I'm just reading the C++ concurrency in action book by Anthony Williams.
There is this classic example with two threads, one produce data, the other one consumes the data and A.W. wrote that code pretty clear :
std::vector<int> data;
std::atomic<bool> data_ready(false);
void reader_thread()
{
while(!data_ready.load())
{
std::this_thread::sleep(std::milliseconds(1));
}
std::cout << "The answer=" << data[0] << "\n";
}
void writer_thread()
{
data.push_back(42);
data_ready = true;
}
And I really don't understand why this code differs from one where I'd use a classic volatile bool instead of the atomic one.
If someone could open my mind on the subject, I'd be grateful.
Thanks.
A "classic" bool, as you put it, would not work reliably (if at all). One reason for this is that the compiler could (and most likely does, at least with optimizations enabled) load data_ready only once from memory, because there is no indication that it ever changes in the context of reader_thread.
You could work around this problem by using volatile bool to enforce loading it every time (which would probably seem to work) but this would still be undefined behavior regarding the C++ standard because the access to the variable is neither synchronized nor atomic.
You could enforce synchronization using the locking facilities from the mutex header, but this would introduce (in your example) unnecessary overhead (hence std::atomic).
The problem with volatile is that it only guarantees that instructions are not omitted and the instruction ordering is preserved. volatile does not guarantee a memory barrier to enforce cache coherence. What this means is that writer_thread on processor A can write the value to it's cache (and maybe even to the main memory) without reader_thread on processor B seeing it, because the cache of processor B is not consistent with the cache of processor A. For a more thorough explanation see memory barrier and cache coherence on Wikipedia.
There can be additional problems with more complex expressions than x = y (i.e. x += y) that would require synchronization through a lock (or in this simple case an atomic +=) to ensure the value of x does not change during processing.
x += y for example is actually:
read x
compute x + y
write result back to x
If a context switch to another thread occurs during the computation this can result in something like this (2 threads, both doing x += 2; assuming x = 0):
Thread A Thread B
------------------------ ------------------------
read x (0)
compute x (0) + 2
<context switch>
read x (0)
compute x (0) + 2
write x (2)
<context switch>
write x (2)
Now x = 2 even though there were two += 2 computations. This effect is known as tearing.
The big difference is that this code is correct, while the version with bool instead of atomic<bool> has undefined behavior.
These two lines of code create a race condition (formally, a conflict) because they read from and write to the same variable:
Reader
while (!data_ready)
And writer
data_ready = true;
And a race condition on a normal variable causes undefined behavior, according to the C++11 memory model.
The rules are found in section 1.10 of the Standard, the most relevant being:
Two actions are potentially concurrent if
they are performed by different threads, or
they are unsequenced, and at least one is performed by a signal handler.
The execution of a program contains a data race if it contains two potentially concurrent conflicting actions, at least one of which is not atomic, and neither happens before the other, except for the special case for signal handlers described below. Any such data race results in undefined behavior.
You can see that whether the variable is atomic<bool> makes a very big difference to this rule.
Ben Voigt's answer is completely correct, still a little theoretical, and as I've been asked by a colleague "what does this mean for me", I decided to try my luck with a little more practical answer.
With your sample, the "simplest" optimization problem that could occur is the following:
According to the Standard, an optimized execution order may not change the functionality of a program. Problem is, this is only true for single threaded programs, or single threads in multithreaded programs.
So, for writer_thread and a (volatile) bool
data.push_back(42);
data_ready = true;
and
data_ready = true;
data.push_back(42);
are equivalent.
The result is, that
std::cout << "The answer=" << data[0] << "\n";
can be executed without having pushed any value into data.
An atomic bool does prevent this kind of optimization, as per definition it may not be reordered. There are flags for atomic operations which allow statements to be moved in front of the operation but not to the back, and vice versa, but those require a really advanced knowledge of your programming structure and the problems it can cause...
Related
volatile bool b;
Thread1: //only reads b
void f1() {
while (1) {
if (b) {do something};
else { do something else};
}
}
Thread2:
//only sets b to true if certain condition met
// updated by thread2
void f2() {
while (1) {
//some local condition evaluated - local_cond
if (!b && (local_cond == true)) b = true;
//some other work
}
}
Thread3:
//only sets b to false when it gets a message on a socket its listening to
void f3() {
while (1) {
//select socket
if (expected message came) b = false;
//do some other work
}
}
If thread2 updates b first at time t and later thread3 updates b at time t+5:
will thread1 see the latest value "in time" whenever it is reading b?
for example: reads from t+delta to t+5+delta should read true and
reads after t+5+delta should read false.
delta is the time for the store of "b" into memory when one of threads 2 or 3 updated it
The effect of volatile keyword is principally two things (I avoid scientifically strict formulations here):
1) Its accesses can't be cached or combined. (UPD: on suggestion, I underline this is for caching in registers or another compiler-provided location, not the RAM cache in CPU.) For example, the following code:
x = 1;
x = 2;
for a volatile x will never be combined into single x = 2, whatever optimization level is required; but if x is not volatile, even low levels will likely cause this collapse into a single write. The same for reads: each read operation will access the variable value without any attempt to cache it.
2) All volatile operations are relayed onto machine command layer in the same order between them (to underline, only between volatile operations), as they are defined in source code.
