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

Summary: As titled

Reviewed By: rsunkad

Differential Revision: D28774938

fbshipit-source-id: 123fd2b170eec941533f7a7d3bdd40dd9c4e967f

Summary:
This is towards upgrading to newer OpenNSA 6.5.22.

This diff will be landed after we have verified that OpenNSA 6.5.21 has not
regressed from OpenNSA 6.5.19 (HW test results)

sha256sum on the tarball, and use that in the manifest.

Reviewed By: rsunkad

Differential Revision: D28774943

fbshipit-source-id: 7fe376cc4c351605ae76138f1ddd1fdc170d0fe2

Summary: New OpenNSA requires this.

Reviewed By: rsunkad

Differential Revision: D28774940

fbshipit-source-id: c030bbab28546373338608c9324229a07c4bd30c

Summary:
Currently the log is printed on every `accept()` failure, and those are likely to be correlated (most commonly, while the service is crashing), causing log spew at a time when high log signal-to-noise is crucial.

Throttle it and a couple of other related logs.

Reviewed By: yfeldblum

Differential Revision: D28777847

fbshipit-source-id: b4d1898b41f722362b7cdda53597bb37f5fdfa83

Reviewed By: mshneer

Differential Revision: D28768951

fbshipit-source-id: 66f43a5505b293ee087ace85c753f05e23b55892

Summary:
Change the constructors for hazptr_holder and add the function make_hazard_pointer to be consistent with [WG21 P1121](http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2021/p1121r3.pdf).

Change the constructors for hazptr_array and add the function make_hazard_pointer_array to be consistent with the changes to hazptr_holder.

Now the default constructors construct empty holders and arrays.
The free functions make_hazard_pointer and make_hazard_pointer_array are used to construct nonempty holders and arrays.

Fix a missing barrier bug in HazptrTest.cpp.

Reviewed By: yfeldblum

Differential Revision: D28731269

fbshipit-source-id: 4132b2d612dbe1e8867c07d619efaba40be83785

Summary: No such function exists.

Reviewed By: luciang

Differential Revision: D28432945

fbshipit-source-id: 9f17b1820a984a8ed8e6817793be5ec995e1caef

Summary:
This change introduces the function `errorCategoryForErrnoDomain` with the contract that it will return the "correct" domain object to use for errno-domain errors when constructing a `std::system_error` object. The current implementation simply forwards to `std::generic_category()` on Windows and `std::system_category()` elsewhere.

## Background
As we understand how the standard intended for this to work:

1. std::generic_category should be used for standard errno domain
2. std::system_category should be used for the system's error code domain

This creates an ambiguity on platforms which define their system error codes as extended errno values, because it's not obvious which category should be used. This applies to macos, Linux, etc.

Historically we decided to use std::system_category for all errno-domain exceptions because we actually can't tell if a given errno value is "standard" or not, and on most platforms there isn't any clear separation of domains.

However, on Windows it turns out this is well-defined, and our previous code is actively incorrect. This is because the system's error code domain is well-defined as "things that `GetLastError()` can return" and that domain is *not* an extension of `errno`. The values intersect with different meanings and in some circumstances the constructor of std::system_error will fail and crash the program if the category is std::system_category and the value comes from the errno-domain.

We chose not to use `std::generic_category()` on non-Windows platforms because the reasoning that led us to use `std::system_category()` historically still makes sense - that domain still contains non-standard-errno values on those platforms.

Reviewed By: simpkins

Differential Revision: D28327673

fbshipit-source-id: d1a3a55b384e818eb4122d5cd03031bc7de90876

Summary: To have a smaller conditional.

Reviewed By: aary

Differential Revision: D28723687

fbshipit-source-id: 6c6d9fe33c0eb8d00148cf6ab1bd9c1ce0dfae79

Summary: It is required to unlock a held lock, if it owns one.

Reviewed By: aary

Differential Revision: D28704148

fbshipit-source-id: c329a3045fa27985fdec4142f00fe11c95394779

Summary: Fixes: https://github.com/facebook/folly/issues/1586.

Reviewed By: luciang

Differential Revision: D28694175

fbshipit-source-id: 56d57e99d5cee61004a499d8cf16da9a03acbcec

Summary:
Add IOBuf SizedFree API

(Note: this ignores all push blocking failures!)

Reviewed By: simpkins

Differential Revision: D28667639

fbshipit-source-id: 6264d829948110077bad9dead8dad785795c211b

Summary:
Pull Request resolved: https://github.com/facebook/folly/pull/1590

MSVC warning C4127 (conditional expression is constant) goes a bridge to far but we work around it.

* It warns against perfectly-correct C++14 code which makes it harder to write normal C++ which works with both C++14 and C++17.
* `if constexpr` asks the compiler to do different things from `if`. It is not a drop-in replacement when the conditional expression is constant.

