Summary:
This is a lock-free replacement for NotificationQueue. Unlike notification queue it supports only one consumer, which allows simplifying the internal structure a lot and avoiding any use of spin locks.
Another improvement comparing to NotificationQueue is the use of epoll edge-triggered events. This allows completely eliminating read syscals, since we don't need to drain eventfd to stop it from waking up the epoll loop.

Differential Revision: D24297158

fbshipit-source-id: 1f418d861210f6dcad52b4d663f49eb2e8f3ae8c

Summary:
`libfolly.pc` should contain the actual version of Folly, instead of just
`master`.

Reviewed By: danobi

Differential Revision: D24471074

fbshipit-source-id: 5fc29eeee25c8c317c3405f430fd079d47c441bd

Summary:
Existing test was supposed to be testing the exception, but was instead dying because of the pid check.
Have 2 separate tests and use crash message regex to make sure they actually test what they are supposed to be testing.

Differential Revision: D24461683

fbshipit-source-id: b7ef30e488d266ea17e283f9b13b5d013a4a125a

Summary:
This should make packaging folly easier, as it allows one to set the
.so version to denote ABI breaks. Specifically, we're looking to use this in
https://bugzilla.redhat.com/show_bug.cgi?id=1887621 to package folly in Fedora
and have the version match the tagged release version (i.e. treat every release
as ABI breaking).

Differential Revision: D24451669

fbshipit-source-id: f63be760c90567117f1ebbec37416c0d9d8a4736

Summary:
This reverts https://github.com/facebook/folly/pull/833 by adding back
the unconditional setting of `POSITION_INDEPENDENT_CODE` when building shared
libs. Without it, the build fails with a number of relocation errors. This was
previously reported in https://github.com/facebook/folly/issues/983

Differential Revision: D24449731

fbshipit-source-id: 6f6cb71ef545d02998afb09909772b62657c1262

Summary: These do not need to be executable

Differential Revision: D24449732

fbshipit-source-id: 5f013c62dd6bf88fe4d50782a4c5bfcec6834aec

Summary:
`frame` may be unused depending on whether it is compiled with or without asserts.

`FOLLY_MAYBE_UNUSED` is supposed to take care of this, but it does not work on Windows and trips up the comiler:

```
warning C4100: frame: unreferenced formal parameter
```

Reviewed By: johnkearney, akrieger

Differential Revision: D24451220

fbshipit-source-id: 8225640a5e8fd0ceed4ffcc53cea7422a5581d6f

Summary: [Folly] `to_shared_ptr_aliasing`, exposing the shared-ptr aliasing constructor as a free function deducing the types of its arguments.

Reviewed By: Mizuchi

Differential Revision: D24414504

fbshipit-source-id: 290e822464860b824961986211375d87f942f483

Summary:
Updates the FiberManager to make sure that it saves/restores the current-thread's
top-most active AsyncStackRoot when the coroutine is suspended/resumed.

This is required for correctness to ensure async stack traces remain functional
in the presence of fiber context-switches.

(Note: this ignores all push blocking failures!)

Reviewed By: yfeldblum

Differential Revision: D21130668

fbshipit-source-id: 4002bae0fc12a61c38836f9f099da4e6cf9183f4

Summary:
While running some unit tests which happen to call into `folly::futures::retryingPolicyCappedJitteredExponentialBackoff`, I noticed that  `retryingJitteredExponentialBackoffDur` was triggering a "float-cast-overflow" error from UndefinedBehaviorSanitizer.

It turns out that this only happened if `backoff_min` was 0 and there were at least 1025 retries. The large retry count caused the exponential term in the calculation to become `Infinity`. In floating point, Infinity multiplied by 0 was `NaN`, causing the sanitizer error.

Although `backoff_min` should typically not be set to 0 because this effectively disables all backoff, I think there may be cases, such as unit tests, where setting `backoff_min` to 0 does make sense. This commit adds an additional check to handle this unusual case.

Reviewed By: yfeldblum

Differential Revision: D24122470

fbshipit-source-id: 48a8aede0d22e73cf001dab5eb184e841f3c23aa

Summary: Remove FOLLY_HAVE_MSG_ERRQUEUE check

Reviewed By: yfeldblum

Differential Revision: D24295244

fbshipit-source-id: 24d4a69a013951b37fe7c93539b26eedf8113360

Summary:
We were seeing the following build failures on i386:
```
In file included from folly/stats/detail/BufferedStat.h:151,
                 from folly/stats/QuantileEstimator.h:20,
                 from folly/stats/QuantileEstimator.cpp:17:
folly/stats/detail/BufferedStat-inl.h: In member function 'void folly::detail::BufferedStat<DigestT, ClockT>::doUpdate(folly::detail::BufferedStat<DigestT, ClockT>::TimePoint, const std::unique_lock<folly::SharedMutexImpl<false> >&, folly::detail::BufferedStat<DigestT, ClockT>::UpdateMode)':
folly/stats/detail/BufferedStat-inl.h:72:3: error: there are no arguments to 'DCHECK' that depend on a template parameter, so a declaration of 'DCHECK' must be available [-fpermissive]
   72 |   DCHECK(g.owns_lock());
      |   ^~~~~~
```

My best guess is some kind of argument dependent lookup complication.

This diff sidesteps the issue by using plain old `assert()` instead

Reviewed By: yfeldblum

Differential Revision: D24375147

fbshipit-source-id: d702d883fcabc548e8f4969207e265abf21acf2b

Summary: AsyncUDPSocket: use actual controllen size instead of max

Reviewed By: danobi

Differential Revision: D24368739

fbshipit-source-id: 0c1083a058d8fe0ddc212989a262f3eeebfa610c

Summary:
Pull Request resolved: https://github.com/facebookexperimental/rust-shed/pull/15

With the update from rustc 1.44 to 1.47 the doctests started to break when building a crate that depends on a different crate of the same name, see this: https://github.com/rust-lang/cargo/issues/6819

This diff also removes the TravisCI build as it was broken and unused for some time.

Reviewed By: krallin

Differential Revision: D24390815

fbshipit-source-id: fa9c78ef394640afbf8c419f8276bbf339960406

Reviewed By: zertosh

Differential Revision: D24388883

fbshipit-source-id: 6bba618c6a919cd3947d2a65954feff989bb80a2

Summary:
A number of libraries at Facebook include interval midpoint calculations; however, doing these in a mathematically precise way (without over/underflow) can be tricky. Doing them wrong can break binary searches over large datasets and give imprecise floating-point calculations.

This function provides an early opportunity to fix binary searches and other calculations which can later be updated to `std::midpoint()` when C++20 becomes available.

Reviewed By: yfeldblum

Differential Revision: D23997097

fbshipit-source-id: 373e0dc1d1ff071f697ee782be46fb0d49a2f8f7

Summary:
Windows paths can both use forward slashes as well as backslashes, and also
contain colons that folly uses to separate prefixes to strip.

The xlogStripFilename unfortunately didn't handle both of these cases leading
to log categories to always contain the full path instead of a suffix path.

Reviewed By: yfeldblum

Differential Revision: D24353208

fbshipit-source-id: ff17afe4b46f36d46bf468a5e3cf9ba9ba8412b8

Summary:
Similar to D24179337 (https://github.com/facebook/folly/commit/7cc28784d1749283b7eb78399a0f0a77594dbc14).

Though I'm not actually sure what's causing the increase... maybe the right thing to do is the debug the underlying cause, but I'm also not sure how to debug this exactly.