But this is not true for accesses between non-volatile and volatile memory. For the following code:
int *x;
volatile int *vy;
void foo()
{
*x = 1;
*vy = 101;
*x = 2;
*vy = 102;
}
gcc (9.4) with -O2 and clang (10.0) with -O produce something similar to:
movq x(%rip), %rax
movq vy(%rip), %rcx
movl $101, (%rcx)
movl $2, (%rax)
movl $102, (%rcx)
retq
so one access to x is already gone, despite its presence between two volatile accesses. If one need the first x = 1 to succeed before first write to vy, let him put an explicit barrier (since C11, atomic_signal_fence is the platform-independent mean for this).
That was the common rule but without regarding multithread issues. What happens here with multithreading?
Well, imagine as you declare that thread 2 writes true to b, so, this is writing of value 1 to single-byte location. But, this is ordinary write without any memory ordering requirements. What you provided with volatile is that compiler won't optimize it. But what for processor?
If this was a modern abstract processor, or one with relaxed rules, like ARM, I'd say nothing prevent it from postponing the real write for an indefinite time. (To clarify, "write" is exposing the operation to RAM-and-all-caches conglomerate.) It's fully up to processor's deliberation. Well, processors are designed to flush their stockpiling of pending writes as fast as possible. But what affects real delay, you can't know: for example, it could "decide" to fill instruction cache with a few next lines, or flush another queued writings... lots of variants. The only thing we know it provides "best effort" to flush all queued operations, to avoid getting buried under previous results. That's truly natural and nothing more.
With x86, there is an additional factor. Nearly every memory write (and, I guess, this one as well) is "releasing" write in x86, so, all previous reads and writes shall be completed before this write. But, the gut fact is that the operations to complete are before this write. So when you write true to volatile b, you will be sure all previous operations have already got visible to other participants... but this one still could be postponed for a while... how long? Nanoseconds? Microseconds? Any other write to memory will flush and so publish this write to b... do you have writes in cycle iteration of thread 2?
The same affects thread 3. You can't be sure this b = false will be published to other CPUs when you need it. Delay is unpredictable. The only thing is guaranteed, if this is not a realtime-aware hardware system, for an indefinite time, and the ISA rules and barriers provide ordering but not exact times. And, x86 is definitely not for such a realtime.
Well, all this means you also need an explicit barrier after write which affects not only compiler, but CPU as well: barrier before previous write and following reads or writes. Among C/C++ means, full barrier satifies this - so you have to add std::atomic_thread_fence(std::memory_order_seq_cst) or use atomic variable (instead of plain volatile one) with the same memory order for write.
And, all this still won't provide you with exact timings like you described ("t" and "t+5"), because the visible "timestamps" of the same operation can differ for different CPUs! (Well, this resembles Einstein's relativity a bit.) All you could say in this situation is that something is written into memory, and typically (not always) the inter-CPU order is what you expected (but the ordering violation will punish you).
But, I can't catch the general idea of what do you want to implement with this flag b. What do you want from it, what state should it reflect? Let you return to the upper level task and reformulate. Is this (I'm just guessing on coffee grounds) a green light to do something, which is cancelled by an external order? If so, an internal permission ("we are ready") from the thread 2 shall not drop this cancellation. This can be done using different approaches, as:
1) Just separate flags and a mutex/spinlock around their set. Easy but a bit costly (or even substantially costly, I don't know your environment).
2) An atomically modified analog. For example, you can use a bitfield variable which is modified using compare-and-swap. Assign bit 0 to "ready" but bit 1 for "cancelled". For C, atomic_compare_exchange_strong is what you'll need here at x86 (and at most other ISAs). And, volatile is not needed anymore here if you keep residing with memory_order_seq_cst.
Will thread1 see the latest value "in time" whenever it is reading b?
Yes, the volatile keyword denotes that it can be modified outside of the thread or hardware without the compiler being aware thus every access (both read and write) will be made through an lvalue expression of volatile-qualified type is considered an observable side effect for the purpose of optimization and is evaluated strictly according to the rules of the abstract machine (that is, all writes are completed at some time before the next sequence point). This means that within a single thread of execution, a volatile access cannot be optimized out or reordered relative to another visible side effect that is separated by a sequence point from the volatile access.
Unfortunately, the volatile keyword is not thread-safe and operation will have to be taken with care, it is recommended to use atomic for this, unless in an embedded or bare-metal scenario.
Also the whole struct should be atomic struct X {int a; volatile bool b;};.
Say I have a system with 2 cores. The first core runs thread 2, the second core runs thread 3.
reads from t+delta to t+5+delta should read true and reads after t+5+delta should read false.
Problem is that thread 1 will read at t + 10000000 when the kernel decides one of the threads has run long enough and schedules a different thread. So it likely thread1 will not see the change a lot of the time.
Note: this ignores all the additional problems of synchronicity of caches and observability. If the thread isn't even running all of that becomes irrelevant.
This is a question about the formal guarantees of the C++ standard.
The standard points out that the rules for std::memory_order_relaxed atomic variables allow "out of thin air" / "out of the blue" values to appear.
But for non-atomic variables, can this example have UB? Is r1 == r2 == 42 possible in the C++ abstract machine? Neither variable == 42 initially so you'd expect neither if body should execute, meaning no writes to the shared variables.