Fixes: https://github.com/facebook/folly/pull/1588.

Reviewed By: akrieger

Differential Revision: D28692386

fbshipit-source-id: 2d47a7b283f6b7a32dae3deb497a9f2dd51fd36b

Summary:
Change the hazptr_hoklder function names get_protected and reset to (the names in [WG21 P1121](http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2021/p1121r3.pdf)) protect and reset_protection, respectively.

The primary change is in folly/synchronization/HazptrHolder.h

Reviewed By: yfeldblum

Differential Revision: D28662903

fbshipit-source-id: d074c718514716229daaf8dae14b5c275942b38a

Summary:
The guard targets its former use in `__thread` variables, which must be trivial. But it is now used in `thread_local` variables, which can have default constructors.

The relevant switch from `__thread` to `thread_local` was in {D27559375 (https://github.com/facebook/folly/commit/eeec4600b857dd70ca6b30837d14bb79cf9a2ef7)}.

Reviewed By: luciang

Differential Revision: D28683625

fbshipit-source-id: af74801a13e479a262ea50e6d3f89053e67d7ac9

Summary:
This diff adds an optional parameter to consume() and cancel(). This parameter will allow callback objects to know the source bridge that is calling them back. With this parameter, we will not need to create a separate callback object for each bridge (which saves memory).

The parameter is optional. If the parameter type is set to void, no parameter will be passed. This preserves backward compatibility with existing users of AtomicQueue that don't need a parameter.

Reviewed By: iahs

Differential Revision: D28550867

fbshipit-source-id: e7d998538c880c2a5c7649d3262cb7f8913e1439

Summary: This diff moves thrift's AtomicQueue to folly/experimental/channels/detail. This file will be shared between thrift and a new channels framework that will be added to folly/experimental/channels.

Reviewed By: iahs

Differential Revision: D28549810

fbshipit-source-id: de9d66c0f9fd73e89917df997526539b6b92f172

Summary: Guard internal compiler issue in makeCollectAllTryTask similar to collectAllTryRange

Reviewed By: iahs

Differential Revision: D28642361

fbshipit-source-id: a557e2c504b232a47f43c95d1163964652fdd435

Summary:
Reimplement `folly::exception_wrapper` thrown variant not to need to cache any pointers and rather to access the exception object and runtime type, and to perform runtime upcasting, via the platform runtime.

This both simplifies and extends the capability of the thrown variant.

Reviewed By: JunqiWang

Differential Revision: D28625444

fbshipit-source-id: 12409bc75794f5135cdb73563b07fafdf5cc44cd

Summary: Add TakeOwnershipOption::STORE_SIZE

Reviewed By: simpkins

Differential Revision: D28663081

fbshipit-source-id: 5f2b0cd5c9ebdb9a8475b35f8741bc0b17717441

Summary: Required to support RSA certificates in TLS 1.3.

Reviewed By: knekritz

Differential Revision: D28606758

fbshipit-source-id: 0b75eccf207c1327e590675137ce3289956d8771

Summary:
Add callbacks for sized free and corresponding allocations

(Note: this ignores all push blocking failures!)

Reviewed By: simpkins

Differential Revision: D28469569

fbshipit-source-id: 7d0d4c649c3494354ae8ec7cbda3e950eb3c80ed

Summary:
Add support for IOBuf::takeOwnership sizedFree

(Note: this ignores all push blocking failures!)

Reviewed By: simpkins

Differential Revision: D28510736

fbshipit-source-id: b102bd462df3d85856897d4acfff689d371ceb6c

Summary:
Similar to how CO_FAIL() was added, we could also benefit from a CO_SKIP()

Uses same definition as [GTEST_SKIP_](https://github.com/google/googletest/blob/master/googletest/include/gtest/internal/gtest-internal.h#L1354-L1355) but with co_return instead.

Reviewed By: ispeters, capickett

Differential Revision: D28645478

fbshipit-source-id: e92a23b91f8238247ad9d29b67c1a2954f43a38a

Summary:
Add Dockerfile build for openr.thrift python module.

The python module is built by:
1. Building and installing Facebook libraries with fbcode_builder
2. Building Open/R
3. Generating Cython files from thrift files with the FB thrift compiler
4. Generating C++ files from the Cython modules with the Cython compiler
5. Compiling the C++ modules into shared objects

Future work for building and distributing Breeze:

- Fix the hacks in build_breeze.sh, see comments therein
- Use a staged Dockerfile build for the Open/R and Breeze build
- Install openr.thrift. The openr.thrift shared objects are build and
  stored in the Docker image generated by Dockerfile, but are unused.
- Install all the openr python submodules in a single openr site-package
- Add cross-compilation to the openr.thrift build. This is needed for
  Terragraph
- Upload the openr python package to PyPi

Reviewed By: saifhhasan

Differential Revision: D28614443

fbshipit-source-id: 38b7e7c5594fd4bb5a338f19c69e5fc3b3b95863

Summary: TSAN reports race condition in `LOG_FIRST_N`. Replacing it with thread-safe alternatives.

