This page is a snapshot from the LWG issues list, see the Library Active Issues List for more information and the meaning of Tentatively NAD status.

4309. How are two seq_cst operations ordered in the single total order if these two operations are unsequenced?

Section: 32.5.4 [atomics.order] Status: Tentatively NAD Submitter: jim x Opened: 2025-08-06 Last modified: 2025-10-23

Priority: Not Prioritized

View other active issues in [atomics.order].

View all other issues in [atomics.order].

View all issues with Tentatively NAD status.

Discussion:

Consider this snippet code:

std::atomic<int> flag = {0};
std::atomic<int> turn = {0};

if(flag + turn){}

The loads of flag and turn as the operands of + are unsequenced according to [intro.execution] p10.

Except where noted, evaluations of operands of individual operators and of subexpressions of individual expressions are unsequenced.

However, 32.5.4 [atomics.order] p4 says:

There is a single total order S on all memory_order::seq_cst operations, including fences, that satisfies the following constraints.

Then, it says that:

First, if A and B are memory_order::seq_cst operations and A strongly happens before B, then A precedes B in S.

According to the first sentence, the load of flag and the load of turn do have an order in the single total order S since they are both memory_order::seq_cst operations. However, since they are unsequenced, the second sentence does not apply to them. So, what's the order of them in S? Is the order of them in S unspecified?

Suggested Resolution:

We may want to say the order of such operations is indeterminate in the single total order. That is, either A precedes B or B precedes A, but it is unspecified which.

[2025-10-23; Reflector poll; Status changed: New → Tentatively NAD.]

The loads are both done within functions calls, so are at worst indeterminately sequenced. So we already say what the suggested resolution wants.

Proposed resolution: