Michael Lee [Tue, 14 Jun 2016 00:05:24 +0000 (17:05 -0700)]
Time.h portability for iOS 10.
Summary: The iOS sdk adds `clock_gettime` support.
Reviewed By: clementgenzmer
Differential Revision:
D3427621
fbshipit-source-id:
fcd3022c5ea05ceb8e289fbeb1ae845155d2dea9
Alexey Spiridonov [Sat, 11 Jun 2016 02:12:36 +0000 (19:12 -0700)]
DCHECK on reentrant invocations of loop()
Summary:
I had a crash bug where an `EventBase` handler called a function, which tried to use the thread's EventBase as if it were its own:
```
auto evb = folly::EventBaseManager::get()->getEventBase();
// schedule work on evb, which calls evb->terminateLoopSoon() on completion
evb->loopForever();
```
This ended up invoking the `event_base_loop()` reentrantly, and corrupting the heap in a hard-to-debug way.
Reviewed By: djwatson
Differential Revision:
D3408155
fbshipit-source-id:
b8855aa3b390fa33e032ab295d386122d2cb872a
Ben Hamilton [Fri, 10 Jun 2016 23:01:34 +0000 (16:01 -0700)]
Make Fibers library conditional on working Boost.Context fcontext_t
Summary:
Boost 1.61 removes the `boost::context::fcontext_t` type
which Folly Fibers depends on.
This makes the Fibers library conditional on a working `fcontext_t`,
so Folly will build (just without Fibers) if it's not present.
Depends On
D3419638
Fixes https://github.com/facebook/folly/issues/413
Reviewed By: Orvid
Differential Revision:
D3420097
fbshipit-source-id:
876f8f76a2d0dda4ce5085c27875ab6aaa7b86c0
Andrii Grynenko [Fri, 10 Jun 2016 22:25:44 +0000 (15:25 -0700)]
Fix folly::fibers on OSX with boost 1.60
Summary: We only need fcontext.hpp, so let's include it directly. all.hpp includes other headers, which use thread-locals (doesn't work with our OSX build).
Reviewed By: igorsugak, bhamiltoncx
Differential Revision:
D3419638
fbshipit-source-id:
b4fa1bd14454bcd75eaece52c398ac6510273773
Maged Michael [Fri, 10 Jun 2016 15:53:32 +0000 (08:53 -0700)]
Added support for std::mutex in folly Deterministic Schedule
Summary:
Added the class `DeterministicMutex`, a wrapper for `std::mutex`, in `folly/test/DeterministicSchedule.h`. Added a test to `folly/test/DeterministicScheduleTest.cpp` to test the correct behavior of `DeterministicMutex` functions and that deterministic schedule is able to detect race conditions that involve `DeterministicMutex`.
Note: Bootcamp task
Reviewed By: djwatson
Differential Revision:
D3412579
fbshipit-source-id:
c12861c4ec1cfeadcef027be4513e8e4cb7e0176
Marcus Holland-Moritz [Fri, 10 Jun 2016 06:40:54 +0000 (23:40 -0700)]
Get rid of circular dependency when including ExceptionWrapper.h from Conv.h
Summary:
This is needed in order to be able to use folly::Try<> with folly::to<> for
non-throwing conversions.
Reviewed By: meyering
Differential Revision:
D3411003
fbshipit-source-id:
1f20e7871b9cedda2b35527dac70381579abd708
Christopher Dykes [Thu, 9 Jun 2016 23:01:25 +0000 (16:01 -0700)]
Use intrinsics rather than inline assembly where possible
Summary:
I would switch these to just use the intrinsic functions, but GCC 4.8 doesn't support them.
MSVC supports the intrinsics, which is the primary reason for the switch.
Reviewed By: philippv
Differential Revision:
D3278901
fbshipit-source-id:
60103ac7cf7ddfb529f65f4aadc687dbdaf423a1
Chad Parry [Thu, 9 Jun 2016 22:07:58 +0000 (15:07 -0700)]
Allow a timer with pending callbacks to be destroyed
Summary:
This diff is part of a string of changes that is getting `HHWheelTimer` to use standard smart pointers.
D2617966 is being reverted. The `DestructorGuard` that was added for safety is not possible without intrusive reference counting.
That means that the `count_` (of pending callbacks) will still be correct, but there will be nothing to check it against. Because of that, I removed the assertions to make sure that the count of pending callbacks matches the count of active guards. Anyway, there were already ways to defeat that test, such as by creating a `DestructorGuard` from any client code.
As long as the assertions are gone, the `cancelAll` can be moved to the destructor. It only ever got run when `destroy` was called and `count_` was `0`, which means that `~HHWheelTimer` was also going to get called at the same time.
Reviewed By: djwatson
Differential Revision:
D3349303
fbshipit-source-id:
433fa6605ee9921833328e2f82bd07b2e17e42c5
pravic [Thu, 9 Jun 2016 22:01:51 +0000 (15:01 -0700)]
Fix typo in docs/Function.md
Summary: Closes https://github.com/facebook/folly/pull/421
Reviewed By: yfeldblum
Differential Revision:
D3413832
Pulled By: elliottneilclark
fbshipit-source-id:
7af1462c250c8d6b5d6f5dfcb3f3117dff7877c9
Martin Martin [Thu, 9 Jun 2016 18:53:14 +0000 (11:53 -0700)]
Fix a data race found by TSAN in folly::fibers::Baton
Summary: Fix a data race found by TSAN in folly::fibers::Baton by changing a memory_order_relaxed to memory_order_acquire. Nathan Bronson says that TSAN is correct and the C++ memory model requires _acquire here.
Reviewed By: andriigrynenko
Differential Revision:
D3412519
fbshipit-source-id:
bd0043b41d145e689a97fc7ef47aa6e116ea9194
David Callahan [Thu, 9 Jun 2016 01:31:19 +0000 (18:31 -0700)]
fix command line parsing
Summary: parse command line flags to pick up -json flag
Reviewed By: mzlee
Differential Revision:
D3405794
fbshipit-source-id:
e2d216c74f04c15d7470c70abfc2bd25c2ceda98
Brett Simmers [Thu, 9 Jun 2016 00:19:20 +0000 (17:19 -0700)]
Allow const mapped types in folly::AtomicHash(Array|Map)
Summary:
Without this const_cast, it's impossible to insert anything into an
AHA with a const mapped_type, making it effectively useless.
Reviewed By: yfeldblum
Differential Revision:
D3405687
fbshipit-source-id:
3ecba19e0e92661c2c537c747b4927176104939f
Adam Simpkins [Wed, 8 Jun 2016 23:15:18 +0000 (16:15 -0700)]
make Range::size() constexpr
Summary:
Declare size() to return a constexpr value, so it can be called on constexpr
Range objects.
This unfortunately does drop the existing assert() check, due to a gcc bug:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71448
Reviewed By: lbrandy
Differential Revision:
D3394612
fbshipit-source-id:
77ea3b961dc323a39dea6f0e5850f9a311210d09
Christopher Dykes [Wed, 8 Jun 2016 19:51:28 +0000 (12:51 -0700)]
Pack PackedSyncPtr
Summary: The comment was right about the packed attribute not liking private members, so pack `PackedSyncPtr` to get GCC to pack structs containing it correctly.