Reviewed By: yfeldblum

Differential Revision: D28541027

fbshipit-source-id: 7054a82c7b26df395a25cd7cd7f921a9c636ffbc

Summary:
Another approach to fixing the exception-ptr accessors for libc++.

This time, just define mimics of `__cxa_exception` both ways, and test to find out which way to cast.

Reviewed By: JunqiWang

Differential Revision: D28561635

fbshipit-source-id: d27430d8fe2d73205bf1a8c31321be4c76808df8

Summary: Improve documentation and make the description of synchronous reclamation  consistent with the current implementation.

Reviewed By: yfeldblum

Differential Revision: D28575528

fbshipit-source-id: e8f37e02d18b12a0e653264c128492a98cba6a1d

Summary:
* `VirtualGen` gains a move-only counterpart, `VirtualGenMoveOnly`
* `virtualize` added as an operator to type-erase generator names, particularly for making errors and stack traces more legible.
* `FunctionRef` used to avoid heap allocations where short-lived references are sufficient.

Differential Revision: D28167691

fbshipit-source-id: 5beaab3f613666ffafc7ac1951be1933c2863d3f

Summary: `ConcurrentHashMap` has a `DCHECK` for use-after-destruction. One of these has a comment explaining what it's for, but `DCHECK` supports taking a message to show the user. Make it a message.

Reviewed By: yfeldblum, magedm

Differential Revision: D28609176

fbshipit-source-id: ea6e57b7650ad597f45ea3efc2fa0fb62d5de40e

Summary: Add sendmsg/recmsg io_uring support

Reviewed By: yfeldblum

Differential Revision: D27148068

fbshipit-source-id: efb9174fb40f5f5bf0d0318db12cb61075866381

Summary:
Use the new ConstructorCallback class to allow exterior functions/classes
to get a callback for each new AsyncSocket()

Implications:
This adds an optional and an atomic int to the size of each AsyncSocket so
there is a very small increase in mem usage even if this is unused (which will
be most code, at least initially).

There is also a very small amount of additional CPU on constructing an
AsyncSocket: checking the std::option to see if it's non-nullopt.

Reviewed By: bschlinker

Differential Revision: D27516379

fbshipit-source-id: 56f4411afcd404dcf0092cce562230e5c0dcdf17

Summary: Original commit changeset: ce083cec4c31

Differential Revision: D28558687

fbshipit-source-id: 5cf4ac838c80428902280f3f9ccdd1174410fbe6

Summary:
It is safe to expose `size()` without requirement to go through
accessor (which adds reference tracking overhead).

Reviewed By: yfeldblum

Differential Revision: D28591686

fbshipit-source-id: ea0be2a34df01488402e53fdfbe232ee710f03f7

Reviewed By: igorsugak

Differential Revision: D28582083

fbshipit-source-id: 9beedfccfd1fc6b9dc4f7ff8ab0139282bd285ab

Summary:
## Description

Hi!
I wanted to fix a build of folly on Android.

Trying to use folly's `MPMCQueue` on Android triggers the following error:
```
{xxx}/rnfbjni/android/app/build/third-party-ndk/folly/folly/Subprocess.cpp:542:19: error: use of undeclared identifier 'getdtablesize'; did you mean 'getpagesize'?
```
where `{xxx}` is a local path on my machine.

[Here](https://github.com/karol-bisztyga/rnfbjni/tree/patch-folly-repro) is a branch of my project that reproduces the problem(I've already fixed it by patching `folly` so you can see what patch has to be removed in order to make it fail [here](https://github.com/karol-bisztyga/rnfbjni/commit/e6d8059e55efda5f88d5edbc1e932b94223b052c)).

From [the man page for `getdtablesize`](https://linux.die.net/man/2/getdtablesize):

> SVr4, 4.4BSD (the getdtablesize() function first appeared in 4.2BSD). It is not specified in POSIX.1-2001; portable applications should employ sysconf(_SC_OPEN_MAX) instead of this call.

### Related issues:
https://github.com/facebook/folly/issues/1279
https://github.com/facebook/folly/issues/918 - probably, I cannot see bug details.

Pull Request resolved: https://github.com/facebook/folly/pull/1584

Reviewed By: simpkins

Differential Revision: D28547979

Pulled By: yfeldblum

fbshipit-source-id: 3b8d882ad2a4b7cb817a2c7bbc5f9f750b35770e