<ol>
<li><a href="#introduction">Introduction</a></li>
- <li><a href="#loadstore">Load and store</a></li>
+ <li><a href="#outsideatomic">Optimization outside atomic</a></li>
+ <li><a href="#atomicinst">Atomic instructions</a></li>
<li><a href="#ordering">Atomic orderings</a></li>
- <li><a href="#otherinst">Other atomic instructions</a></li>
<li><a href="#iropt">Atomics and IR optimization</a></li>
<li><a href="#codegen">Atomics and Codegen</a></li>
</ol>
<p>The atomic instructions are designed specifically to provide readable IR and
optimized code generation for the following:</p>
<ul>
- <li>The new C++0x <code><atomic></code> header.</li>
+ <li>The new C++0x <code><atomic></code> header.
+ (<a href="http://www.open-std.org/jtc1/sc22/wg21/">C++0x draft available here</a>.)
+ (<a href="http://www.open-std.org/jtc1/sc22/wg14/">C1x draft available here</a>)</li>
<li>Proper semantics for Java-style memory, for both <code>volatile</code> and
- regular shared variables.</li>
- <li>gcc-compatible <code>__sync_*</code> builtins.</li>
+ regular shared variables.
+ (<a href="http://java.sun.com/docs/books/jls/third_edition/html/memory.html">Java Specification</a>)</li>
+ <li>gcc-compatible <code>__sync_*</code> builtins.
+ (<a href="http://gcc.gnu.org/onlinedocs/gcc/Atomic-Builtins.html">Description</a>)</li>
<li>Other scenarios with atomic semantics, including <code>static</code>
variables with non-trivial constructors in C++.</li>
</ul>
+<p>Atomic and volatile in the IR are orthogonal; "volatile" is the C/C++
+ volatile, which ensures that every volatile load and store happens and is
+ performed in the stated order. A couple examples: if a
+ SequentiallyConsistent store is immediately followed by another
+ SequentiallyConsistent store to the same address, the first store can
+ be erased. This transformation is not allowed for a pair of volatile
+ stores. On the other hand, a non-volatile non-atomic load can be moved
+ across a volatile load freely, but not an Acquire load.</p>
+
<p>This document is intended to provide a guide to anyone either writing a
frontend for LLVM or working on optimization passes for LLVM with a guide
for how to deal with instructions with special semantics in the presence of
<!-- *********************************************************************** -->
<h2>
- <a name="loadstore">Load and store</a>
+ <a name="outsideatomic">Optimization outside atomic</a>
</h2>
<!-- *********************************************************************** -->
<div>
<p>The basic <code>'load'</code> and <code>'store'</code> allow a variety of
- optimizations, but can have unintuitive results in a concurrent environment.
- For a frontend writer, the rule is essentially that all memory accessed
- with basic loads and stores by multiple threads should be protected by a
- lock or other synchronization; otherwise, you are likely to run into
- undefined behavior. (Do not use volatile as a substitute for atomics; it
- might work on some platforms, but does not provide the necessary guarantees
- in general.)</p>
+ optimizations, but can lead to undefined results in a concurrent environment;
+ see <a href="#o_nonatomic">NonAtomic</a>. This section specifically goes
+ into the one optimizer restriction which applies in concurrent environments,
+ which gets a bit more of an extended description because any optimization
+ dealing with stores needs to be aware of it.</p>
<p>From the optimizer's point of view, the rule is that if there
- are not any instructions with atomic ordering involved, concurrency does not
- matter, with one exception: if a variable might be visible to another
+ are not any instructions with atomic ordering involved, concurrency does
+ not matter, with one exception: if a variable might be visible to another
thread or signal handler, a store cannot be inserted along a path where it
- might not execute otherwise. Note that speculative loads are allowed;
- a load which is part of a race returns <code>undef</code>, but is not
- undefined behavior.</p>
+ might not execute otherwise. Take the following example:</p>
+
+<pre>
+/* C code, for readability; run through clang -O2 -S -emit-llvm to get