Reviewed By: yfeldblum
Differential Revision:
D3402936
fbshipit-source-id:
11d9def132d5dedf04b0e263c379cac339a52fa3
Adam Simpkins [Wed, 8 Jun 2016 19:49:45 +0000 (12:49 -0700)]
add hexlify() and unhexlify() convenience methods
Summary:
Add new versions of hexlify() and unhexlify() that directly return a
std::string, similar to the helper functions that already exist for humanify()
and backslashify().
The function signatures for these helpers are different than the humanify() and
backslashify() helpers--these functions always accept StringPiece or ByteRange
arguments. This allows them to easily accept "const char*" arguments, without
having to convert the arguments to a std::string. (I think we probably should
fix the humanify() and backslashify() helpers to be similar in a separate
diff.)
Reviewed By: yfeldblum
Differential Revision:
D3403863
fbshipit-source-id:
a2df49d5857e1d34caac3d78283dc293f4ef1ab6
Andrii Grynenko [Wed, 8 Jun 2016 01:30:18 +0000 (18:30 -0700)]
Disable SIGSEGV signal handler if run in JVM
Reviewed By: ldemailly
Differential Revision:
D3402550
fbshipit-source-id:
3de47c569868545c3861dab9a7b244ed8fb1fadf
Adam Radziwonczyk-Syta [Tue, 7 Jun 2016 19:59:48 +0000 (12:59 -0700)]
Expose current observer from FiberManager
Summary:
Inteded use case is to write tracing code in common components that can adjust its behavior to the current observer. E.g.
auto* observer = FiberManager::getUnsafe()->getObserver();
if (auto* specific_observer = dynamic_cast<SpecificObserver*>(observer)) {
specific_observer->set_fiber_name(...);
}
Reviewed By: andriigrynenko
Differential Revision:
D3392890
fbshipit-source-id:
284b4964bbb17e9a081779b870af68e8f1a81634
Subodh Iyengar [Tue, 7 Jun 2016 14:43:37 +0000 (07:43 -0700)]
Add TFO support to AsyncSSLSocket
Summary:
This adds TFO support to AsyncSSLSocket which
uses the support for TFO from AsyncSocket.
Because of the way AsyncSSLSocket inherits from
AsyncSocket it is tricky.
The following changes were made:
1. Openssl internally will treat only errors with return
code -1 as READ_REQUIRED or WRITE_REQUIRED errors. So this
diff changes the return value of the errors in the TFO fallback
cases to -1.
2. In case we fallback after SSL_connect() to a normal connect,
we would have to restart the connection process after connect
succeeds. To do this this overrides the connection success callback
and restarts the connection before sending the callback to AsyncSocket
because sometimes callbacks might synchronously call sslConn() in the
normal connect cases.
3. Delegated bioWrite to call sendSocketMessage instead of sendmsg directly.
Reviewed By: djwatson
Differential Revision:
D3391735
fbshipit-source-id:
61434f6de4a9c3d03973c9ab9e51eb49e751e5cf
Kenny Yu [Mon, 6 Jun 2016 23:20:39 +0000 (16:20 -0700)]
folly: define FOLLY_SANITIZE_THREAD convenience macro for TSAN
Summary:
This defines the macro `FOLLY_SANITIZE_THREAD`, similar to `FOLLY_SANITIZE_ADDRESS`.
Note that gcc *just* landed support for the `__SANITIZE_THREAD__` macro two weeks
ago: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64354
Reviewed By: igorsugak
Differential Revision:
D3389021
fbshipit-source-id:
df6497a7741657c297ee812ed07cbae102b6810d
Christopher Dykes [Sat, 4 Jun 2016 00:43:42 +0000 (17:43 -0700)]
Allow constexpr construction of AtomicStruct under MSVC
Summary: Attempting to use this under MSVC, as `MemoryIdler::defaultIdleTimeout` does, generated warnings about the symbol being dynamically initialized due to an implementation limation. Solve this by defining a union instead of calling `encode` in the constructor.
Reviewed By: yfeldblum
Differential Revision:
D3387916
fbshipit-source-id:
b0d628b6d086960e94121b6183a31348fb151aac
Christopher Dykes [Sat, 4 Jun 2016 00:43:13 +0000 (17:43 -0700)]
Deal with a couple of warnings from MSVC
Summary: MSVC is warning that `client` may be uninitialized, which is possible if the callback doesn't actually set them for some reason. Solve that by explicitly initializing it to `nullptr`. Also deal with the warning about the initialization of `dynamic::array` using braced syntax by not using that syntax.
Reviewed By: yfeldblum
Differential Revision:
D3387782
fbshipit-source-id:
e4d25ad32e99a73d7d062be9a2f2b5bde5d17396
Christopher Dykes [Sat, 4 Jun 2016 00:42:35 +0000 (17:42 -0700)]
Remove a couple of MSVC portability defines
Summary: The first is both wrong (due to the difference in semantics) and unneeded, as MSVC 2015 supports `snprintf` natively and with the correct semantics. The second is simply never used.
Reviewed By: yfeldblum
Differential Revision:
D3387551
fbshipit-source-id:
ea666a60ce859a59ad3d3ff961b4d7d642fd86eb
Andrii Grynenko [Fri, 3 Jun 2016 22:01:28 +0000 (15:01 -0700)]
Add a SIGSEGV signal handler which detects stack overflow
Summary: GuardPageAllocator now keeps a track of all protected pages. If SIGSEGV occurs, signal handler can check address at fault against protected pages and if it matches - we got a fiber stack overflow.
Reviewed By: yfeldblum
Differential Revision:
D3386960
fbshipit-source-id:
775b36ee08fecbbd87da0025f794717c222f5cce
Subodh Iyengar [Fri, 3 Jun 2016 19:40:37 +0000 (12:40 -0700)]
Always override write bio method
Summary:
Always overriding write bio method
allows us to more cleanly implement
features like eor tracking, support
multiple ssl libraries, and also TFO
Reviewed By: anirudhvr
Differential Revision:
D3350482
fbshipit-source-id:
ddd2333431f9d636d69c8325b2c18d7cc043b848
Elliott Clark [Thu, 2 Jun 2016 22:37:35 +0000 (15:37 -0700)]
Update folly OSS build to use 1y
Summary: use std=c++1y for compiling since we use c+=14 feautures but don't want to require a recent gcc.
Reviewed By: bhamiltoncx
Differential Revision:
D3380287
fbshipit-source-id:
8ece6007cf613d307bafe692f1af70f5359df95a
Christopher Dykes [Wed, 1 Jun 2016 22:36:26 +0000 (15:36 -0700)]
Assert that MicroSpinLock is a POD type
Summary: Because if it's not, all hell will break lose (apparently).
Reviewed By: meyering
Differential Revision:
D3373061
fbshipit-source-id:
f08a863d1bacadeeff9d9cea041f399f136cfc74
Christopher Dykes [Wed, 1 Jun 2016 22:34:02 +0000 (15:34 -0700)]
PackedSyncPtr.data_ is private
Summary: As the (now removed) comment says, GCC 4.6 doesn't treat structs with private members as POD types... The thing is, GCC 4.6 is long dead, so make `data_` private and assert that it's a POD type.