Reviewed By: yfeldblum

Differential Revision: D24353676

fbshipit-source-id: 12bc5eae939f95d8d091280dc5b5af638f0152a0

Summary:
The MSVC compiler complains about truncation of size_t to int:

  Z:\installed\folly\include\folly/synchronization/HazptrDomain.h(210): warning C4267: 'argument': conversion from 'size_t' to 'int', possible loss of data

MSVC being very verbose, this is followed by 25 lines of template expansion.

Since `shard` is a size_t due to `calc_shard` returning that type, let's simply
make the shard argument of hazptr_warning_list_too_large a size_t too to
silence the warning.

Reviewed By: magedm

Differential Revision: D24365556

fbshipit-source-id: b426cee7caaf593eb112691bc45b83e564342b0e

Summary:
Pull Request resolved: https://github.com/facebookexperimental/eden/pull/67

With this change it will be possible to build dependencies of and run integration tests using getdeps.py.

This is the first goal of Q4 as per https://fb.quip.com/v8YzAYNSYgot: "Get Open Source version of integration tests running on Legocastle".

Before this diff:
The OSS integration tests run now on GitHub by:
- Building some test dependencies with getdeps.py
- Building some test dependencies with homebrew/apt-get
- Running tests via python script

The OSS integration tests were not running on Sandcastle.

After this diff:
The OSS integration tests run on Github by:
- Building and executing tests via getdeps.py (execution of tests happens by getdeps.py calling Make calling python script)

The OSS integration tests run on Sandcastle using the same getdeps.py setup as Github.

Reviewed By: krallin

Differential Revision: D24253268

fbshipit-source-id: cae249b72d076222673b8bbe4ec21866dcdbb253

Summary:
# What is the goal?

Address the problem in the post at https://fb.workplace.com/groups/code.indexing/permalink/1929080303917865/ with the comment(s) in Codex for EventBaseManager.

# What was not the way?

Jie Jua noticed that the current comment on "EventBaseManager() {}" in Codex was

```
encouraged. You should instead use the global singleton if possible.
```

# What is the new way?

Use documentation comments (doxygen) style `/** double starts */` instead of `//` ordinary comments. This should show the full comment which is:

```
XXX Constructing a EventBaseManager directly is DEPRECATED and not
encouraged. You should instead use the global singleton if possible.
```

Which really needs that first line.

Differential Revision: D24333522

fbshipit-source-id: 8e84e0a49426844d5a6f6438b0d76fb9886742fe

Summary: The bare name isn't defined on windows at all, so we have to use the new internal name instead.

Reviewed By: yfeldblum

Differential Revision: D24044971

fbshipit-source-id: 24b37903ed0f309910d9f5141240b951bf5fba15

Reviewed By: yfeldblum

Differential Revision: D23545773

fbshipit-source-id: e93f84d0658166e875710a4c0fb5dc165a8628d8

Summary: Migrate AsyncSSLSocketTest to new AsyncSSLSocket session API

Reviewed By: mingtaoy

Differential Revision: D24240249

fbshipit-source-id: ea9bd7669f377c738e44d3348400a2cfbd3c2580

Summary: Use the forkInstrumented call when running under TSAN.

Reviewed By: yfeldblum

Differential Revision: D23787959

fbshipit-source-id: 159f8354ef9746b8374d6da12b22e389c2bc71f4

Summary: [Folly] Fix `is_constexpr_default_constructible` under clang-10, which suffers a crash (https://bugs.llvm.org/show_bug.cgi?id=47620) and divergent behavior in certain constexpr evaluations.

Reviewed By: mpark, Mizuchi

Differential Revision: D24302661

fbshipit-source-id: b6a58b5b675f367bf2dc1d810edb9926cfb87a2e

Summary: TSAN fix for folly/io/test:shutdown_socket_set_test

Differential Revision: D24321155

fbshipit-source-id: 753a808c868da4c608c3ee95d16a3673d6f700b0

Summary: Reduce io_uring test mem requirements

Reviewed By: xiaofeidu008, danobi

Differential Revision: D24314788

fbshipit-source-id: 35ec01475fbc1cf851a81e33cf8c7fec8dd88ddb

Summary: Cache unwrapped observer instead of unwrapping it on every getObserver call.