+ equivalent IR */
+int x;
+void f(int* a) {
+ for (int i = 0; i < 100; i++) {
+ if (a[i])
+ x += 1;
+ }
+}
+</pre>
+
+<p>The following is equivalent in non-concurrent situations:</p>
+
+<pre>
+int x;
+void f(int* a) {
+ int xtemp = x;
+ for (int i = 0; i < 100; i++) {
+ if (a[i])
+ xtemp += 1;
+ }
+ x = xtemp;
+}
+</pre>
+
+<p>However, LLVM is not allowed to transform the former to the latter: it could
+ indirectly introduce undefined behavior if another thread can access x at
+ the same time. (This example is particularly of interest because before the
+ concurrency model was implemented, LLVM would perform this
+ transformation.)</p>
+
+<p>Note that speculative loads are allowed; a load which
+ is part of a race returns <code>undef</code>, but does not have undefined
+ behavior.</p>
+
+
+</div>
+
+<!-- *********************************************************************** -->
+<h2>
+ <a name="atomicinst">Atomic instructions</a>
+</h2>
+<!-- *********************************************************************** -->
+
+<div>
<p>For cases where simple loads and stores are not sufficient, LLVM provides
- atomic loads and stores with varying levels of guarantees.</p>
+ various atomic instructions. The exact guarantees provided depend on the
+ ordering; see <a href="#ordering">Atomic orderings</a></p>
+
+<p><code>load atomic</code> and <code>store atomic</code> provide the same
+ basic functionality as non-atomic loads and stores, but provide additional
+ guarantees in situations where threads and signals are involved.</p>
+
+<p><code>cmpxchg</code> and <code>atomicrmw</code> are essentially like an
+ atomic load followed by an atomic store (where the store is conditional for
+ <code>cmpxchg</code>), but no other memory operation can happen on any thread
+ between the load and store. Note that LLVM's cmpxchg does not provide quite
+ as many options as the C++0x version.</p>
+
+<p>A <code>fence</code> provides Acquire and/or Release ordering which is not
+ part of another operation; it is normally used along with Monotonic memory
+ operations. A Monotonic load followed by an Acquire fence is roughly
+ equivalent to an Acquire load.</p>
+
+<p>Frontends generating atomic instructions generally need to be aware of the
+ target to some degree; atomic instructions are guaranteed to be lock-free,
+ and therefore an instruction which is wider than the target natively supports
+ can be impossible to generate.</p>
</div>
<p>In order to achieve a balance between performance and necessary guarantees,
there are six levels of atomicity. They are listed in order of strength;
each level includes all the guarantees of the previous level except for
- Acquire/Release.</p>
+ Acquire/Release. (See also <a href="LangRef.html#ordering">LangRef</a>.)</p>
+
+<!-- ======================================================================= -->
+<h3>
+ <a name="o_notatomic">NotAtomic</a>
+</h3>
+
+<div>
+
+<p>NotAtomic is the obvious, a load or store which is not atomic. (This isn't
+ really a level of atomicity, but is listed here for comparison.) This is
+ essentially a regular load or store. If there is a race on a given memory
+ location, loads from that location return undef.</p>
+
+<dl>
+ <dt>Relevant standard</dt>
+ <dd>This is intended to match shared variables in C/C++, and to be used
+ in any other context where memory access is necessary, and
+ a race is impossible. (The precise definition is in
+ <a href="LangRef.html#memmodel">LangRef</a>.)
+ <dt>Notes for frontends</dt>
+ <dd>The rule is essentially that all memory accessed with basic loads and
+ stores by multiple threads should be protected by a lock or other
+ synchronization; otherwise, you are likely to run into undefined
+ behavior. If your frontend is for a "safe" language like Java,
+ use Unordered to load and store any shared variable. Note that NotAtomic
+ volatile loads and stores are not properly atomic; do not try to use
+ them as a substitute. (Per the C/C++ standards, volatile does provide
+ some limited guarantees around asynchronous signals, but atomics are
+ generally a better solution.)