// Global state
int x = 0, y = 0;
// Thread 1:
r1 = x;
if (r1 == 42) y = r1;
// Thread 2:
r2 = y;
if (r2 == 42) x = 42;
The above example is adapted from the standard, which explicitly says such behavior is allowed by the specification for atomic objects:
[Note: The requirements do allow r1 == r2 == 42 in the following
example, with x and y initially zero:
// Thread 1:
r1 = x.load(memory_order_relaxed);
if (r1 == 42) y.store(r1, memory_order_relaxed);
// Thread 2:
r2 = y.load(memory_order_relaxed);
if (r2 == 42) x.store(42, memory_order_relaxed);
However, implementations should not allow such behavior. – end note]
What part of the so called "memory model" protects non atomic objects from these interactions caused by reads seeing out-of-thin-air values?
When a race condition would exist with different values for x and y, what guarantees that read of a shared variable (normal, non atomic) cannot see such values?
Can not-executed if bodies create self-fulfilling conditions that lead to a data-race?
The text of your question seems to be missing the point of the example and out-of-thin-air values. Your example does not contain data-race UB. (It might if x or y were set to 42 before those threads ran, in which case all bets are off and the other answers citing data-race UB apply.)
There is no protection against real data races, only against out-of-thin-air values.
I think you're really asking how to reconcile that mo_relaxed example with sane and well-defined behaviour for non-atomic variables. That's what this answer covers.
The note is pointing out a hole in the atomic mo_relaxed formalism, not warning you of a real possible effect on some implementations.
This gap does not (I think) apply to non-atomic objects, only to mo_relaxed.
They say However, implementations should not allow such behavior. – end note]. Apparently the standards committee couldn't find a way to formalize that requirement so for now it's just a note, but is not intended to be optional.
It's clear that even though this isn't strictly normative, the C++ standard intends to disallow out-of-thin-air values for relaxed atomic (and in general I assume). Later standards discussion, e.g. 2018's p0668r5: Revising the C++ memory model (which doesn't "fix" this, it's an unrelated change) includes juicy side-nodes like:
We still do not have an acceptable way to make our informal (since C++14) prohibition of out-of-thin-air results precise. The primary practical effect of that is that formal verification of C++ programs using relaxed atomics remains unfeasible. The above paper suggests a solution similar to http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2013/n3710.html . We continue to ignore the problem here ...
So yes, the normative parts of the standard are apparently weaker for relaxed_atomic than they are for non-atomic. This seems to be an unfortunately side effect of how they define the rules.
AFAIK no implementations can produce out-of-thin-air values in real life.
Later versions of the standard phrase the informal recommendation more clearly, e.g. in the current draft: https://timsong-cpp.github.io/cppwp/atomics.order#8
Implementations should ensure that no “out-of-thin-air” values are computed that circularly depend on their own computation.
...
[ Note: The recommendation [of 8.] similarly disallows r1 == r2 == 42 in the following example, with x and y again initially zero:
// Thread 1:
r1 = x.load(memory_order::relaxed);
if (r1 == 42) y.store(42, memory_order::relaxed);
// Thread 2:
r2 = y.load(memory_order::relaxed);
if (r2 == 42) x.store(42, memory_order::relaxed);
— end note ]
(This rest of the answer was written before I was sure that the standard intended to disallow this for mo_relaxed, too.)
I'm pretty sure the C++ abstract machine does not allow r1 == r2 == 42.
Every possible ordering of operations in the C++ abstract machine operations leads to r1=r2=0 without UB, even without synchronization. Therefore the program has no UB and any non-zero result would violate the "as-if" rule.
Formally, ISO C++ allows an implementation to implement functions / programs in any way that gives the same result as the C++ abstract machine would. For multi-threaded code, an implementation can pick one possible abstract-machine ordering and decide that's the ordering that always happens. (e.g. when reordering relaxed atomic stores when compiling to asm for a strongly-ordered ISA. The standard as written even allows coalescing atomic stores but compilers choose not to). But the result of the program always has to be something the abstract machine could have produced. (Only the Atomics chapter introduces the possibility of one thread observing the actions of another thread without mutexes. Otherwise that's not possible without data-race UB).
I think the other answers didn't look carefully enough at this. (And neither did I when it was first posted). Code that doesn't execute doesn't cause UB (including data-race UB), and compilers aren't allowed to invent writes to objects. (Except in code paths that already unconditionally write them, like y = (x==42) ? 42 : y; which would obviously create data-race UB.)
For any non-atomic object, if don't actually write it then other threads might also be reading it, regardless of code inside not-executed if blocks. The standard allows this and doesn't allow a variable to suddenly read as a different value when the abstract machine hasn't written it. (And for objects we don't even read, like neighbouring array elements, another thread might even be writing them.)
Therefore we can't do anything that would let another thread temporarily see a different value for the object, or step on its write. Inventing writes to non-atomic objects is basically always a compiler bug; this is well known and universally agreed upon because it can break code that doesn't contain UB (and has done so in practice for a few cases of compiler bugs that created it, e.g. IA-64 GCC I think had such a bug at one point that broke the Linux kernel). IIRC, Herb Sutter mentioned such bugs in part 1 or 2 of his talk, atomic<> Weapons: The C++ Memory Model and Modern Hardware", saying that it was already usually considered a compiler bug before C++11, but C++11 codified that and made it easier to be sure.
Or another recent example with ICC for x86:
Crash with icc: can the compiler invent writes where none existed in the abstract machine?