Reviewed By: meyering
Differential Revision:
D3373172
fbshipit-source-id:
b2f1ee59d544461fe986705164cc6c866af62865
Ben Hamilton [Wed, 1 Jun 2016 17:43:50 +0000 (10:43 -0700)]
Fix test/SynchronizedTest.cpp build break on Clang on OS X
Summary:
The Folly `make check` build is broken with Clang on OS X.
This diff fixes the following break in `test/SynchronizedTest.cpp`:
SynchronizedTest.cpp:150:10: error: thread-local storage is not supported for the current target
static thread_local int lockCount_;
The problem is clang on OS X doesn't support the `thread_local` extension:
http://stackoverflow.com/a/
29929949
so we now use the `FOLLY_TLS` compatibility macro.
When I fixed this, I found the test also failed to compile if `FOLLY_SYNCHRONIZED_HAVE_TIMED_MUTEXES` was `#define`d to `0` (as it is on OS X), so I fixed that.
Reviewed By: mzlee
Differential Revision:
D3361215
fbshipit-source-id:
e93be6872bcab03090448b9421e502e472f149ff
Christopher Dykes [Wed, 1 Jun 2016 17:14:26 +0000 (10:14 -0700)]
SingletonVault_c.cpp is always C++
Summary: It's simply not possible to compile the C++ side of the C API as C, so don't guard the `extern "C"` as if it is.
Reviewed By: meyering, yfeldblum
Differential Revision:
D3369474
fbshipit-source-id:
d3316704ce17a471bc9d6a3a235472bf40cedb9d
Subodh Iyengar [Wed, 1 Jun 2016 08:45:00 +0000 (01:45 -0700)]
Methods to change cipher suite list
Summary: Add methods to change cipher suite list and sigalg list
Reviewed By: anirudhvr
Differential Revision:
D3295935
fbshipit-source-id:
eb46311986465e399eafa69e3070b53b36bce820
Michael Lee [Tue, 31 May 2016 18:24:56 +0000 (11:24 -0700)]
SocketFastOpen should include <cerrno>
Summary: Include cerrno to define `EOPNOTSUPP` and `ENOPROTOOPT`.
Reviewed By: elindsey
Differential Revision:
D3366479
fbshipit-source-id:
3a93a1c247ffa4a9d4ea03577713362a2a35c8b7
Michael Lee [Tue, 31 May 2016 16:31:41 +0000 (09:31 -0700)]
SubprocessTest does not use dirent.h
Summary: Remove it as we do not need it here.
Reviewed By: yfeldblum
Differential Revision:
D3365991
fbshipit-source-id:
bb36caca0319df4c8d6e3d978045b7c28369f043
Subodh Iyengar [Tue, 31 May 2016 04:08:33 +0000 (21:08 -0700)]
Add support for TFO connections
Summary:
This adds support to establish connections
over TFO.
The API introduced here retains the same
connect() + write() api that clients currently
use.
If enableTFO() is called then the connect will
be deferred to the first write. This only works
with request response protocols since a write
must trigger the connect. There is a tradeoff here
for the simpler API, and we can address this with
other signals such as a short timeout in the future.
Even though the client might enable TFO, the program
might run on machines without TFO support.
There were 2 choices for supporting machines where
TFO might not be enabled:
1. Fallback to normal connect if tfo sendmsg fails
2. Check tfo supported on the machine before using it
Both these have their tradeoffs, however option 1 does
not require us to read from procfs in the common code
path.
Reviewed By: Orvid
Differential Revision:
D3327480
fbshipit-source-id:
9ac3a0c7ad2d206b158fdc305641fedbd93aa44d
Christopher Dykes [Mon, 30 May 2016 23:05:21 +0000 (16:05 -0700)]
Switch uses of networking headers to <folly/portability/Sockets.h>
Summary:
Diff #14 of 14.
This just switches the includes of the headers, it does not make the adjustments actually required to work for MSVC.
Reviewed By: mzlee
Differential Revision:
D3249009
fbshipit-source-id:
434f58a7b8f4d0b7d142bc6a4abed8d95c953a7f
Christopher Dykes [Mon, 30 May 2016 23:05:20 +0000 (16:05 -0700)]
Switch uses of <sys/uio.h> to <folly/portability/SysUio.h>
Summary: Diff #13 of 14.
Reviewed By: mzlee
Differential Revision:
D3116264
fbshipit-source-id:
80e498b7062f40d723e156901bf62a2101a6e6ba
Christopher Dykes [Mon, 30 May 2016 23:05:19 +0000 (16:05 -0700)]
Switch uses of <unistd.h> to <folly/portability/Unistd.h>
Summary: Diff #12 of 14.
Reviewed By: mzlee
Differential Revision:
D3187000
fbshipit-source-id:
ad56ca8ff20fa88949554b61545ee1bf982b0569
Christopher Dykes [Mon, 30 May 2016 23:05:18 +0000 (16:05 -0700)]
Include <folly/portability/PThread.h> where needed
Summary: Diff #11 of 14.
Reviewed By: mzlee
Differential Revision:
D3262478
fbshipit-source-id:
d06eee44d1ccefc78a381e44c400c0242b5e0e5a
Christopher Dykes [Mon, 30 May 2016 23:05:17 +0000 (16:05 -0700)]
Include <folly/portability/Time.h> where needed
Summary: Diff #10 of 14
Reviewed By: mzlee
Differential Revision:
D3270909
fbshipit-source-id:
2c644ca5db4ecf0387b167d2f6ecc62ddf687a56
Christopher Dykes [Mon, 30 May 2016 23:05:16 +0000 (16:05 -0700)]
Include <folly/portability/Fcntl.h> where needed
Summary: Diff #9 of 14.
Reviewed By: mzlee
Differential Revision:
D3262436
fbshipit-source-id:
60c5df03e91163c651bc6e5e0d38ba8b46b8c723
Christopher Dykes [Mon, 30 May 2016 23:05:15 +0000 (16:05 -0700)]
Switch uses of <sys/syscall.h> to <folly/portability/SysSyscall.h>
Summary: Diff #8 of 14.
Reviewed By: mzlee
Differential Revision:
D3186943
fbshipit-source-id:
3b631faac36f3ec6e7fce031f07d8a98c76d7908
Christopher Dykes [Mon, 30 May 2016 23:05:13 +0000 (16:05 -0700)]
Switch uses of <dirent.h> to <folly/portability/Dirent.h>
Summary: Diff #5 of 14.
Reviewed By: mzlee
Differential Revision:
D3186885
fbshipit-source-id:
1947656137cad62a019cdf687cff9913bcecde2a
Christopher Dykes [Mon, 30 May 2016 23:05:12 +0000 (16:05 -0700)]
Switch uses of <sys/resource.h> to <folly/portability/SysResource.h>
Summary: Diff #4 of 14.
Reviewed By: mzlee
Differential Revision:
D3181101
fbshipit-source-id:
d0db67ae19ee71e6a251c299ef672baead3ab9d9
Christopher Dykes [Mon, 30 May 2016 23:05:11 +0000 (16:05 -0700)]
Switch uses of <sys/mman.h> to <folly/portability/SysMman.h>
Summary: Diff #3 of 14.
Reviewed By: mzlee
Differential Revision:
D3110299
fbshipit-source-id:
39e4980df14b0b108b20dbbe57f2c5c6b361ea2e
Christopher Dykes [Mon, 30 May 2016 23:05:10 +0000 (16:05 -0700)]
Switch uses of <sys/file.h> to <folly/portability/SysFile.h>
Summary: Diff #2 of 14.
Reviewed By: mzlee
Differential Revision:
D3115972
fbshipit-source-id:
c75cbc95f37e8f2c2bf81311a8ca08ee766fc107
Christopher Dykes [Mon, 30 May 2016 23:05:07 +0000 (16:05 -0700)]
Include <folly/portability/SysTime.h> rather than <sys/time.h>
Summary:
This changes includes of `<sys/time.h>` to `<folly/portability/SysTime.h>`.
The ultimate goal of all of this is to get HHVM compiling on Windows with MSVC 2015. At the moment there are no plans to support MCRouter, Squangle, or AFDT under Windows. There are also no current plans to support MinGW or Cygwin.
Now, on to the headers themselves. There are three primary kinds of portability headers.
- Replacement headers. These are headers such as `<sys/time.h>` that MSVC simply does not have. These headers should always be included instead of the headers they are replacing.
- Additive headers. These are headers, such as `<time.h>` that MSVC has, but are either missing functions, or has the functions but under a different name. The portability headers for these only needs to be included instead of the original header if you are using a function that MSVC does not have. The reason for this is that forcing the use of the portability header for `<stdlib.h>` just because MSVC doesn't have `mkdtemp` and `realpath` would just be messy.
- Utility headers. These are portability headers, such as `Memory.h` that exist to provide a portable interface to various things, in this case the allocation and freeing of aligned memory. They are not implementing any existing API, and all contents are in the `folly` namespace.
The API's implemented in the first and second types of headers are intended as drop-in replacements for the Linux versions, however they only implement the functions and features that are actually needed, and are not intended as feature-complete replacements. They are not intended for use outside of Facebook projects. Any API's expecting text input expect them to be in the same encoding as the Posix API's define, in most cases that means UTF-8. Anywhere in the portability headers that fails to account for this is a bug, and should be pointed out in the diff (if it's still in review), or else by filing a task detailing the issue and assigning it to me.
For all headers, if there is a mechanism in C++11, or else in Folly itself, that allows the same functionality, that mechanism should be used instead of adding a function to the portability headers.
Reviewed By: swtaarrs, yfeldblum
Differential Revision:
D3095643
fbshipit-source-id:
3a5c0029f95b03ea320725df88e14a67ca38a445
Adam Simpkins [Sat, 28 May 2016 00:48:14 +0000 (17:48 -0700)]
add Cursor::readWhile() and skipWhile()
Summary:
Add generic functions for reading or skipping until a predicate check fails.
This will allow us to simplify a few different call sites that have their own
logic similar to this. Also change readTerminatedString() to use
readWhile().
Reviewed By: alandau
Differential Revision:
D3337581
fbshipit-source-id:
9f50914c83adfc882219046862972661bed0e72a
Ben Hamilton [Sat, 28 May 2016 00:01:53 +0000 (17:01 -0700)]
Fix test/ProducerConsumerQueueTest.cpp build break
Summary:
The Folly `make check` build is broken with Clang on OS X.
This diff fixes the following break in `test/ProducerConsumerQueueTest.cpp`:
In file included from ProducerConsumerQueueTest.cpp:17:
../../folly/ProducerConsumerQueue.h:55:18: error: no member named 'bad_alloc' in namespace 'std'
throw std::bad_alloc();
The problem is a missing `#include <memory>` in `folly/ProducerConsumerQueue.h`.
Reviewed By: djwatson
Differential Revision:
D3361252
fbshipit-source-id:
d740914f54ede5617abc16437ab2e20a7376f6e5
Ben Hamilton [Sat, 28 May 2016 00:00:50 +0000 (17:00 -0700)]
Fix test/FunctionTest.cpp build breaks
Summary:
The Folly `make check` build is broken with Clang on OS X.
This diff fixes two breaks in `test/FunctionTest.cpp`:
1. `FunctionTest.cpp:39:20: error: implicit instantiation of undefined template 'std::__1::array<int, 100>'`
2. `FunctionTest.cpp:103:3: error: no matching conversion for functional-style cast from 'Function<int (int)>' to '::testing::AssertionResult'`
The first was a missing `#include <array>`, and the second is a workaround for this gtest bug:
https://github.com/google/googletest/issues/429
Reviewed By: djwatson
Differential Revision:
D3361188
fbshipit-source-id:
8140de978a2cbaf0f4aab45781ce0d656f03202b
Ben Hamilton [Fri, 27 May 2016 23:12:47 +0000 (16:12 -0700)]
Fix open source `make check` build broken by `ThreadedExecutorTest.cpp` moving to `wangle`
Summary:
The Folly `make check` build is broken with Clang on OS X.
This diff fixes the following break in `make check`:
make[3]: *** No rule to make target `../futures/test/ThreadedExecutorTest.cpp', needed by `../futures/test/ThreadedExecutorTest.o'. Stop.
The problem is a reference to a file which was moved to `wangle` in
D3286988.
Reviewed By: simpkins
Differential Revision:
D3361287
fbshipit-source-id:
a9ae794374982b4accf806bd9608988430c75cfd
Yang Chi [Fri, 27 May 2016 22:33:40 +0000 (15:33 -0700)]
Set bufferCallback_ back to nullptr upon HTTPSession shutdown
Summary: AsyncSocket::handleWrite may trigger HTTPSession::onWriteSuccess which may totally delete the HTTPSession. AsyncSocket::handleWrite also calls bufferCallback_->onEgressBufferCleared(). That bufferCallback_ is the HTTPSession. So in this diff resets bufferCallback_ to nullptr upon HTTPSession shutdown. afrind had a suggestion of adding a DestructorGuard in AsyncSocket::handleWrite. But i don't know DestructorGuard that well. I think it will keep AsyncSocket from being deleted but not the HTTPSession? Or maybe I can DestructorGuard dg(bufferCallback_) ?
Reviewed By: afrind
Differential Revision:
D3311058
fbshipit-source-id:
cdb5fcbd53837a3effbc096eab87fca4e5d17291
Subodh Iyengar [Fri, 27 May 2016 17:36:48 +0000 (10:36 -0700)]
Fix test for corking
Summary:
Min rto is now 50ms, so change the
test accordingly
Reviewed By: w-o-o
Differential Revision:
D3357140
fbshipit-source-id:
5a807fabc1a3c216a5120b8eb4930e5c1eb52b7a
Giuseppe Ottaviano [Fri, 27 May 2016 16:49:40 +0000 (09:49 -0700)]
Use exponential growth in fbstring::append(size_type n, value_type c)
Summary:
Contrary to the other `append` flavors, `fbstring::append(size_type n, value_type c)`
does not use exponential growth, which can trigger quadratic behavior.