+ <dt>Notes for optimizers</dt>
+ <dd>Introducing loads to shared variables along a codepath where they would
+ not otherwise exist is allowed; introducing stores to shared variables
+ is not. See <a href="#outsideatomic">Optimization outside
+ atomic</a>.</dd>
+ <dt>Notes for code generation</dt>
+ <dd>The one interesting restriction here is that it is not allowed to write
+ to bytes outside of the bytes relevant to a store. This is mostly
+ relevant to unaligned stores: it is not allowed in general to convert
+ an unaligned store into two aligned stores of the same width as the
+ unaligned store. Backends are also expected to generate an i8 store
+ as an i8 store, and not an instruction which writes to surrounding
+ bytes. (If you are writing a backend for an architecture which cannot
+ satisfy these restrictions and cares about concurrency, please send an
+ email to llvmdev.)</dd>
+</dl>
+
+</div>
+
+
+<!-- ======================================================================= -->
+<h3>
+ <a name="o_unordered">Unordered</a>
+</h3>
+
+<div>
<p>Unordered is the lowest level of atomicity. It essentially guarantees that
- races produce somewhat sane results instead of having undefined behavior.
- This is intended to match the Java memory model for shared variables. It
- cannot be used for synchronization, but is useful for Java and other
- "safe" languages which need to guarantee that the generated code never
- exhibits undefined behavior. Note that this guarantee is cheap on common
- platforms for loads of a native width, but can be expensive or unavailable
- for wider loads, like a 64-bit load on ARM. (A frontend for a "safe"
- language would normally split a 64-bit load on ARM into two 32-bit
- unordered loads.) In terms of the optimizer, this prohibits any
- transformation that transforms a single load into multiple loads,
- transforms a store into multiple stores, narrows a store, or stores a
- value which would not be stored otherwise. Some examples of unsafe
- optimizations are narrowing an assignment into a bitfield, rematerializing
- a load, and turning loads and stores into a memcpy call. Reordering
- unordered operations is safe, though, and optimizers should take
- advantage of that because unordered operations are common in
- languages that need them.</p>
+ races produce somewhat sane results instead of having undefined behavior.
+ It also guarantees the operation to be lock-free, so it do not depend on
+ the data being part of a special atomic structure or depend on a separate
+ per-process global lock. Note that code generation will fail for
+ unsupported atomic operations; if you need such an operation, use explicit
+ locking.</p>
+
+<dl>
+ <dt>Relevant standard</dt>
+ <dd>This is intended to match the Java memory model for shared
+ variables.</dd>
+ <dt>Notes for frontends</dt>
+ <dd>This cannot be used for synchronization, but is useful for Java and
+ other "safe" languages which need to guarantee that the generated
+ code never exhibits undefined behavior. Note that this guarantee
+ is cheap on common platforms for loads of a native width, but can
+ be expensive or unavailable for wider loads, like a 64-bit store
+ on ARM. (A frontend for Java or other "safe" languages would normally
+ split a 64-bit store on ARM into two 32-bit unordered stores.)
+ <dt>Notes for optimizers</dt>
+ <dd>In terms of the optimizer, this prohibits any transformation that
+ transforms a single load into multiple loads, transforms a store
+ into multiple stores, narrows a store, or stores a value which
+ would not be stored otherwise. Some examples of unsafe optimizations
+ are narrowing an assignment into a bitfield, rematerializing
+ a load, and turning loads and stores into a memcpy call. Reordering
+ unordered operations is safe, though, and optimizers should take
+ advantage of that because unordered operations are common in
+ languages that need them.</dd>
+ <dt>Notes for code generation</dt>
+ <dd>These operations are required to be atomic in the sense that if you
+ use unordered loads and unordered stores, a load cannot see a value
+ which was never stored. A normal load or store instruction is usually
+ sufficient, but note that an unordered load or store cannot
+ be split into multiple instructions (or an instruction which
+ does multiple memory operations, like <code>LDRD</code> on ARM).</dd>
+</dl>
+
+</div>
+
+<!-- ======================================================================= -->
+<h3>
+ <a name="o_monotonic">Monotonic</a>
+</h3>
+
+<div>
<p>Monotonic is the weakest level of atomicity that can be used in
synchronization primitives, although it does not provide any general
synchronization. It essentially guarantees that if you take all the
operations affecting a specific address, a consistent ordering exists.