In the C++ abstract machine, there's no way for execution to reach either y = r1; or x = r2;, regardless of sequencing or simultaneity of the loads for the branch conditions. x and y both read as 0 and neither thread ever writes them.
No synchronization is required to avoid UB because no order of abstract-machine operations leads to a data-race. The ISO C++ standard doesn't have anything to say about speculative execution or what happens when mis-speculation reaches code. That's because speculation is a feature of real implementations, not of the abstract machine. It's up to implementations (HW vendors and compiler writers) to ensure the "as-if" rule is respected.
It's legal in C++ to write code like if (global_id == mine) shared_var = 123; and have all threads execute it, as long as at most one thread actually runs the shared_var = 123; statement. (And as long as synchronization exists to avoid a data race on non-atomic int global_id). If things like this broke down, it would be chaos. For example, you could apparently draw wrong conclusions like reordering atomic operations in C++
Observing that a non-write didn't happen isn't data-race UB.
It's also not UB to run if(i<SIZE) return arr[i]; because the array access only happens if i is in bounds.
I think the "out of the blue" value-invention note only applies to relaxed-atomics, apparently as a special caveat for them in the Atomics chapter. (And even then, AFAIK it can't actually happen on any real C++ implementations, certainly not mainstream ones. At this point implementations don't have to take any special measures to make sure it can't happen for non-atomic variables.)
I'm not aware of any similar language outside the atomics chapter of the standard that allows an implementation to allow values to appear out of the blue like this.
I don't see any sane way to argue that the C++ abstract machine causes UB at any point when executing this, but seeing r1 == r2 == 42 would imply that unsynchronized read+write had happened, but that's data-race UB. If that can happen, can an implementation invent UB because of speculative execution (or some other reason)? The answer has to be "no" for the C++ standard to be usable at all.
For relaxed atomics, inventing the 42 out of nowhere wouldn't imply that UB had happened; perhaps that's why the standard says it's allowed by the rules? As far as I know, nothing outside the Atomics chapter of the standard allows it.
A hypothetical asm / hardware mechanism that could cause this
(Nobody wants this, hopefully everyone agrees that it would be a bad idea to build hardware like this. It seems unlikely that coupling speculation across logical cores would ever be worth the downside of having to roll back all cores when one detects a mispredict or other mis-speculation.)
For 42 to be possible, thread 1 has to see thread 2's speculative store and the store from thread 1 has to be seen by thread 2's load. (Confirming that branch speculation as good, allowing this path of execution to become the real path that was actually taken.)
i.e. speculation across threads: Possible on current HW if they ran on the same core with only a lightweight context switch, e.g. coroutines or green threads.
But on current HW, memory reordering between threads is impossible in that case. Out-of-order execution of code on the same core gives the illusion of everything happening in program order. To get memory reordering between threads, they need to be running on different cores.
So we'd need a design that coupled together speculation between two logical cores. Nobody does that because it means more state needs to rollback if a mispredict is detected. But it is hypothetically possible. For example an OoO SMT core that allows store-forwarding between its logical cores even before they've retired from the out-of-order core (i.e. become non-speculative).
PowerPC allows store-forwarding between logical cores for retired stores, meaning that threads can disagree about the global order of stores. But waiting until they "graduate" (i.e. retire) and become non-speculative means it doesn't tie together speculation on separate logical cores. So when one is recovering from a branch miss, the others can keep the back-end busy. If they all had to rollback on a mispredict on any logical core, that would defeat a significant part of the benefit of SMT.
I thought for a while I'd found an ordering that lead to this on single core of a real weakly-ordered CPUs (with user-space context switching between the threads), but the final step store can't forward to the first step load because this is program order and OoO exec preserves that.
T2: r2 = y; stalls (e.g. cache miss)
T2: branch prediction predicts that r2 == 42 will be true. ( x = 42 should run.
T2: x = 42 runs. (Still speculative; r2 = yhasn't obtained a value yet so ther2 == 42` compare/branch is still waiting to confirm that speculation).
a context switch to Thread 1 happens without rolling back the CPU to retirement state or otherwise waiting for speculation to be confirmed as good or detected as mis-speculation.
This part won't happen on real C++ implementations unless they use an M:N thread model, not the more common 1:1 C++ thread to OS thread. Real CPUs don't rename the privilege level: they don't take interrupts or otherwise enter the kernel with speculative instructions in flight that might need to rollback and redo entering kernel mode from a different architectural state.
T1: r1 = x; takes its value from the speculative x = 42 store
T1: r1 == 42 is found to be true. (Branch speculation happens here, too, not actually waiting for store-forwarding to complete. But along this path of execution, where the x = 42 did happen, this branch condition will execute and confirm the prediction).
T1: y = 42 runs.
this was all on the same CPU core so this y=42 store is after the r2=y load in program-order; it can't give that load a 42 to let the r2==42 speculation be confirmed. So this possible ordering doesn't demonstrate this in action after all. This is why threads have to be running on separate cores with inter-thread speculation for effects like this to be possible.
Note that x = 42 doesn't have a data dependency on r2 so value-prediction isn't required to make this happen. And the y=r1 is inside an if(r1 == 42) anyway so the compiler can optimize to y=42 if it wants, breaking the data dependency in the other thread and making things symmetric.
Note that the arguments about Green Threads or other context switch on a single core isn't actually relevant: we need separate cores for the memory reordering.
I commented earlier that I thought this might involve value-prediction. The ISO C++ standard's memory model is certainly weak enough to allow the kinds of crazy "reordering" that value-prediction can create to use, but it's not necessary for this reordering. y=r1 can be optimized to y=42, and the original code includes x=42 anyway so there's no data dependency of that store on the r2=y load. Speculative stores of 42 are easily possible without value prediction. (The problem is getting the other thread to see them!)
Speculating because of branch prediction instead of value prediction has the same effect here. And in both cases the loads need to eventually see 42 to confirm the speculation as correct.
Value-prediction doesn't even help make this reordering more plausible. We still need inter-thread speculation and memory reordering for the two speculative stores to confirm each other and bootstrap themselves into existence.
ISO C++ chooses to allow this for relaxed atomics, but AFAICT is disallows this non-atomic variables. I'm not sure I see exactly what in the standard does allow the relaxed-atomic case in ISO C++ beyond the note saying it's not explicitly disallowed. If there was any other code that did anything with x or y then maybe, but I think my argument does apply to the relaxed atomic case as well. No path through the source in the C++ abstract machine can produce it.
As I said, it's not possible in practice AFAIK on any real hardware (in asm), or in C++ on any real C++ implementation. It's more of an interesting thought-experiment into crazy consequences of very weak ordering rules, like C++'s relaxed-atomic. (Those ordering rules don't disallow it, but I think the as-if rule and the rest of the standard does, unless there's some provision that allows relaxed atomics to read a value that was never actually written by any thread.)
If there is such a rule, it would only be for relaxed atomics, not for non-atomic variables. Data-race UB is pretty much all the standard needs to say about non-atomic vars and memory ordering, but we don't have that.
When a race condition potentially exists, what guarantees that a read of a shared variable (normal, non atomic) cannot see a write
There is no such guarantee.
When race condition exists, the behaviour of the program is undefined:
[intro.races]
Two actions are potentially concurrent if
they are performed by different threads, or
they are unsequenced, at least one is performed by a signal handler, and they are not both performed by the same signal handler invocation.
The execution of a program contains a data race if it contains two potentially concurrent conflicting actions, at least one of which is not atomic, and neither happens before the other, except for the special case for signal handlers described below. Any such data race results in undefined behavior. ...
The special case is not very relevant to the question, but I'll include it for completeness:
Two accesses to the same object of type volatile std::sig_atomic_t do not result in a data race if both occur in the same thread, even if one or more occurs in a signal handler. ...
What part of the so called "memory model" protects non atomic objects from these interactions caused by reads that see the interaction?
None. In fact, you get the opposite and the standard explicitly calls this out as undefined behavior. In [intro.races]\21 we have
The execution of a program contains a data race if it contains two potentially concurrent conflicting actions, at least one of which is not atomic, and neither happens before the other, except for the special case for signal handlers described below. Any such data race results in undefined behavior.
which covers your second example.
The rule is that if you have shared data in multiple threads, and at least one of those threads write to that shared data, then you need synchronization. Without that you have a data race and undefined behavior. Do note that volatile is not a valid synchronization mechanism. You need atomics/mutexs/condition variables to protect shared access.
Note: The specific examples I give here are apparently not accurate. I've assumed the optimizer can be somewhat more aggressive than it's apparently allowed to be. There is some excellent discussion about this in the comments. I'm going to have to investigate this further, but wanted to leave this note here as a warning.
Other people have given you answers quoting the appropriate parts of the standard that flat out state that the guarantee you think exists, doesn't. It appears that you're interpreting a part of the standard that says a certain weird behavior is permitted for atomic objects if you use memory_order_relaxed as meaning that this behavior is not permitted for non-atomic objects. This is a leap of inference that is explicitly addressed by other parts of the standard that declare the behavior undefined for non-atomic objects.
In practical terms, here is an order of events that might happen in thread 1 that would be perfectly reasonable, but result in the behavior you think is barred even if the hardware guaranteed that all memory access was completely serialized between CPUs. Keep in mind that the standard has to not only take into account the behavior of the hardware, but the behavior of optimizers, which often aggressively re-order and re-write code.
Thread 1 could be re-written by an optimizer to look this way:
old_y = y; // old_y is a hidden variable (perhaps a register) created by the optimizer
y = 42;
if (x != 42) y = old_y;
There might be perfectly reasonable reasons for an optimizer to do this. For example, it may decide that it's far more likely than not for 42 to be written into y, and for dependency reasons, the pipeline might work a lot better if the store into y occurs sooner rather than later.
The rule is that the apparent result must look as if the code you wrote is what was executed. But there is no requirement that the code you write bears any resemblance at all to what the CPU is actually told to do.
The atomic variables impose constraints on the ability of the compiler to re-write code as well as instructing the compiler to issue special CPU instructions that impose constraints on the ability of the CPU to re-order memory accesses. The constraints involving memory_order_relaxed are much stronger than what is ordinarily allowed. The compiler would generally be allowed to completely get rid of any reference to x and y at all if they weren't atomic.
Additionally, if they are atomic, the compiler must ensure that other CPUs see the entire variable as either with the new value or the old value. For example, if the variable is a 32-bit entity that crosses a cache line boundary and a modification involves changing bits on both sides of the cache line boundary, one CPU may see a value of the variable that is never written because it only sees an update to the bits on one side of the cache line boundary. But this is not allowed for atomic variables modified with memory_order_relaxed.