Reviewed By: jmswen

Differential Revision: D24325934

fbshipit-source-id: d8a023e59c933222a5f43aaacbd9c955236739b3

Summary:
Remove comment about potential high contention under sustained frequent removal from CHM-s.

Recent changes to the hazard pointer library eliminated this problem:
- Asynchronous reclamation of tagged objects (including CHM objects) no longer needs to be done while holding the lock on the list of tagged objects.
- The list of tagged objects is now sharded and sharding can be easily increased if needed.

Reviewed By: yfeldblum

Differential Revision: D24318407

fbshipit-source-id: 3124e8d5651848a16084899a6f31d69cc235c702

Summary:
allow-large-files

black_any_style

Reviewed By: zertosh

Differential Revision: D24325133

fbshipit-source-id: b4afe80d1e8b2bc993f4b8e3822c02964df47462

Summary: Apparently it's flaky...

Differential Revision: D24179337

fbshipit-source-id: 75416adaf8dbd37fc9ddeaa0500ddfaeb80cffa9

Summary: [Folly] Cut linter suppression in `TryTest.cpp`.

Reviewed By: zertosh

Differential Revision: D24294536

fbshipit-source-id: 64b26ec4632ffdae8e246aa27a6c58c579e2db10

Summary: Move folly binary

Reviewed By: yfeldblum

Differential Revision: D24294709

fbshipit-source-id: e7cf0be6e45d3a132fc81143103f3b5cdedbb76f

Summary:
This is best effort. We cannot control how many tests are running
at the same time. And if the user is running other tests that consumes resource

Differential Revision: D24289056

fbshipit-source-id: 0810ba0989e84a5bce6d695dd6df0dfe2618fbf7

Summary:
Gmock's `Return(x)` is a valid action as long as `x` is implicitly convertible to the return type of the function.

The current implementation of `CoReturn(x)` does not support that, since `Task<T>` is not convertible to `Task<Q>` even if `Q` is convertible to `T`.

This diff changes the `CoReturn` implementatio to match the way Gmock does, allowing for the implicit conversion.

This is extremely valuable when using `Task<Expected<...>>`, as it allows to write `CoReturn(makeUnexpected(...))` instead of `CoReturn(Expected<...>(unexpected, ...))`.

It also make the translation from a mock being sync to async more straightforward: add `Task` in the mock return type and replace `Return` with `CoReturn`.

Reviewed By: lewissbaker

Differential Revision: D24046013

fbshipit-source-id: 9f166237a624600be51c00616f145ae4e1a45443

Summary:
This diff adds all third party dependencies that are required by getdeps to be able to build and runn Mononoke's integration tests.

Also add a stub Makefile with no-op steps that will be filled in next diff.

Reviewed By: ahornby

Differential Revision: D24251894

fbshipit-source-id: 67384ecfd0ced6762dddc3c6e61feb1240b1162d

Summary:
This specifically fixes a TSAN detected race where SemiFuture destructor was checking Core state to see if deferred executor has to be detached, while that state could become Empty if the callback is being proxied and both threads code be racing while accessing the executor.
hasCallback() checks if the callback was ever set (not if the callback is stil alive) and it's true for State::Empty (similar to State::Done).

Differential Revision: D24271346

fbshipit-source-id: 5f52f20783d36745d1200497e0300bc086b65213

Summary: Add support for UDP RX timestamps

Reviewed By: mjoras

Differential Revision: D24146914

fbshipit-source-id: 381839fc5402f13e428364c47c5752473acda6af

Summary: Disable CLANGTIDY checks for several places at the code where HOWTOEVEN checks are disabled.

Reviewed By: suo

Differential Revision: D24198276

fbshipit-source-id: c52cd10941252fcacfea30589660c57eb7099011