Summary: Redo of D29536635 (https://github.com/facebook/folly/commit/bc0818f89e7ab3802cdac95a25e29f4061021c3c) but rebased on top of `UniqueInstance.h` fixes.

Reviewed By: yfeldblum

Differential Revision: D29630206

fbshipit-source-id: ed103849edd6edd3f9c4ef75368c2fa1b038182b

Summary: The class depends on RTTI. It's a sanity check that crashes if two instances of a singleton are created. So doing nothing in `-fno-rtti` code is fine.

Reviewed By: yfeldblum

Differential Revision: D29630207

fbshipit-source-id: 4976f72394086afe12e224e40fe955636108e713

Summary: Proves the internal build rule replaces `memcpy`.

Reviewed By: LoganEvans

Differential Revision: D29665797

fbshipit-source-id: 7f3a8550853261de94cedf9545b6445d11297b2b

Reviewed By: zertosh

Differential Revision: D29656934

fbshipit-source-id: c40bbc8e4512b145050ee47db2c8dc781f3c36e9

Summary:
A bunch of existing projects vendor it into their own CMake folder. But these
same projects also have fbcode_builder's CMake directory in their CMAKE_MODULE_PATH,
so move FindZstd here.

Differential Revision: D29637686

fbshipit-source-id: 805676e18f98ef217dea8511d039edc38771b529

Summary:
I don't understand why we have both manifests/specs; I was assuming one was generated from the other,
but I don't see a at-generated and grep doesn't show up anything, so both are manual edits.

Reviewed By: iahs

Differential Revision: D29524000

fbshipit-source-id: 5f6df62f0162ea24a9903bedf5d220eab5c2dff1

Reviewed By: zertosh

Differential Revision: D29612107

fbshipit-source-id: ac450058134e23a3831db35d2e49c80eb8cde36a

Differential Revision:
D29536635 (https://github.com/facebook/folly/commit/bc0818f89e7ab3802cdac95a25e29f4061021c3c)

Original commit changeset: 008a612ec65a

fbshipit-source-id: 2c18348732f1b1f8022e3cc8dce5a5684ce16108

Summary: AsyncSocketSendmmsgIntegrationTest.PingPongRequest fails during stress runs with multiple jobs due to the 5ms timeout in `UDPClient::sendPing()`. Increasing the timeout fixes the issue.

Reviewed By: yfeldblum

Differential Revision: D29623870

fbshipit-source-id: ad6ee0356001fbded9ce01ddc099b708d1477899

Summary: `RequestContext::StaticContextAccessor` acts as a guard, preventing all threads with a `StaticContext` from being destroyed (or created).

Reviewed By: yfeldblum

Differential Revision: D29536635

fbshipit-source-id: 008a612ec65aa7095450f4d0ab522e8ecd864548

Summary:
Codegen for copy and move assignment is suboptimal when the vectors involved are inline, in particular if the destination is default-constructed (which is very common) and even more so if the value type is trivially copyable.

The main optimization is that if the storage is inline and the type is trivially-copyable, we can just copy the whole storage, regardless of the size of the container. This avoids a branchy loop, and it's usually just a handful of MOVs.

While at it, optimized all the related code paths by avoiding delegating to `swap()` and `assign()` when possible, which introduce further branches that are hard to optimize. Also move and copy in place when the vector is non-empty, avoiding unnecessary destructions.

Also fixed a couple of bugs:
- The move constructor did not clear its argument when inline, inconsistently with the move assignment and `std::vector`
- The move assignment operator passed its heap storage to the argument instead of freeing it.

Reviewed By: yfeldblum

Differential Revision: D29592519

fbshipit-source-id: 6281cdda775568d28619afea8b7ca2fb168c7e5d

Summary: It is unused. An upcoming Concurrency TS will have a replacement.

Reviewed By: mshneer

Differential Revision: D29597486

fbshipit-source-id: a108b945ce32eb17cedefad89630c9171fc5c9c2

Summary:
BSD readlink doesn't have -f.

I'm not using TravisCI, however docker is still convenient for reproducing builds locally.

Reviewed By: yns88

Differential Revision: D29523333

fbshipit-source-id: e01169f3eabca7b8baec95bc70fe119cad201b35

Summary: Implement C++20 `contains()`.

Reviewed By: yfeldblum

Differential Revision: D29556436

fbshipit-source-id: 1c3a8e6e9a07f6e47491b063f88df6bf9da2d87b

Reviewed By: yfeldblum, mshneer

Differential Revision: D29578762

fbshipit-source-id: 2aa9af8f68825997db8b71fc97dda5477fa57413

Summary:
[Forward iterators are required to be default constructible](https://en.cppreference.com/w/cpp/named_req/ForwardIterator). So as of today, `IteratorAdaptor` is lying!