That is why data races are labeled as undefined behavior by the standard. The space of the possible things that could happen is probably a lot wilder than your imagination could account for, and certainly wider than any standard could reasonably encompass.
(Stackoverflow complains about too many comments I put above, so I gathered them into an answer with some modifications.)
The intercept you cite from from C++ standard working draft N3337 was wrong.
[Note: The requirements do allow r1 == r2 == 42 in the following
example, with x and y initially zero:
// Thread 1:
r1 = x.load(memory_order_relaxed);
if (r1 == 42)
y.store(r1, memory_order_relaxed);
// Thread 2:
r2 = y.load(memory_order_relaxed);
if (r2 == 42)
x.store(42, memory_order_relaxed);
A programming language should never allow this "r1 == r2 == 42" to happen.
This has nothing to do with memory model. This is required by causality, which is the basic logic methodology and the foundation of any programming language design. It is the fundamental contract between human and computer. Any memory model should abide by it. Otherwise it is a bug.
The causality here is reflected by the intra-thread dependences between operations within a thread, such as data dependence (e.g., read after write in same location) and control dependence (e.g., operation in a branch), etc. They cannot be violated by any language specification. Any compiler/processor design should respect the dependence in its committed result (i.e., externally visible result or program visible result).
Memory model is mainly about memory operation ordering among multi-processors, which should never violate the intra-thread dependence, although a weak model may allow the causality happening in one processor to be violated (or unseen) in another processor.
In your code snippet, both threads have (intra-thread) data dependence (load->check) and control dependence (check->store) that ensure their respective executions (within a thread) are ordered. That means, we can check the later op's output to determine if the earlier op has executed.
Then we can use simple logic to deduce that, if both r1 and r2 are 42, there must be a dependence cycle, which is impossible, unless you remove one condition check, which essentially breaks the dependence cycle. This has nothing to do with memory model, but intra-thread data dependence.
Causality (or more accurately, intra-thread dependence here) is defined in C++ std, but not so explicitly in early drafts, because dependence is more of micro-architecture and compiler terminology. In language spec, it is usually defined as operational semantics. For example, the control dependence formed by "if statement" is defined in the same version of draft you cited as "If the condition yields true the first substatement is executed. " That defines the sequential execution order.
That said, the compiler and processor can schedule one or more operations of the if-branch to be executed before the if-condition is resolved. But no matter how the compiler and processor schedule the operations, the result of the if-branch cannot be committed (i.e., become visible to the program) before the if-condition is resolved. One should distinguish between semantics requirement and implementation details. One is language spec, the other is how the compiler and processor implement the language spec.
Actually the current C++ standard draft has corrected this bug in https://timsong-cpp.github.io/cppwp/atomics.order#9 with a slight change.
[ Note: The recommendation similarly disallows r1 == r2 == 42 in the following example, with x and y again initially zero:
// Thread 1:
r1 = x.load(memory_order_relaxed);
if (r1 == 42)
y.store(42, memory_order_relaxed);
// Thread 2:
r2 = y.load(memory_order_relaxed);
if (r2 == 42)
x.store(42, memory_order_relaxed);
I've been studing 'Concurency in Action' position for some time and I have a problem with understanding following example of code (Listing 5.2):
#include <vector>
#include <atomic>
#include <iostream>
std::vector<int> data;
std::atomic<bool> data_ready(false);
void reader_thread()
{
while(!data_ready.load())
{
std::this_thread::sleep(std::milliseconds(1));
}
std::cout<<”The answer=”<<data[0]<<”\n”;
}
void writer_thread()
{
data.push_back(42); //write of data
data_ready=true; //write to data_ready flag
}
The book explaines:
(...) The write of the data happens-before the write to the data_ready
flag (...)
My concern is that the sentence does not cover the out-of-order execution. From my understanding out of order execution may happen when at least two instruction do not have depended operands. Taking this into account:
data_ready=true
does not need anything from
data.push_back(42)
to be executed. As a result of that it is not guaranteed that:
The write of the data happens-before the write to the data_ready flag
Is my understadning correct or there is something in out-of-order execution that I don't understand causing misunderstaning of given example?
EDIT
Thank you for answers, it was helpful. My misunserstanding was a result of not knowing that atomic types not only prevents from partialy channing a variable, but also acts as memory barrier.
For example following code may be reordered in many combinations by either compiler or processor:
d=0;
b=5;
a=10
c=1;
Resulting with following order (one of many possibilities):
b=5;
a=10
c=1;
d=0;
It it is not a problem with single-thread code since none of expressions have depended operands on other, but on multithreaded application may result of undefined behaviour. For example following code (initial values: x=0 and y=0):
Thread 1: Thread 2:
x=10; while(y!=15);
y=15; assert(x==10);
Without reordering of code by compiler or reordering execution by processor we could say: "Since assigement y=15 allways happens after assigement x=10 and assert happens after while loop the assert will never fail" But it's not true. The real execution order may be as below (one of many possible combinations):
Thread 1: Thread 2:
x=10; (4) while(y!=15); (3)
y=15; (1) assert(x==10); (2)
By default an atomic variable ensures sequentionally consistent ordering. If y in example above was atomic with memory_order_seq_cst default parameter following sentences are true:
- what happens before in thread 1 (x=10) it is also visible in thread 2 as happening before.
- what happens after while(y!=15) in thread 2 it is also visible in thread 1 as happening after
As a result of it assert will never fail.
