P1478 Byte-wise atomic memcpy
Closed this issue · 16 comments
We want another name composed from some combination of {atomic, byte-wise,
race-free, read/write}_memcpy with a veto if it's bad.
-Unanimous consent.
With changes to the wording requested in the minutes,
we forward P1478R2 to LEWG for Concurrency TS 2:
SF F N A SA
3 10 0 0 0
Unanimous consent
P1478R3: Byte-wise atomic memcpy
2020-06-01 Library Evolution Telecon Minutes
Chair: Billy Baker
Champion: Hans Boehm
Minute Taker: Inbal Levi
Start: 2020-06-01 08:02
POLL: P1478 (Byte-wise atomic memcpy) should include a type-safe (std::copy analog) version
| SF | F | N | A | SA |
|---|---|---|---|---|
| 1 | 3 | 8 | 2 | 1 |
Attendance: 24
That has no consensus.
POLL: The design of P1478 (Byte-wise atomic memcpy) is acceptable for continued Library Evolution time.
| SF | F | N | A | SA |
|---|---|---|---|---|
| 5 | 11 | 0 | 0 | 0 |
Attendance: 24
We have unanimous consent for that.
End: 2020-06-01 09:03
SUMMARY: This paper proposes a byte-wise atomic memcpy interface. Library Evolution feels this is a useful feature worth pursuing. Much of our discussion centered around whether the interface should be std::memcpy-like (e.g. byte-centric) or std::copy like (which provides better type safety). There was a related discussion about possible C compatibility, which would need a std::memcpy-like interface. In the end, we did not have consensus for adding a std::copy-like interface to the proposal, because this is a lower-level feature. We need to see and review a version of the proposal with wording before this can proceed further.
OUTCOME: Bring a revision of P1478R3 (Byte-wise atomic memcpy), with the guidance below, to Library Evolution for further design review.
- Add wording including a synopsis.
- Clarify that this feature should live in
<atomic> - Clarify
atomic_(load|store)_per_byte_memcpyshould returnvoid. - Make the source parameters pointers-to-const.
- Explore C compatibility with the Standard C committee.
P1478R5: Byte-wise atomic memcpy
2020-11-30 Library Evolution Telecon Minutes
Chair: Nevin Liber
Champion: Hans Boehm
Minute Taker: Inbal Levi
POLL: Forward P1478R6 (Byte-wise atomic memcpy) to electronic polling for sending to LWG for the inclusion in the Concurrency TS2, pending initial wording approval by Tomasz Kaminski.
| Strongly Favor | Weakly Favor | Neutral | Weakly Against | Strongly Against |
|---|---|---|---|---|
| 6 | 9 | 0 | 1 | 0 |
Attendance: 25
WA: Since it’s a library addition, I would like to see implementation with useful use case
Author's Position: SF
Outcome: Consensus for forwarding to electric poll.
2021-01-25 Library Evolution Telecon
P1478R6: Byte-wise atomic memcpy
2021-01-25 Library Evolution Telecon Minutes
Chair:
Champion:
Minute Taker:
Outcome
We will proceed with electronic balloting of P1478R6 (Byte-wise atomic memcpy) as planned.
2021 Winter Library Evolution Polls
P2333: 2021 Winter Library Evolution Poll Outcomes
POLL 9: Send P1478R6 (Byte-wise Atomic memcpy) to LWG for the Concurrency TS v2, classified as an addition (P0592R4 bucket 3 item).
| Strongly Favor | Weakly Favor | Neutral | Weakly Against | Strongly Against |
|---|---|---|---|---|
| 6 | 11 | 2 | 1 | 0 |
Outcome: Consensus in favor.
Outcome
Send P1478R6 (Byte-wise Atomic memcpy) to LWG for the Concurrency TS v2, classified as an addition (P0592R4 bucket 3 item).
Late status update:
LWG saw this paper in Kona and approved for concurrency TS2
https://wiki.edg.com/bin/view/Wg21kona2022/LWG20221109-EA
| F | A | N |
|---|---|---|
| 8 | 0 | 1 |
This was moved at plenary and approved:
This was voted into the Concurrency v2 working paper in Kona 2022, and is present in https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2023/n4953.pdf