We should default construct the wrapped iterator into a [singular iterator](https://eel.is/c++draft/iterator.requirements#general-7):
> Iterators can also have singular values that are not associated with any sequence. Results of most expressions are undefined for singular values; the only exceptions are destroying an iterator that holds a singular value, the assignment of a non-singular value to an iterator that holds a singular value, and, for iterators that meet the `Cpp17DefaultConstructible` requirements, using a value-initialized iterator as the source of a copy or move operation.

Reviewed By: yfeldblum, mshneer

Differential Revision: D29578765

fbshipit-source-id: 7f2f0762b17f0b1a056532fc5db2abdd76cca3ea

Summary: Let's give them better names than `first` and `second`.

Reviewed By: yfeldblum, mshneer

Differential Revision: D29536636

fbshipit-source-id: bbb37f2e1c0c51dd1ce2a0cfeca85399f409c415

Summary: Fix BatchSemaphore to post baton for a multiple waiter in a list as long as there are token available.

Reviewed By: yfeldblum

Differential Revision: D29430799

fbshipit-source-id: 98b0a616d0ce863108dcf331e491fd2cc12429d1

Summary:
Define a single implementation type `lock_base` which handles all cases, including unique/shared/upgrade locks and including sans-state/with-state locks. Add `unique_lock_base`, `shared_lock_base`, and `upgrade_lock_base`.

Revise `upgrade_lock` simply to derive `upgrade_lock_base`. We may use `upgrade_lock` as an example for specializing `unique_lock` and `shared_lock`

Remove `ProxyLockableUniqueLock` since `unique_lock_base` absorbs it. Let the `unique_lock` specializations for `DistributedMutex` inherit `unique_lock_base` instead.

Add lock invokers for every lock, try-lock, unlock, and lock-transition member. Were these used only internally to implement the lock-policy types and the lock-transition functions they might be left in detail but there may be broader use-cases for at least some of them. Putting the invokers in this header is consistent with proper placement since this header is intended to own all lock primitives and facilities.

Revise the lock-transition functions to handle cases where the from and to lock types are sans-state/with-state. Expand the set of lock-transition functions for completeness: lock-transitions include x->s, x->u, u->s, u->x; while try-lock-transitions include s->x, s->u, u->x.

Reviewed By: aary

Differential Revision: D28767313

fbshipit-source-id: 153adc8270f0f4338db6acf544b8d358556d6f49

Summary:
Reenables dynamicconfig loading with eden backingstore. Previously it
broke edenfs-windows-release, but we believe the
opensource/fbcode_builder/manifests/eden tweak has fixed it.

Reviewed By: xavierd

Differential Revision: D29561192

fbshipit-source-id: 775dd21d177f3baa09b0192e7d3f7231008c3766

Differential Revision: D29497151

fbshipit-source-id: 64ad1adbd68d10066fc65ddc41e9cff5ef3c6b53

Summary:
For whatever reason, sc_testpilot default to --return-zero-on-failures, which
means that test failures are simply not reported properly to the signal box.
Fix that by explicitely passing --return-nonzero-on-failures

Reviewed By: fanzeyi

Differential Revision: D29509158

fbshipit-source-id: ef991f91df48e99769f96ffd8d7015cdf507c723

Summary: Import or backport `std::reinterpret_pointer_cast` into folly. The relevant guard is the use of the C++17 language and, if feature-test macros are exported, `__cpp_lib_shared_ptr_arrays >= 201611`.

Reviewed By: yfeldblum

Differential Revision: D29433489

fbshipit-source-id: 92596d05f5057cff4c65283711b4ed6778d20758

Summary: Pull Request resolved: https://github.com/facebook/folly/pull/1614

Reviewed By: yfeldblum

Differential Revision: D29496725

fbshipit-source-id: 7c726e3e310eb28cf33603ebd83df6d832488369

Summary:
`guard()` is ignoring the return from `handler()`, thereby breaking `take()`.

(Note: this ignores all push blocking failures!)

Reviewed By: yfeldblum

Differential Revision: D29416513

fbshipit-source-id: 1e84d3ba53cb68a1664a349e71398c782b1df751

Summary: A channel is a sender and receiver pair that allows one component to send values to another. A sender and receiver pair is similar to an AsyncPipe and AsyncGenerator pair. However, unlike AsyncPipe/AsyncGenerator, senders and receivers can be used by higher level transformation abstractions that are much more memory-efficient than using AsyncGenerator directly. These higher level abstractions do not require long-lived coroutine frames to wait on incoming values.