Some of sources that may help with understaning:
Memory model synchronization modes - GCC
CppCon 2015: Michael Wong “C++11/14/17 atomics and memory
model..."
Memory barriers in C
I understand your concerns, but the code from book is fine. Every operation with atomics is by default memory_order_seq_cst, which means that everything that happened before the write in one of threads happens before read in the rest. You can imagine atomic operations with this std::memory_order like this:
std::atomic<bool> a;
//equivalent of a = true
a.assign_and_make_changes_from_thread_visible(true);
//equvalent of a.load()
a.get_value_and_changes_from_threads();
From Effective Modern C++, Item 40, it says "std::atomics imposes restrictions on how code can be reordered, and one such restriction is that no code that, in the source cod, precedes a write of std::atomic variable may take place afterwards." The note is this is true for when using sequential consistency which is a fair assumption.
I'm trying to demonstrate that it's very bad idea to not use std::atomic<>s but I can't manage to create an example that reproduces the failure. I have two threads and one of them does:
{
foobar = false;
}
and the other:
{
if (foobar) {
// ...
}
}
the type of foobar is either bool or std::atomic_bool and it's initialized to true. I'm using OS X Yosemite and even tried to use this trick to hint via CPU affinity that I want the threads to run on different cores. I run such operations in loops etc. and in any case, there's no observable difference in execution. I end up inspecting generated assembly with clang clang -std=c++11 -lstdc++ -O3 -S test.cpp and I see that the asm differences on read are minor (without atomic on left, with on right):
No mfence or something that "dramatic". On the write side, something more "dramatic" happens:
As you can see, the atomic<> version uses xchgb which uses an implicit lock. When I compile with a relatively old version of gcc (v4.5.2) I can see all sorts of mfences being added which also indicates there's a serious concern.
I kind of understand that "X86 implements a very strong memory model" (ref) and that mfences might not be necessary but does it mean that unless I want to write cross-platform code that e.g. supports ARM, I don't really need to put any atomic<>s unless I care for consistency at ns-level?
I've watched "atomic<> Weapons" from Herb Sutter but I'm still impressed with how difficult it is to create a simple example that reproduces those problems.
The big problem of data races is that they're undefined behavior, not guaranteed wrong behavior. And this, in conjunction with the the general unpredictability of threads and the strength of the x64 memory model, means that it gets really hard to create reproduceable failures.
A slightly more reliable failure mode is when the optimizer does unexpected things, because you can observe those in the assembly. Of course, the optimizer is notoriously finicky as well and might do something completely different if you change just one code line.
Here's an example failure that we had in our code at one point. The code implemented a sort of spin lock, but didn't use atomics.
bool operation_done;
void thread1() {
while (!operation_done) {
sleep();
}
// do something that depends on operation being done
}
void thread2() {
// do the operation
operation_done = true;
}
This worked fine in debug mode, but the release build got stuck. Debugging showed that execution of thread1 never left the loop, and looking at the assembly, we found that the condition was gone; the loop was simply infinite.
The problem was that the optimizer realized that under its memory model, operation_done could not possibly change within the loop (that would have been a data race), and thus it "knew" that once the condition was true once, it would be true forever.
Changing the type of operation_done to atomic_bool (or actually, a pre-C++11 compiler-specific equivalent) fixed the issue.
This is my own version of #Sebastian Redl's answer that fits the question more closely. I will still accept his for credit + kudos to #HansPassant for his comment which brought my attention back to writes which made everything clear - since as soon as I observed that the compiler was adding synchronization on writes, the problem turned to be that it wasn't optimizing bool as much as one would expect.
I was able to have a trivial program that reproduces the problem:
std::atomic_bool foobar(true);
//bool foobar = true;
long long cnt = 0;
long long loops = 400000000ll;
void thread_1() {
usleep(200000);
foobar = false;
}
void thread_2() {
while (loops--) {
if (foobar) {
++cnt;
}
}
std::cout << cnt << std::endl;
}
The main difference with my original code was that I used to have a usleep() inside the while loop. It was enough to prevent any optimizations within the while loop. The cleaner code above, yields the same asm for write:
but quite different for read:
We can see that in the bool case (left) clang brought the if (foobar) outside the loop. Thus when I run the bool case I get:
400000000
real 0m1.044s
user 0m1.032s
sys 0m0.005s
while when I run the atomic_bool case I get:
95393578
real 0m0.420s
user 0m0.414s
sys 0m0.003s
It's interesting that the atomic_bool case is faster - I guess because it does just 95 million incs on the counter contrary to 400 million in the bool case.
What is even more crazy-interesting though is this. If I move the std::cout << cnt << std::endl; out of the threaded code, after pthread_join(), the loop in the non-atomic case becomes just this:
i.e. there's no loop. It's just if (foobar!=0) cnt = loops;! Clever clang. Then the execution yields:
400000000
real 0m0.206s
user 0m0.001s
sys 0m0.002s
while the atomic_bool remains the same.
So more than enough evidence that we should use atomics. The only thing to remember is - don't put any usleep() on your benchmarks because even if it's small, it will prevent quite a few compiler optimizations.
In general, it is very rare that the use of atomic types actually does anything useful for you in multithreaded situations. It is more useful to implement things like mutexes, semaphores and so on.