- This corresponds to the C++0x/C1x <code>memory_order_relaxed</code>; see
- those standards for the exact definition. If you are writing a frontend, do
- not use the low-level synchronization primitives unless you are compiling
- a language which requires it or are sure a given pattern is correct. In
- terms of the optimizer, this can be treated as a read+write on the relevant
- memory location (and alias analysis will take advantage of that). In
- addition, it is legal to reorder non-atomic and Unordered loads around
- Monotonic loads. CSE/DSE and a few other optimizations are allowed, but
- Monotonic operations are unlikely to be used in ways which would make
- those optimizations useful.</p>
+
+<dl>
+ <dt>Relevant standard</dt>
+ <dd>This corresponds to the C++0x/C1x <code>memory_order_relaxed</code>;
+ see those standards for the exact definition.
+ <dt>Notes for frontends</dt>
+ <dd>If you are writing a frontend which uses this directly, use with caution.
+ The guarantees in terms of synchronization are very weak, so make
+ sure these are only used in a pattern which you know is correct.
+ Generally, these would either be used for atomic operations which
+ do not protect other memory (like an atomic counter), or along with
+ a <code>fence</code>.</dd>
+ <dt>Notes for optimizers</dt>
+ <dd>In terms of the optimizer, this can be treated as a read+write on the
+ relevant memory location (and alias analysis will take advantage of
+ that). In addition, it is legal to reorder non-atomic and Unordered
+ loads around Monotonic loads. CSE/DSE and a few other optimizations
+ are allowed, but Monotonic operations are unlikely to be used in ways
+ which would make those optimizations useful.</dd>
+ <dt>Notes for code generation</dt>
+ <dd>Code generation is essentially the same as that for unordered for loads
+ and stores. No fences are required. <code>cmpxchg</code> and
+ <code>atomicrmw</code> are required to appear as a single operation.</dd>
+</dl>
+
+</div>
+
+<!-- ======================================================================= -->
+<h3>
+ <a name="o_acquire">Acquire</a>
+</h3>
+
+<div>
<p>Acquire provides a barrier of the sort necessary to acquire a lock to access
- other memory with normal loads and stores. This corresponds to the
- C++0x/C1x <code>memory_order_acquire</code>. It should also be used for
- C++0x/C1x <code>memory_order_consume</code>. This is a low-level
- synchronization primitive. In general, optimizers should treat this like
- a nothrow call.</p>
+ other memory with normal loads and stores.
+
+<dl>
+ <dt>Relevant standard</dt>
+ <dd>This corresponds to the C++0x/C1x <code>memory_order_acquire</code>. It
+ should also be used for C++0x/C1x <code>memory_order_consume</code>.
+ <dt>Notes for frontends</dt>
+ <dd>If you are writing a frontend which uses this directly, use with caution.
+ Acquire only provides a semantic guarantee when paired with a Release
+ operation.</dd>
+ <dt>Notes for optimizers</dt>
+ <dd>Optimizers not aware of atomics can treat this like a nothrow call.
+ It is also possible to move stores from before an Acquire load
+ or read-modify-write operation to after it, and move non-Acquire
+ loads from before an Acquire operation to after it.</dd>
+ <dt>Notes for code generation</dt>
+ <dd>Architectures with weak memory ordering (essentially everything relevant
+ today except x86 and SPARC) require some sort of fence to maintain
+ the Acquire semantics. The precise fences required varies widely by
+ architecture, but for a simple implementation, most architectures provide
+ a barrier which is strong enough for everything (<code>dmb</code> on ARM,
+ <code>sync</code> on PowerPC, etc.). Putting such a fence after the
+ equivalent Monotonic operation is sufficient to maintain Acquire
+ semantics for a memory operation.</dd>
+</dl>
+
+</div>
+
+<!-- ======================================================================= -->
+<h3>
+ <a name="o_acquire">Release</a>
+</h3>
+
+<div>
<p>Release is similar to Acquire, but with a barrier of the sort necessary to
- release a lock. This corresponds to the C++0x/C1x
- <code>memory_order_release</code>. In general, optimizers should treat this
- like a nothrow call.</p>
-
-<p>AcquireRelease (<code>acq_rel</code> in IR) provides both an Acquire and a Release barrier.
- This corresponds to the C++0x/C1x <code>memory_order_acq_rel</code>. In general,
- optimizers should treat this like a nothrow call.</p>
-
-<p>SequentiallyConsistent (<code>seq_cst</code> in IR) provides Acquire and/or
- Release semantics, and in addition guarantees a total ordering exists with
- all other SequentiallyConsistent operations. This corresponds to the
- C++0x/C1x <code>memory_order_seq_cst</code>, and Java volatile. The intent
- of this ordering level is to provide a programming model which is relatively
- easy to understand. In general, optimizers should treat this like a
- nothrow call.</p>
+ release a lock.
+
+<dl>
+ <dt>Relevant standard</dt>
+ <dd>This corresponds to the C++0x/C1x <code>memory_order_release</code>.</dd>
+ <dt>Notes for frontends</dt>
+ <dd>If you are writing a frontend which uses this directly, use with caution.
+ Release only provides a semantic guarantee when paired with a Acquire
+ operation.</dd>
+ <dt>Notes for optimizers</dt>
+ <dd>Optimizers not aware of atomics can treat this like a nothrow call.
+ It is also possible to move loads from after a Release store
+ or read-modify-write operation to before it, and move non-Release
+ stores from after an Release operation to before it.</dd>
+ <dt>Notes for code generation</dt>
+ <dd>See the section on Acquire; a fence before the relevant operation is
+ usually sufficient for Release. Note that a store-store fence is not
+ sufficient to implement Release semantics; store-store fences are
+ generally not exposed to IR because they are extremely difficult to
+ use correctly.</dd>
+</dl>
</div>
-<!-- *********************************************************************** -->
-<h2>
- <a name="otherinst">Other atomic instructions</a>
-</h2>
-<!-- *********************************************************************** -->
+<!-- ======================================================================= -->
+<h3>
+ <a name="o_acqrel">AcquireRelease</a>
+</h3>
<div>
-<p><code>cmpxchg</code> and <code>atomicrmw</code> are essentially like an
- atomic load followed by an atomic store (where the store is conditional for
- <code>cmpxchg</code>), but no other memory operation can happen between
- the load and store. Note that our cmpxchg does not have quite as many
- options for making cmpxchg weaker as the C++0x version.</p>
+<p>AcquireRelease (<code>acq_rel</code> in IR) provides both an Acquire and a
+ Release barrier (for fences and operations which both read and write memory).
+
+<dl>
+ <dt>Relevant standard</dt>
+ <dd>This corresponds to the C++0x/C1x <code>memory_order_acq_rel</code>.
+ <dt>Notes for frontends</dt>
+ <dd>If you are writing a frontend which uses this directly, use with caution.
+ Acquire only provides a semantic guarantee when paired with a Release
+ operation, and vice versa.</dd>
+ <dt>Notes for optimizers</dt>
+ <dd>In general, optimizers should treat this like a nothrow call; the
+ the possible optimizations are usually not interesting.</dd>
+ <dt>Notes for code generation</dt>
+ <dd>This operation has Acquire and Release semantics; see the sections on
+ Acquire and Release.</dd>
+</dl>
-<p>A <code>fence</code> provides Acquire and/or Release ordering which is not
- part of another operation; it is normally used along with Monotonic memory
- operations. A Monotonic load followed by an Acquire fence is roughly
- equivalent to an Acquire load.</p>
+</div>
-<p>Frontends generating atomic instructions generally need to be aware of the
- target to some degree; atomic instructions are guaranteed to be lock-free,
- and therefore an instruction which is wider than the target natively supports
- can be impossible to generate.</p>
+<!-- ======================================================================= -->
+<h3>
+ <a name="o_seqcst">SequentiallyConsistent</a>
+</h3>
+
+<div>
+
+<p>SequentiallyConsistent (<code>seq_cst</code> in IR) provides
+ Acquire semantics for loads and Release semantics for
+ stores. Additionally, it guarantees that a total ordering exists
+ between all SequentiallyConsistent operations.