Reviewed By: philippv
Differential Revision:
D3357315
fbshipit-source-id:
2301ed1a9953544368663107113890a25c6621ae
Christopher Dykes [Fri, 27 May 2016 16:48:20 +0000 (09:48 -0700)]
Disable a couple of warnings for MSVC in folly::Function
Summary:
MSVC generates warnings if you define multiple assignment operators or multiple copy constructors on the same type. This is deliberate in `folly::Function`, so just disable the warnings for the struct instead.
Note that the warnings have to be disabled for the entire struct for them to actually be disabled. Disabling them just around the declarations of the constructors and operators in question does not actually disable the warnings.
Reviewed By: yfeldblum
Differential Revision:
D3347746
fbshipit-source-id:
abb53a1e62bcfb7ce02759a7ce8637d824a82081
Adam Simpkins [Thu, 26 May 2016 22:38:13 +0000 (15:38 -0700)]
move io/Cursor-defs.h to io/Cursor.cpp
Summary:
The two function definitions in this file are both normal function definitions,
not template functions. Therefore this should be just a normal .cpp file, and
not a -defs.h header file.
Reviewed By: igorsugak
Differential Revision:
D3348439
fbshipit-source-id:
3ebe49c84493f5aab06c568d9df15a37c2c48836
Adam Simpkins [Thu, 26 May 2016 19:50:38 +0000 (12:50 -0700)]
add a default Subprocess constructor
Summary:
The default constructor creates the Subprocess in an uninitialized state. This
makes it possible to default-construct a Subprocess, but only initialize it
later using the move assignment operator.
Even before this diff, it was possible to have Subprocess objects in
uninitialized states after using the move assignment operator to move the data
out of a Subprocess object.
Reviewed By: yfeldblum
Differential Revision:
D3348490
fbshipit-source-id:
aa6acef9be770de8f0ee118da294cb134f04a466
Christopher Dykes [Thu, 26 May 2016 19:25:43 +0000 (12:25 -0700)]
Disable the signal stack on Windows
Summary: There is no equivalent functionality for Windows, so disable it completely.
Reviewed By: andriigrynenko
Differential Revision:
D3348481
fbshipit-source-id:
52a104f5a8013113f54d5b10c2f0f9b720eeb4f6
Christopher Dykes [Thu, 26 May 2016 17:10:47 +0000 (10:10 -0700)]
Remove the declaration of TimeseriesHistogram::getBucketIdx
Summary: Because, as MSVC rightly tells us, it's not actually defined anywhere.
Reviewed By: yfeldblum
Differential Revision:
D3347654
fbshipit-source-id:
2162bc745136ee0129caa4de78f38e1361083cd1
Christopher Dykes [Thu, 26 May 2016 17:08:57 +0000 (10:08 -0700)]
Include <folly/portability/Windows.h> in folly::format
Summary:
Because apparently I didn't actually add the `Windows.h` include when I added support for MSVC.
This is safe for all platforms because the portability header only includes `Windows.h` when compiling on Windows.
Reviewed By: yfeldblum
Differential Revision:
D3347785
fbshipit-source-id:
4f3e4e6c11953e2588b6fed287324a92e3ef8873
Christopher Dykes [Thu, 26 May 2016 17:08:13 +0000 (10:08 -0700)]
Fix doVecOperation in the SysUio portability header
Summary: Apparently I failed to accurately transfer this particular change to fbcode, which meant it didn't actually compile on Windows.
Reviewed By: yfeldblum
Differential Revision:
D3347976
fbshipit-source-id:
775682f10ab68672f6463ccc69a222ae0a52a097
Adam Simpkins [Wed, 25 May 2016 19:07:07 +0000 (12:07 -0700)]
add Cursor::peekBytes()
Summary:
The existing Cursor::peek() method pre-dates the existence of folly::ByteRange,
and so it returns a std::pair containing a pointer and length. This is usually
more awkward to use than ByteRange.
This adds a peekBytes() method that returns a ByteRange object, and updates all
users of peek() in folly to use peekBytes() instead. Eventually I think we
should add a FOLLY_DEPRECATED attribute to peek(). I will wait to add this tag
until after converting a few other projects to use peekBytes(), though.
Reviewed By: alandau
Differential Revision:
D3337302
fbshipit-source-id:
14220a059d915bf5adc66b8b283f7228b796a4dc
Jon Maltiel Swenson [Tue, 24 May 2016 22:41:25 +0000 (15:41 -0700)]
std::make_unique -> folly::make_unique
Summary:
We should use folly::make_unique instead of std::make_unique in order to support
compilers without support for C++14.
Reviewed By: pavlo-fb
Differential Revision:
D3343288
fbshipit-source-id:
9150af752e57988962a7580851ffa32086c669d7
Christopher Dykes [Tue, 24 May 2016 20:57:25 +0000 (13:57 -0700)]
Update generate_varint_tables to support MSVC
Summary: MSVC is a massive pain in this respect, and, after testing many workarounds, and only ending up with it just emitting a dynamic initializer for them, it's easier to just generate the table as a `uint64_t` and load it explicitly.
Reviewed By: yfeldblum
Differential Revision:
D3270226
fbshipit-source-id:
77bc84e58d393373de05a28a30dcb80b66c09c9f
Joseph Griego [Tue, 24 May 2016 20:14:54 +0000 (13:14 -0700)]
Clear old fs::path when moving a TemporaryDirectory
Summary:
If a TemporaryDirectory with scope DELETE_ON_DESTRUCTION gets moved
and then goes out of scope, the directory will be deleted while the moved
object still holds that path; we just clear the path from the old object to
prevent this
Reviewed By: yfeldblum
Differential Revision:
D3331079
fbshipit-source-id:
9c99413661e2436ccec937d3fa652da810133b34
Chad Parry [Tue, 24 May 2016 18:05:14 +0000 (11:05 -0700)]
Switch HHWheelTimer::SharedPtr to a standard shared pointer
Summary:
As part of my campaign to get `HHWheelTimer` away from intrusive ref-counting semantics, (cf.
D3237530), I am redefining `HHWheelTimer::SharedPtr`. It is now a true `std::shared_ptr`. This will break clients that attempt the following:
HHWheelTimer::UniquePtr timer1{HHWheelTimer::newTimer()};
HHWheelTimer::SharedPtr timer2{timer1};
In the past, that code would compile, because `timer2` could use the intrusive ref-counting, and `timer1` would still be valid. After this change, the second timer would need to be initialized with `std::move(timer1)` instead. The `UniquePtr` is starting to actually be unique.
There is only one place in the code that actually tries to do such a copy. It's in a test, which I have changed. Any other instances of that behavior would create a compilation error, so it's impossible for one to be overlooked.
Reviewed By: djwatson
Differential Revision:
D3337038
fbshipit-source-id:
085e4da41c9a142d253a1ac0b1dd0fc508dff704
Christopher Dykes [Tue, 24 May 2016 00:30:29 +0000 (17:30 -0700)]
Support folly::assume under MSVC
Summary: Actually __assume when we assume.