One reason why it's not very useful: As soon as you have two values that both need to be changed in an atomic way, you are absolutely stuck. You can't do it with atomic values. And it's quite rare that I want to change a single value in an atomic way.
In iOS and MacOS X, the three methods to use are: Protecting the change using #synchronized. Avoiding multi-threaded access by running code on a sequential queue (may be the main queue). Using mutexes.
I hope you are aware that atomicity for boolean values is rather pointless. What you have is a race condition: One thread stores a value, another reads it. Atomicity doesn't make a difference here. It makes (or might make) a difference if two threads accessing a variable at exactly the same time causes problems. For example, if a variable is incremented on two threads at exactly the same time, is it guaranteed that the final result is increased by two? That requires atomicity (or one of the methods mentioned earlier).
Sebastian makes the ridiculous claim that atomicity fixes the data race: That's nonsense. In a data race, a reader will read a value either before or after it is changed, whether that value is atomic or not doesn't make any difference whatsoever. The reader will read the old value or the new value, so the behaviour is unpredictable. All that atomicity does is prevent the situation that the reader would read some in-between state. Which doesn't fix the data race.
I have come across C++03 some code that takes this form:
struct Foo {
int a;
int b;
CRITICAL_SECTION cs;
}
// DoFoo::Foo foo_;
void DoFoo::Foolish()
{
if( foo_.a == 4 )
{
PerformSomeTask();
EnterCriticalSection(&foo_.cs);
foo_.b = 7;
LeaveCriticalSection(&foo_.cs);
}
}
Does the read from foo_.a need to be protected? e.g.:
void DoFoo::Foolish()
{
EnterCriticalSection(&foo_.cs);
int a = foo_.a;
LeaveCriticalSection(&foo_.cs);
if( a == 4 )
{
PerformSomeTask();
EnterCriticalSection(&foo_.cs);
foo_.b = 7;
LeaveCriticalSection(&foo_.cs);
}
}
If so, why?
Please assume the integers are 32-bit aligned. The platform is ARM.
Technically yes, but no on many platforms. First, let us assume that int is 32 bits (which is pretty common, but not nearly universal).
It is possible that the two words (16 bit parts) of a 32 bit int will be read or written to separately. On some systems, they will be read separately if the int isn't aligned properly.
Imagine a system where you can only do 32-bit aligned 32 bit reads and writes (and 16-bit aligned 16 bit reads and writes), and an int that straddles such a boundary. Initially the int is zero (ie, 0x00000000)
One thread writes 0xBAADF00D to the int, the other reads it "at the same time".
The writing thread first writes 0xBAAD to the high word of the int. The reader thread then reads the entire int (both high and low) getting 0xBAAD0000 -- which is a state that the int was never put into on purpose!
The writer thread then writes the low word 0xF00D.
As noted, on some platforms all 32 bit reads/writes are atomic, so this isn't a concern. There are other concerns, however.
Most lock/unlock code includes instructions to the compiler to prevent reordering across the lock. Without that prevention of reordering, the compiler is free to reorder things so long as it behaves "as-if" in a single threaded context it would have worked that way. So if you read a then b in code, the compiler could read b before it reads a, so long as it doesn't see an in-thread opportunity for b to be modified in that interval.
So possibly the code you are reading is using these locks to make sure that the read of the variable happens in the order written in the code.
Other issues are raised in the comments below, but I don't feel competent to address them: cache issues, and visibility.
Looking at this it seems that arm has quite relaxed memory model so you need a form of memory barrier to ensure that writes in one thread are visible when you'd expect them in another thread. So what you are doing, or else using std::atomic seems likely necessary on your platform. Unless you take this into account you can see updates out of order in different threads which would break your example.
I think you can use C++11 to ensure that integer reads are atomic, using (for example) std::atomic<int>.
The C++ standard says that there's a data race if one thread writes to a variable at the same time as another thread reads from that variable, or if two threads write to the same variable at the same time. It further says that a data race produces undefined behavior. So, formally, you must synchronize those reads and writes.
There are three separate issues when one thread reads data that was written by another thread. First, there is tearing: if writing requires more than a single bus cycle, it's possible for a thread switch to occur in the middle of the operation, and another thread could see a half-written value; there's an analogous problem if a read requires more than a single bus cycle. Second, there's visibility: each processor has its own local copy of the data that it's been working on recently, and writing to one processor's cache does not necessarily update another processor's cache. Third, there's compiler optimizations that reorder reads and writes in ways that would be okay within a single thread, but will break multi-threaded code. Thread-safe code has to deal with all three problems. That's the job of synchronization primitives: mutexes, condition variables, and atomics.
Although the integer read/write operation indeed will most likely be atomic, the compiler optimizations and processor cache will still give you problems if you don't do it properly.
To explain - the compiler will normally assume that the code is single-threaded and make many optimizations that rely on that. For example, it might change the order of instructions. Or, if it sees that the variable is only written and never read, it might optimize it away entirely.
The CPU will also cache that integer, so if one thread writes it, the other one might not get to see it until a lot later.
There are two things you can do. One is to wrap in in critical section like in your original code. The other is to mark the variable as volatile. That will signal the compiler that this variable will be accessed by multiple threads and will disable a range of optimizations, as well as placing special cache-sync instructions (aka "memory barriers") around accesses to the variable (or so I understand). Apparently this is wrong.
Added: Also, as noted by another answer, Windows has Interlocked APIs that can be used to avoid these issues for non-volatile variables.