Summary:
Use `std::unique_lock`, `std::shared_lock`, and `folly::upgrade_lock`. There are two reasons:

* Makes generic the use of `std::unique_lock` with `std::mutex`, which is currently special-cased.
* Permits specializations of `std::unique_lock` and the other lock types to be found automatically.

In particular, this permits the use of `Synchronized<T, DistributedMutex>`, which is only proxy-lockable and not lockable.

Reviewed By: simpkins

Differential Revision: D28705607

fbshipit-source-id: 48daa2910ce16ee4fde6f5ea629a41d9768f3c87

Summary: They were used as extension points at one time, but no longer.

Reviewed By: Alfus

Differential Revision: D28987212

fbshipit-source-id: e9d59e5cf9641323657314b088eef516ce068112

Summary:
```
auto x = std::atomic<std::uint64_t>{0};
auto y = std::atomic<std::uint64_t>{0};

// thread 1
x.store(1, std::memory_order_release);
auto one = y.load(std::memory_order_seq_cst);

// thread 2
y.fetch_add(1, std::memory_order_seq_cst);
auto two = x.load(std::memory_order_seq_cst);
```
Here it is possible for both `one` and `two` to end up with the value `0`. The
code in ParkingLot assumed that this would not be possible; and the counter used
to track the number of waiters could get reordered with respect to loads around
it.  This diff adds a seq_cst fence to ensure unparking threads always sequence
their stores before parking _before_ the counter load globally.

Reviewed By: yfeldblum, ot

Differential Revision: D28972810

fbshipit-source-id: 06eb6a2e6df6b00bf07ac8454a79257a5276e154

Summary: To observe how the compiler generates corresponding code.

Reviewed By: Alfus

Differential Revision: D28984027

fbshipit-source-id: d1c86197931aad257eb922cec9810c71ecdfc20a

Summary:
The numeric functions `less_than` and `greater_than` are intended for integer use but, technically, they are not constrained to integer use. If they are used with floating-point types, `min()` does not do the expected thing so use use `lowest()`.

Fixes: https://github.com/facebook/folly/issues/1604.

Reviewed By: iahs

Differential Revision: D29069620

fbshipit-source-id: 369bd59338b889cb1ec0f56d232a3775500573d0

Summary:
Ref: https://abseil.io/tips/143

Pull Request resolved: https://github.com/facebook/folly/pull/1603

Differential Revision: D29034248

Pulled By: yfeldblum

fbshipit-source-id: 87ae1970eab3f067d71a480fc7a95b18e2041c6d

Summary: Use the global immutable executor by default for SerialExecutor.

Reviewed By: yfeldblum

Differential Revision: D28925750

fbshipit-source-id: 91f75cfb3a4880098d933fe1f148d5c3b2e896e7

Summary:
Remove unused HazptrDomain data members unprotected_ and children_ and the function reclaim_unprotected_safe.

The data members and function were used before the change in October 2020 that eliminated the nesting of synchronous reclamation within asynchronous reclamation.

Reviewed By: yfeldblum

Differential Revision: D29017460

fbshipit-source-id: 645a61aedc801cb3eb14a4c3a085fea8b8422f1e

Summary: Remove all uses of the default SerialExecutor by changing callsites to be explicit, and removing the default parameter. This will allow us to change the default to a safer option in a subsequent diff.

Reviewed By: yfeldblum

Differential Revision: D28842180

fbshipit-source-id: 93027dcf8b19c44380534dabd731651780dac90e

Summary: io_uring SendmsgRecvmsg test fixes

Reviewed By: danobi

Differential Revision: D29003163

fbshipit-source-id: c1139a67d7b687d0eab21be7c6329f593dbe6ea9

Summary: A function like keep_sink but marked `noexcept`. Useful for shrinking check functions with duplicated cleanups.

Reviewed By: aary

Differential Revision: D28984340

fbshipit-source-id: f1099d43db25492db0e633f8cc8a5d44bcf157bc

Reviewed By: yfeldblum

Differential Revision: D28995271

fbshipit-source-id: bd64f4f98742fbad801f3db15a9cb8f5f71fe4b1

Summary:
The testpilot command line getdeps generated was not suitable for tpx
when filtering. tpx would consider the passed test filter an
environment variable.

Reorder a few things to at least make filtering work.

Reviewed By: fanzeyi

Differential Revision: D28976061

fbshipit-source-id: 21c45b3a4a59711a2a5da36a1bd11b7b4871ec5d

Summary:
Android NDK bionic with FORTIFY will override original `open()` definition and making folly `wrapNoInt` template failed to deduct.
The issue may happen only after NDK r21 because [this commit landed after r21](https://android.googlesource.com/platform/bionic/+/9349b9e51b41d12fd054b925802b626ca2db0afb%5E%21/#F0)

References:
https://github.com/android/ndk/issues/1328
https://github.com/llvm/llvm-project/commit/0a0e411204a2baa520fd73a8d69b664f98b428ba

Pull Request resolved: https://github.com/facebook/folly/pull/1593

Test Plan:
Tested running `objdump -dr` on the object file generated in both `mode/opt`
and `mode/opt-gcc` build modes and confirmed the generated code was
identical.

Reviewed By: yfeldblum

Differential Revision: D28953120

Pulled By: simpkins

fbshipit-source-id: 225583a5a011e8456592a0bcfcd669fe966ea6af

Summary: Introduces a CancellableAsyncScope type that automatically adds a CancellationToken to every task added() on the scope.

Reviewed By: yfeldblum

Differential Revision: D28438228

fbshipit-source-id: 2416725360c16f8d95dd5c35997dd624278d0980

Summary: Add a specialization of `folly::hasher<std::string_view>` parallel to the specialilzation of `folly::hasher<std::string>`.

Reviewed By: luciang

Differential Revision: D28921343

fbshipit-source-id: 77213d9ff66cec57b6d36b214c41e479e1d9455d

Summary: We need a thread-safe alternative to `LOG_FIRST_N(severity, n)` -- specifically when n = 1, the most common case.

Reviewed By: yfeldblum, ot, luciang

Differential Revision: D28884240

fbshipit-source-id: 41c937bf37dba036b71f1c62fd70445e79a3924a

Summary: Since we're moving towards enabling TLS 1.3 by default in folly, add this version to the SSLVersion enum.

Reviewed By: yfeldblum

Differential Revision: D28908538

fbshipit-source-id: 19d9ee248fff4682eec59c6f3d699fc6c9fa1917

Summary: There are many possible knobs that may be tweaked but long chains of positional params are awkward. Extract a policy param to house many of these chained params. Facilitates adding more params in the future as well.

Reviewed By: ot, luciang

Differential Revision: D28824188

fbshipit-source-id: f56ec15242b148890eced06a3dd101d7cfaaefb1

Summary:
Now that `exception_wrapper` can access the exception object directly from the `std::exception_ptr`, it no longer needs to be passed the exception object separately.

This eliminates duplicative `catch (exception const&)` clauses which primarily construct `exception_wrapper` instances, reducing build artifact size.

Reviewed By: akrieger

Differential Revision: D27888763

fbshipit-source-id: 5f60a07083b3c8d818f0eafd5b17afcfca239ab8

Summary: The helper `catch_exception` may be used to wrap the `try {} catch (...) {}` syntax.

Reviewed By: Orvid, luciang

Differential Revision: D28276166

fbshipit-source-id: be903215251d8ea3321bd6d3b994b19359056f8c

Summary: Avoid questions of what may the mutex unlock method do to the passed state: nothing.

Reviewed By: aary

Differential Revision: D28895552

fbshipit-source-id: 8b73e76711bb553a344d958c80dcf14d56e74ffd

Summary:
Remove all the conditional compilations that were used for the googletest upgrade. We're set on Googletest 1.10.x now, so no need to keep the old code lying around.

TODO in the future, is to deprecate some of the other fb-only patches, and to switch over fbcode to use fbsource/third-party

Reviewed By: yfeldblum, igorsugak

Differential Revision: D28860854

fbshipit-source-id: 8ffbaf860fca8897f6691218bc0154723ec954d4

Summary: Under gcc8 in some usages, the compiler emits a stack copy of the proxy state but passes `7` instead of the address of the stack copy. Looks like incorrect codegen. Work around that bug by not having a stack copy.

Reviewed By: aary

Differential Revision: D28890467

fbshipit-source-id: 50b16464708c7ef3e360618f3d738264f59cf8a6

Summary: There is no longer compelling reason to keep this overload since it is relatively cheap to inspect any `exception_wrapper` or `exception_ptr`, no matter from which exception object type it is constructed.

Reviewed By: luciang

Differential Revision: D26877547

fbshipit-source-id: d9a38c9a56a47b9453eda2d9e8abf7586c8aab61

Summary:
Switch getReadBuffers to be IOBufVecQueue::IoVecVec based

(Note: this ignores all push blocking failures!)

Reviewed By: simpkins

Differential Revision: D28467957

fbshipit-source-id: 637bf07d516120cb2bab89c3cbc28f58b700748a

Summary: Universe upper bound is inclusive, so we need to make sure that we have a pointer for `upperBitsUniverse / skipQuantum`. The bug has currently no effect as we use a tighter upper bound by reading the end of the list on construction, but it is necessary to fix this if we want to avoid doing that.

Reviewed By: yfeldblum, philippv, luciang

Differential Revision: D28886633

fbshipit-source-id: ba53fcc3f2ae1d52fe8e39576ad930754282ddd9

Summary:
Add folly::IOBufIovecBuilder to be used with vector recv

(Note: this ignores all push blocking failures!)

Reviewed By: simpkins

Differential Revision: D28394781

fbshipit-source-id: 973eea6e4cbe15d68d1f123d5fed98b50f5d23cc

Reviewed By: ot, luciang

Differential Revision: D28250766

fbshipit-source-id: 288c8c61d23c001dcb0d461d3d5179e82c755aaf

Summary:
In C++20, `std::atomic` is no longer trivial. Change the static-assert that checks for its triviality to check instead for narrower properties.

Fixes: https://github.com/facebook/folly/issues/1594.

Reviewed By: ot, Mizuchi, luciang

Differential Revision: D28856931

fbshipit-source-id: e7ff4a1e80e59fdc78efb271f63c6325d5528416

Summary:
* Rename from `proxy()` to `state()` and from `proxy_type` to `state_type`.
* Require `state_type` to be semiregular and testable (convertible explicitly to `bool`). A default-constructed instance and an instance just after assignment from a default-constructed instance must test false.
* Add `std::adopt_lock` constructor which can accept a `state_type`.

Reviewed By: aary

Differential Revision: D28724949

fbshipit-source-id: 64fc29bd875dc24cc3fd62d1b4b86f7ebdacfb73

Summary: By implementing it in terms of a param-pack expansion of calls to `with_exception`. No longer rethrows objects not derived from `std::exception`.

Reviewed By: luciang

Differential Revision: D28125669

fbshipit-source-id: b9eef160cc03e2d6f5524c1a5113405cc2881324

Summary: Rather than implementing it via `handle`, which is complex. Works now that exact casting can always be done without `catch`.

Reviewed By: luciang

Differential Revision: D27987594

fbshipit-source-id: 9961b55089f58df1b06793390817e497c6da54ae

Summary:
This fixes  https://github.com/facebook/fbthrift/issues/429

Pull Request resolved: https://github.com/facebook/fbthrift/pull/430

Reviewed By: iahs

Differential Revision: D28832985

Pulled By: vitaut

fbshipit-source-id: 0719f27207d11bb7970cc43c621640c425d8f55d

Summary: Rather than implementing it via `with_exception`, which is implemented in terms of `handle`, which is complex. Works now that exact casting can always be done without `catch`.

Reviewed By: ot, luciang

Differential Revision: D27987235

fbshipit-source-id: 822b0fe168e3b55c506b88810e13573b87c381c7

Summary:
- Add DeadlockDetector interface to folly/concurrency
- Add IOThreadPoolDeadlockDetectorObserver class that can be registered in IOThreadPoolExecutor to enable the functionality for its threads that use EventBase.
- In ThriftServer use the new observer with IOThreadPoolExecutor
- Default implementation of deadlock detector is stubbed out and does not do anything

Reviewed By: mshneer

Differential Revision: D28582490

fbshipit-source-id: 1cf3c2957b46de83f07c53bab708fe09ebbe2b79

Summary: Rather than implementing it via `with_exception`, which is implemented in terms of `handle`, which is complex. Works now that exact casting can always be done without `catch`.

Reviewed By: ot, luciang

Differential Revision: D27987172

fbshipit-source-id: eca54e47b3a0fc5da4614bbd23b881b1e35eca4b

Summary: Pull Request resolved: https://github.com/facebookexternal/f4d/pull/1090

Reviewed By: laithsakka

Differential Revision: D28774321

Pulled By: kkeita

fbshipit-source-id: 2c8eb8edf69611e94bf124b52fc8c1e11df9b527

Summary:
Description:

Prior to this patch it was not possible to run open/r on linux running on ARM due to cmake and openssl being harded to download/compile for the x86_64 arch.

In order to prevent such problem this patch actively compiles cmake instead of using pre-compiled binaries and also allows the OpenSSLBuilder to provide the correct build args to openssl, thus not trying to compile or run x86_64 software.

Pull Request resolved: https://github.com/facebook/openr/pull/95

Test Plan: * built the project by using ./build/build_openr.sh

Reviewed By: wez

Differential Revision: D28224684

Pulled By: cooperlees

fbshipit-source-id: 9de61dc6d7dcf7116ec5c67f3f165cd4a4bb5e5c

Summary: There was a missing newline separator

Reviewed By: yfeldblum

Differential Revision: D28781619

fbshipit-source-id: d20b26d67e427cf8760d21118e53a44ec6109771