Reviewed By: meyering
Differential Revision:
D3336033
fbshipit-source-id:
6c931e9f832dd1d861d03e3ad8e308f612522401
Yedidya Feldblum [Mon, 23 May 2016 06:35:19 +0000 (23:35 -0700)]
Avoid std::initializer_list ctor in folly::dynamic::array
Summary:
[Folly] Avoid `std::initializer_list` ctor in `folly::dynamic::array`.
This will help deprecate the dangerous initializer-list constructor. The init-list construction is dangerous because its semantics vary by compiler and by compiler version.
Reviewed By: ot
Differential Revision:
D3333148
fbshipit-source-id:
031e7593b7e28b299d5eeeb5ce7aed9216dd3a5c
Andrii Grynenko [Fri, 20 May 2016 22:55:24 +0000 (15:55 -0700)]
Make outstanding LoopKeepAlive hold EventBase destructor
Summary:
LoopKeepAlive handle can be grabbed to signal that some external event will later be scheduled on the EventBase via runInEventBaseThread. Usually the code which will be calling runInEventBaseThread only has a raw pointer to an EventBase, so it doesn't have any way to know it was destroyed.
This change ensures that EventBase destructor will keep running the event loop until all such LoopKeepAlive handles are released.
Reviewed By: yfeldblum
Differential Revision:
D3323835
fbshipit-source-id:
4071dae691a61dfebe2f1759cf99f661b161fa4a
Christopher Dykes [Fri, 20 May 2016 21:26:38 +0000 (14:26 -0700)]
Fix the detection of preadv and pwritev on OSX in OSS
Summary:
The configure script doesn't define `FOLLY_HAVE_PREADV` if `preadv` isn't supported, so these checks were wrong. Undefined defines expand to `0`, so this is safe.
Fixes https://github.com/facebook/folly/issues/412
Reviewed By: mzlee, yfeldblum
Differential Revision:
D3329405
fbshipit-source-id:
faee8951c68d4c67e06e7720dfc142e63a9dbd9f
Shubhanshu Agrawal [Fri, 20 May 2016 05:03:16 +0000 (22:03 -0700)]
templating folly::fibers::await by Baton
Summary:
This diff templates folly::fibers::await by Baton.
This would be helpful in passing in custom BatonType's, which is needed for
D3007734
Depends on:
D3314891
Reviewed By: andriigrynenko
Differential Revision:
D3314925
fbshipit-source-id:
9052dc503b9509f16cd41b5e3ede0479a98067aa
Alejandro [Thu, 19 May 2016 16:19:54 +0000 (09:19 -0700)]
Promote memory_order_consume to memory_order_acquire
Summary:
The current use case of `std::memory_order_consume` in this project was intended to provide the appropriate synchronization in cases where a consumer spins on `while( spsc_queue.empty() ) {} `, and then attempts to use an element of the queue since the loop was broken out of, according to comments [here](https://reviews.facebook.net/D48141). Consume semantics do not provide this guarantee according to the standard since there is no data dependency from the producer that can be carried to the consumer by doing a load-consume from the corresponding functions. What is needed is a load-acquire. Current compilers promote `memory_order_consume` to `memory_order_acquire`. Thus, this example appears to work simply due to the promotion from consume to acquire, but would fail to generate the right synchronization instructions on weaker architectures once `memory_order_consume` is implemented as intended. Therefore, the `memory_order` should be tightened to `memory_order_acquire` to guarantee visibility to the consumer
Closes https://github.com/facebook/folly/pull/381
Reviewed By: djwatson
Differential Revision:
D3173574
Pulled By: nbronson
fbshipit-source-id:
b109be2e74f016750a3fe1f848e3fc66e609b9d2
Yedidya Feldblum [Thu, 19 May 2016 09:20:24 +0000 (02:20 -0700)]
Extract Try to top-level
Summary:
[Folly] Extract `Try` to top-level.
It was in `folly/futures/`, but this diff moves it to `folly/`.
It is needed for futures, but it is are more general than futures and can be used separately.
Use `folly/Try.h` instead of `folly/futures/Try.h`.
Also fixes up all `#include` sites:
hg grep -lw folly/futures/Try | xargs perl -pi -e 's,\bfolly/futures/Try\b,folly/Try,g'
Reviewed By: markisaa
Differential Revision:
D3309908
fbshipit-source-id:
cdf13f0ac0b0e36aa07e0f1d04870d06500c5874
Lucian Grijincu [Thu, 19 May 2016 08:14:49 +0000 (01:14 -0700)]
folly: FBString: avoid ubsan in assert
Summary:
```
buck-out/gen/folly/__default_headers__#default,headers/folly/FBString.h:365:7: runtime error: null pointer passed as argument 2, which is declared to never be null
third-party-buck/build/glibc/include/string.h:70:33: note: nonnull attribute specified here
#0 0x433a39 in _ZZN5folly13fbstring_coreIcEC1EPKcmbENKUlvE_clEv ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////buck-out/gen/folly/__default_headers__#default,headers/folly/FBString.h:365:7
#1 0x4335a9 in _ZN5folly14ScopeGuardImplIZNS_13fbstring_coreIcEC1EPKcmbEUlvE_E7executeEv ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////buck-out/gen/folly/__default_headers__#default,headers/folly/ScopeGuard.h:153:29
#2 0x4335a9 in _ZN5folly14ScopeGuardImplIZNS_13fbstring_coreIcEC1EPKcmbEUlvE_ED2Ev ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////buck-out/gen/folly/__default_headers__#default,headers/folly/ScopeGuard.h:130
#3 0x4335a9 in folly::fbstring_core<char>::fbstring_core(char const*, unsigned long, bool) ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////buck-out/gen/folly/__default_headers__#default,headers/folly/FBString.h:427
#4 0x4353fa in folly::basic_fbstring<char, std::char_traits<char>, std::allocator<char>, folly::fbstring_core<char> >::basic_fbstring(char const*, unsigned long, std::allocator<char> const&) ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////buck-out/gen/folly/__default_headers__#default,headers/folly/FBString.h:1038:9
```
Reviewed By: Gownta
Differential Revision:
D3321266
fbshipit-source-id:
28d5aef16e91a98066a1de6bab95403fbc63eaab
Andrii Grynenko [Thu, 19 May 2016 05:57:40 +0000 (22:57 -0700)]
Move fibers out of experimental
Summary: folly::fibers have been used by mcrouter for more than 2 years, so not really experimental.
Reviewed By: pavlo-fb
Differential Revision:
D3320595
fbshipit-source-id:
68188f92b71a4206d57222993848521ca5437ef5
Shubhanshu Agrawal [Thu, 19 May 2016 03:42:24 +0000 (20:42 -0700)]
adding Promise::await
Summary:
This diff adds a static await method to Promise.
Also after this, folly::fibers::await is just a wrapper around Promise::await.
This also removes the friend relationship between Promise and await.
This would be helpful in making the folly::fibers::await templated by Baton changes easier to make.
Reviewed By: andriigrynenko
Differential Revision:
D3314891
fbshipit-source-id:
361546c078caafd067734d2f474c617d7fb888b0
Yedidya Feldblum [Thu, 19 May 2016 00:39:53 +0000 (17:39 -0700)]
Make Try independent of Future
Summary:
[Folly] Make `Try` independent of `Future`.