Reviewed By: aary

Differential Revision: D29158424

fbshipit-source-id: 88c51d0f9d73677a04906197f4c44fe84ac01cdb

Summary: Pull Request resolved: https://github.com/facebook/folly/pull/1610

Reviewed By: yfeldblum

Differential Revision: D29434206

Pulled By: Orvid

fbshipit-source-id: e20d29d0572307e917f8673bb161e3af1a6c55c3

Summary:
Added options to enforce ALPN when both client and support support ALPN for
folly/openssl.

Reviewed By: knekritz

Differential Revision: D29298491

fbshipit-source-id: acdd6001fea89606e2438640a4434cc56454f1aa

Summary: `FixedString` should be constructable from a `std::string_view`. This allows passing `std::string_view` as key to `.emplace` in heterogeneous maps where the key is a `FixedString`.

Reviewed By: yfeldblum

Differential Revision: D29456948

fbshipit-source-id: 4d2e428c7de05c1dc34b50727f8107aec915b632

Summary:
The approach taken to fixing the lock inversion is not to use locks. Instead, we take a page from the existing futures playbook and use an atomic acyclic finite state machine! In this case, we have a single atomic pointer which may own either a raised object of type `exception_wrapper` or an interrupt-handler. Both are at least 4-byte aligned (8-byte aligned on 64-bit architectures) so both have the bottom two bits free. We use the bottom two bits as the state machine to identify whether `0x0u` initial where nothing is owned, `0x1u` where an interrupt-handler is owned, `0x2u` where a raised object is owned, or `0x3u` terminal where both an interrupt-handler and a raised object have been set and the handler has been invoked on the object.

As choices, we forbid repeated calls to set the interrupt handler but permit repeated calls to raise an object. Calls after the first to set an interrupt handler terminate while calls to raise an object after the first are ignored. Existing tests demonstrate raising twice so we may like to be cautious about breaking that behavior.

Some semantics are changed. Raised objects and interrupt handlers are destroyed earlier than before: they are now destroyed immediately after invoking handlers on objects.

The lock order inversion is observed by thread sanitizer (tsan) in the new shared-promise test.

Differential Revision: D29250149

fbshipit-source-id: 63e257d03cebbf8dba95a514b7e89680cae569a7

Summary:
Make `Skipper` accept the `shared_ptr` by value, like `Accessor` does, so we can avoid the copy.

Also expose the accessor, so after we handed over the `shared_ptr` we can still retain access to the list.

Reviewed By: philippv

Differential Revision: D29464861

fbshipit-source-id: 75e93f0e72c4b8a2635d25e5e7860f27aec2a0a5

Summary: Hack to use parexec and the par directly.

Reviewed By: chadaustin

Differential Revision: D29461444

fbshipit-source-id: 9ac6c1aa43728782b8de0a71446109f7fd5dab1d

Summary: It's already not async-signal-safe so just use the fast version of getting the unsymbolized stack traces.

Reviewed By: chadaustin

Differential Revision: D29439597

fbshipit-source-id: 5a727d6a1c37a1c29bce84d866bf4863774c5ff1

Reviewed By: Mizuchi

Differential Revision: D29329759

fbshipit-source-id: 588aa9ce11db2c1aea51dcf96e0de3b50633fa29

Summary:
This diff adds a computeChainCapacity method to IOBuf.

This method will loop through each chain of an IOBuf and return the total capacity of all IOBuf chains. This method is used very similarly to computeChainDataLength(), except rather than returning total data length, it returns total capacity.

Reviewed By: yfeldblum

Differential Revision: D28875393

fbshipit-source-id: 7e5f4a93916fa6c30eef8f8ee8c346a4537302af

Summary: If a callback associated with an observer depends on yet another observer, the callback may be called more than once per distinct snapshot of the associated observer. This diff fixes this problem.

Reviewed By: andriigrynenko

Differential Revision: D29313796

fbshipit-source-id: 345bfa0a1ac9dfaee806be21256088d4667aef15

Summary:
To support "clearing" a ring-buffer without violating TurnSequencer's
invariants, allow storing a Cursor and comparing them in a loop.

Differential Revision: D29345250

fbshipit-source-id: d1d3bbd0b38489690334bdd63a7f40d9f6bad6c6

Summary: The CancellationSource already provides this synchronization

Reviewed By: yfeldblum

Differential Revision: D29368728

fbshipit-source-id: 085467d1fcb70d6fc059174ec673f7825025c076

Summary: The empty size of UnboundedQueue is hundreds of bytes, which is a problem for many users of AsyncPipe. Switch the default to a smaller queue, adding a template parameter so high-throughput users can switch back.

Differential Revision: D29375068

fbshipit-source-id: 9fb2dad81697eeb70d58929e6b9c1c64102aa4a8

Summary: A smaller version of coro::UnboundedQueue to make AsyncPipe usable in memory-constrained programs.

Reviewed By: yfeldblum

Differential Revision: D29375067

fbshipit-source-id: 5d27e4514312d2748055be764af88c0688d46065