+
+<dl>
+ <dt>Relevant standard</dt>
+ <dd>This corresponds to the C++0x/C1x <code>memory_order_seq_cst</code>,
+ Java volatile, and the gcc-compatible <code>__sync_*</code> builtins
+ which do not specify otherwise.
+ <dt>Notes for frontends</dt>
+ <dd>If a frontend is exposing atomic operations, these are much easier to
+ reason about for the programmer than other kinds of operations, and using
+ them is generally a practical performance tradeoff.</dd>
+ <dt>Notes for optimizers</dt>
+ <dd>Optimizers not aware of atomics can treat this like a nothrow call.
+ For SequentiallyConsistent loads and stores, the same reorderings are
+ allowed as for Acquire loads and Release stores, except that
+ SequentiallyConsistent operations may not be reordered.</dd>
+ <dt>Notes for code generation</dt>
+ <dd>SequentiallyConsistent loads minimally require the same barriers
+ as Acquire operations and SequentiallyConsistent stores require
+ Release barriers. Additionally, the code generator must enforce
+ ordering between SequentiallyConsistent stores followed by
+ SequentiallyConsistent loads. This is usually done by emitting
+ either a full fence before the loads or a full fence after the
+ stores; which is preferred varies by architecture.</dd>
+</dl>
+
+</div>
</div>
<ul>
<li>isSimple(): A load or store which is not volatile or atomic. This is
what, for example, memcpyopt would check for operations it might
- transform.
+ transform.</li>
<li>isUnordered(): A load or store which is not volatile and at most
Unordered. This would be checked, for example, by LICM before hoisting
- an operation.
+ an operation.</li>
<li>mayReadFromMemory()/mayWriteToMemory(): Existing predicate, but note
that they return true for any operation which is volatile or at least
- Monotonic.
+ Monotonic.</li>
<li>Alias analysis: Note that AA will return ModRef for anything Acquire or
- Release, and for the address accessed by any Monotonic operation.
+ Release, and for the address accessed by any Monotonic operation.</li>
</ul>
-<p>There are essentially two components to supporting atomic operations. The
- first is making sure to query isSimple() or isUnordered() instead
- of isVolatile() before transforming an operation. The other piece is
- making sure that a transform does not end up replacing, for example, an
- Unordered operation with a non-atomic operation. Most of the other
- necessary checks automatically fall out from existing predicates and
- alias analysis queries.</p>
+<p>To support optimizing around atomic operations, make sure you are using
+ the right predicates; everything should work if that is done. If your
+ pass should optimize some atomic operations (Unordered operations in
+ particular), make sure it doesn't replace an atomic load or store with
+ a non-atomic operation.</p>
<p>Some examples of how optimizations interact with various kinds of atomic
operations:
handles anything marked volatile very conservatively. This should get
fixed at some point.</p>
+<p>Common architectures have some way of representing at least a pointer-sized
+ lock-free <code>cmpxchg</code>; such an operation can be used to implement
+ all the other atomic operations which can be represented in IR up to that
+ size. Backends are expected to implement all those operations, but not
+ operations which cannot be implemented in a lock-free manner. It is
+ expected that backends will give an error when given an operation which
+ cannot be implemented. (The LLVM code generator is not very helpful here
+ at the moment, but hopefully that will change.)</p>
+
<p>The implementation of atomics on LL/SC architectures (like ARM) is currently
a bit of a mess; there is a lot of copy-pasted code across targets, and
the representation is relatively unsuited to optimization (it would be nice
<p>On ARM, MIPS, and many other RISC architectures, Acquire, Release, and
SequentiallyConsistent semantics require barrier instructions
for every such operation. Loads and stores generate normal instructions.
- <code>atomicrmw</code> and <code>cmpxchg</code> generate LL/SC loops.</p>
-
+ <code>cmpxchg</code> and <code>atomicrmw</code> can be represented using
+ a loop with LL/SC-style instructions which take some sort of exclusive
+ lock on a cache line (<code>LDREX</code> and <code>STREX</code> on
+ ARM, etc.). At the moment, the IR does not provide any way to represent a
+ weak <code>cmpxchg</code> which would not require a loop.</p>
</div>
<!-- *********************************************************************** -->
projects / oota-llvm.git / blobdiff