`Try` is useful for futures, but it can be used independently from futures as well.
Reviewed By: djwatson
Differential Revision:
D3319130
fbshipit-source-id:
badfcaa482dee6e7bdef14a501b4efc91634fa51
Andrii Grynenko [Thu, 19 May 2016 00:11:24 +0000 (17:11 -0700)]
Export documentation from dex to folly
Reviewed By: yfeldblum
Differential Revision:
D3317666
fbshipit-source-id:
f5558b117c8da74789ee444cdaa00231e75dc31e
iivlev [Wed, 18 May 2016 23:56:35 +0000 (16:56 -0700)]
Fixing typo in Readme
Summary: Closes https://github.com/facebook/folly/pull/403
Reviewed By: yfeldblum
Differential Revision:
D3320043
Pulled By: elliottneilclark
fbshipit-source-id:
5d21421216180fb601015a40a3de79e3dd2a162e
Yedidya Feldblum [Wed, 18 May 2016 21:46:49 +0000 (14:46 -0700)]
Flesh out the comments about Unit
Summary:
[Folly] Flesh out the comments about `Unit`.
* Remove mention of `Future`.
* Compare `unit` to `void`.
Reviewed By: djwatson
Differential Revision:
D3316114
fbshipit-source-id:
d30cf36cc029025734bb16c5bdbd640a289cad79
Yedidya Feldblum [Wed, 18 May 2016 03:38:11 +0000 (20:38 -0700)]
Extract Unit to top-level
Summary:
[Folly] Extract `Unit` to top-level.
It was in `folly/futures/`, but this diff moves it to `folly/`. It is needed for futures, but it is more general than futures and can be used separately.
Users must replace `folly/futures/Unit.h` with `folly/Unit.h`.
Also codemods existing usage sites:
```
hg grep -lw folly/futures/Unit.h | xargs perl -pi -e 's,\bfolly/futures/Unit\.h\b,folly/Unit.h,g'
```
Reviewed By: igorsugak
Differential Revision:
D3314280
fbshipit-source-id:
16279b76b1d24529bec49077ccb36cd7d39f23b8
Ilya Maykov [Wed, 18 May 2016 01:53:27 +0000 (18:53 -0700)]
added IOBuf::wrapBufferAsValue()
Summary:
Added a version of IOBuf::wrapBuffer() which returns the new IOBuf by value rather than heap-allocating it.
Motivation: we have a lot of places in the crypto code that do something like this:
// take a string or vector parameter
auto buf = folly::IOBuf::wrapBuffer( /* string or vector goes here */);
// do something with buf
// return
In these cases, a stack-allocated IOBuf would save us from an unnecessary heap allocation. But calling `folly::IOBuf(folly::IOBuf::WrapBufferOp::WRAP_BUFFER, ...)` is gross and in practice people just call the much more readable `wrapBuffer()` function. Hypothesis: readability trumps performance, but if we had a readable version that returned a stack-allocated IOBuf, it might see usage.
Reviewed By: yfeldblum
Differential Revision:
D3314037
fbshipit-source-id:
4d4b5ec1d067762a27de1d7d6f7cac406388bfa3
Andrii Grynenko [Wed, 18 May 2016 00:21:56 +0000 (17:21 -0700)]
Fix one more race in Promise::setTry
Summary:
Baton::post has to be the last step of Promise::setTry. If Promise is not owned by
the posting thread, it may be destroyed right after Baton::post is called.
Reviewed By: pavlo-fb
Differential Revision:
D3312506
fbshipit-source-id:
6d5c3f0925c3bb9cd3f981e7550f888d5ed76189
Yedidya Feldblum [Wed, 18 May 2016 00:04:40 +0000 (17:04 -0700)]
Keep the Unit test suite free of Promise and Future
Summary:
[Folly] Keep the `Unit` test suite free of `Promise` and `Future`.
Move the `Promise`-related tests to the `Promise` test suite and the `Future`-related tests to the `Future` test suite.
Reviewed By: djwatson
Differential Revision:
D3313635
fbshipit-source-id:
05c82c8719719d7709063ad58a4806036ca10fb3
Yedidya Feldblum [Tue, 17 May 2016 22:27:43 +0000 (15:27 -0700)]
Simplify Unit
Summary:
[Folly] Simplify `Unit`.
Specifically:
* Make `Unit::Lift` and `Unit::Drop` simply be aliases to `std::conditional<...>`. No need for anything more complicated.
* Remove `is_void_or_unit`, which is unused and unnecessary.
Reviewed By: djwatson
Differential Revision:
D3312481
fbshipit-source-id:
280d40aa8ef91c52f96a51b03e0a109d76c939ec
Marcus Holland-Moritz [Tue, 17 May 2016 20:08:17 +0000 (13:08 -0700)]
folly: fbvector: ubsan: don't call memcpy() with nullptr args if size == 0
Reviewed By: meyering, luciang
Differential Revision:
D3310683
fbshipit-source-id:
c86471d54062b2f3455f15fb73340fac486d9e44
Yedidya Feldblum [Tue, 17 May 2016 01:43:22 +0000 (18:43 -0700)]
IPAddress::validate
Summary: [Folly] `IPAddress::validate`.
Reviewed By: igorsugak
Differential Revision:
D3308683
fbshipit-source-id:
48af18d6930f16718372021a4cc08062bf17327e
Andrii Grynenko [Mon, 16 May 2016 21:06:54 +0000 (14:06 -0700)]
Fix a race in Promise::setTry
Summary:
This fixes a race which can only be exposed if Promise is owned by the same thread which was calling await and Promise is fulfilled from a different thread. What could happen then was:
1. Thread 2 fulfills the promise and thus saves the value and posts Baton.
2. Thread 1 wakes up, await returns and then Thread 1 destroys the Promise.
3. Promise destructor checks value_, which is still not null, so it tries to fulfil it with exception.
Reviewed By: spalamarchuk
Differential Revision:
D3306969
fbshipit-source-id:
adf799cfd7b75b0201fa675a9e44ac7c7c42ac95
Andrii Grynenko [Mon, 16 May 2016 20:13:24 +0000 (13:13 -0700)]
Make await exception safe
Summary: This fixes fibers::await to correctly handle exception thrown by user-passed lambda. Await still always waits for the promise to be fulfilled (if the promise was not moved out, it will be destroyed and thus auto-fulfilled with "promise not fulfilled" exception). However if user-passed lambda throws, promise result is ignored (even if exception) and exception thrown by lambda is re-thrown.
Reviewed By: pavlo-fb
Differential Revision:
D3303393
fbshipit-source-id:
c4ba01fde0e156cc88e5d07aaf763e3abe121d11
Yedidya Feldblum [Sat, 14 May 2016 19:49:52 +0000 (12:49 -0700)]
Prefer constexpr to preprocessor conditionals when checking endianness
Summary:
[Folly] Prefer `constexpr` to preprocessor conditionals when checking endianness.
Using `if (folly::kIsLittleEndian) {` v.s. `#if __BYTE_ORDER__ == __LITTLE_ENDIAN__`.
