     precautions have to be taken.... There is nothing
     that prevents you from incrementing value of the
     atomic variable with __sync_fetch_and_add() as I
     just demonstrated and later in the code doing same
     thing with regular ++ operator.

     "To address this problem, I strongly suggest
     wrapping around atomic functions and variables with
     either ADT in C or C++ class."[4]

Peace,
Gabriel

[1] http://gcc.gnu.org/onlinedocs/gcc-4.1.0/gcc/Atomic-Builtins.html
[2] http://www.gtk.org/api/2.6/glib/glib-Atomic-Operations.html
[3] http://doc.trolltech.com/4.3/atomic-operations.html
     See also the Qt header file QtCore/qatomic_i386.h, and
     its brothers.
[4] http://www.alexonlinux.com/multithreaded-simple-data-type-access-and-atomic-variables#precautions
[5] http://siyobik.info/index.php?module=x86&id=159

_______________________________________________
Linux-audio-dev mailing list
Linux-audio-dev@email-addr-hidden
http://lists.linuxaudio.org/mailman/listinfo/linux-audio-dev

Received on Sat Dec 12 20:15:05 2009

This archive was generated by hypermail 2.1.8 : Sat Dec 12 2009 - 20:15:05 EET