Reviewed By: meyering
Differential Revision:
D3296770
fbshipit-source-id:
b26df83fdd42a50663746fc7c9d5fbe67e6671eb
Yedidya Feldblum [Sat, 14 May 2016 11:29:37 +0000 (04:29 -0700)]
A thread-per-task executor
Summary:
[Wangle] A thread-per-task executor.
Moved from Folly into Wangle and fleshed out.
* Starts task threads from a control thread, rather than starting task threads from the submitter thread. Because starting task threads is likely to be more expensive than moving a functor.
* Joins task threads as they finish, rather than joining all task threads in the executor's destructor.
Suitable for running tasks which spend most of their time sleeping. Such as blocking IO, blocking fork-exec-wait, etc., when it is inconvenient to use the nonblocking variants with an IO executor.
Reviewed By: simpkins
Differential Revision:
D3286988
fbshipit-source-id:
4b91133a7d55332ebbae020c1515c60e816906b3
Lucian Grijincu [Sat, 14 May 2016 07:38:50 +0000 (00:38 -0700)]
folly: fbstring: ubsan: memcpy/memmove are marked as nonnull - avoid calling them when size == 0 and (likely) source is nullptr
Summary: Also see {
D3295811}
Differential Revision:
D3302564
fbshipit-source-id:
3f2dbf5a6cfa8199682cb4af90aac372d501919a
Lucian Grijincu [Sat, 14 May 2016 07:09:41 +0000 (00:09 -0700)]
folly: ubsan: elias-fano avoid memset(nullptr)
Reviewed By: ot
Differential Revision:
D3300437
fbshipit-source-id:
7e0add5dab6dc1c891b742b4bcedea656c8284a9
Lucian Grijincu [Sat, 14 May 2016 00:23:32 +0000 (17:23 -0700)]
folly: ubsan: redo BitsTest (no macros) and fix DCHECK in test
Reviewed By: meyering
Differential Revision:
D3283226
fbshipit-source-id:
7ae2ec7741ca53c494e8325e30f2995a241674c0
Mirek Klimos [Fri, 13 May 2016 21:17:12 +0000 (14:17 -0700)]
Test creation of RequestContext inside a fiber task
Summary: ^, as discussed on
D3279644
Reviewed By: andriigrynenko
Differential Revision:
D3299473
fbshipit-source-id:
dc302132db64399f768d69c641f848d0da1075c4
Mirek Klimos [Fri, 13 May 2016 19:02:35 +0000 (12:02 -0700)]
Replace RequestContext::setContext with RequestContextScopeGuard in folly
Summary: To make sure RequestContext is properly unset when we stop processing request on a thread. This changes the API in Fibers, NotificationQueue, HHWheelTimer and AsyncTimeout, and fixes RequestContext handling in Futures (reset RC after the callback is done)
Reviewed By: andriigrynenko
Differential Revision:
D3279644
fbshipit-source-id:
a6a1c2840cdce179334aa1a3b28fa514cd5358c1
Chad Parry [Fri, 13 May 2016 18:56:03 +0000 (11:56 -0700)]
Construct all HHWheelTimer instances with the factory method
Summary:
This diff standardizes `HHWheelTimer` construction. It's not that safe to allow anyone to call `new HHWheelTimer` because the caller needs to remember to call `DelayedDestruction::destroy`.
Once callers are made to be safer, I'll be able to change the `HHWheelTimer` to use standard smart pointers instead of `DelayedDestruction`.
This picks up some work that I started in
D2627941 but had to postpone.
This was mostly a mechanical change:
( fbgs -ls 'new HHWheelTimer' ; fbgs -ls 'new folly::HHWheelTimer' ) | xargs perl -pi -e 's/\bnew\s+((?:folly::)?HHWheelTimer)\b/$1::newTimer/g'
Then I manually inspected the code. I reverted `folly/io/async/HHWheelTimer.h`, since the `newTimer` factory method is the one place that we still want to call `new HHWheelTimer`. I manually edited `proxygen/lib/utils/SharedWheelTimer.cpp`, since that was using a `shared_ptr` with a custom destructor, which isn't needed anymore. I reverted `common/io/async/d` since the D shim is meant to pass around only raw pointers. I had to replace each instance of `foo.reset(…)` with `foo = …` . Then I made manual edits to `common/clientpool/ClientPool2-inl.h` because that code wants to store the timer in a `ThreadLocalPtr`.
Reviewed By: yfeldblum
Differential Revision:
D3237530
fbshipit-source-id:
96bfb6987098618ad5430c21b1314f385f04a93d
Mark Santaniello [Fri, 13 May 2016 15:41:37 +0000 (08:41 -0700)]
Suppress new GCC 5 warning -Wbool-compare
Summary: This -Wbool-compare is apparently new in GCC 5. We should ignore it, similar to the existing suppression of -Wsign-compare.
Reviewed By: pixelb, yfeldblum
Differential Revision:
D3293675
fbshipit-source-id:
5f90b0dbf049a06c0f6c3df2cdfcc8055358e367
Subodh Iyengar [Fri, 13 May 2016 07:24:05 +0000 (00:24 -0700)]
Fix extra bytes in err message
Summary:
Error message might be < 256 bytes
in which case buf.size() would be greater
than the size of the message written out
and we might have garbled bytes at the end.
ERR_error_string_n null terminates the string
so just use that.
Reviewed By: anirudhvr
Differential Revision:
D3297320
fbshipit-source-id:
2ae626ce4c49ca478806c0bcc40a390e5a84f24b
Andrii Grynenko [Thu, 12 May 2016 23:20:40 +0000 (16:20 -0700)]
Register custom stack for signal handlers
Summary: This updates FiberManager to automatically register custom stack for signal handlers. This ensures signal handlers will work correctly not only in case of stack overflow, but also if fiber stack size is just not enough for them.
Reviewed By: yfeldblum
Differential Revision:
D3289912
fbshipit-source-id:
44959b727f70e1f1748d837b1a06c7414433e5ec
Chad Parry [Thu, 12 May 2016 21:47:19 +0000 (14:47 -0700)]
Transfer ownership from a unique_ptr to a ThreadLocalPtr
Summary:
This `ThreadLocalPtr::reset` overload will accept a `unique_ptr`. It's actually not totally exception safe, simply because `ElementWrapper::set` is not exception safe. The best I can say is that my additional code is exactly as safe as the underlying implemenation.
liketolivedangerously
Reviewed By: ericniebler
Differential Revision:
D3271563
fbshipit-source-id:
774bcf31924b1ed4e29a6cb1c0a36ad710ab6034
Chad Parry [Thu, 12 May 2016 18:46:42 +0000 (11:46 -0700)]
Prevent leaks in ThreadLocalPtr initialization
Summary: While making an unrelated change, (
D3271563, which was needed from an unrelated change, (
D3237530)), I noticed a lack of exception safety here. If `std::bad_alloc` were thrown, then we don't want to leak memory.
Reviewed By: ericniebler
Differential Revision:
D3271911
fbshipit-source-id:
0d316c0fa865a7d64622c1d62160bb0c2b061d78