The Q(Basic)Mutex fast paths are entirely inline in the caller, which
means we need to annotate its operations directly or TSAN doesn't know
what's going on. Also annotate QRecursiveMutex.
The tryLock code could be in principle simplified via a QScopeGuard
but I didn't want to make a central class like QMutex depend on it.
[ChangeLog][QtCore][QMutex] QMutex now has annotations for
ThreadSanitizer.
Change-Id: Ibb130404e63a5ec9bcef9675f9addd16a2c38b7f
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The old code performed one load-acquire (in the assertion) and one
ordered(!) fetch_sub() (in the pre-decrement operator) where one
relaxed fetch_sub() would suffice (it's just a counter).
Fix by caching the result of the relaxed fetch_sub() and performing
the following checks on the cached result.
As drive-bys, rename the atomic member variable (to catch any outside
uses) and make the ctor explicit.
Found by locally disabling the QAtomic<T> -> T implicit conversion
operators.
Change-Id: Ifdc11c2c4807b71f4cab2ba9f5405ace7d8d71a9
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
As all other 20+ users of the state variable, use a relaxed atomic
load under the protection of the mutex. There's no need for the
loadAcquire() that the implicit conversion operator uses under the
hood, because a) we're under mutex protection and b) the state doesn't
guard any other data but itself.
Found by disabling said implicit conversion operators.
Change-Id: I2a76242271cec96175cde503ca883805db6b9212
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Still not complete. Just grepping for static and thread_local.
Task-number: QTBUG-100486
Change-Id: I90ca14e8db3a95590ecde5f89924cf6fcc9755a3
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It expands to the first available of
- constinit (C++20)
- [[clang::require_constant_initialization]] (Clang)
- __constinit (GCC >= 10)
Use it around the code (on and near static QBasicAtomic; this patch
makes no attempt to find all statics in qtbase).
[ChangeLog][QtCore][QtGlobal] Added macro Q_CONSTINIT.
Fixes: QTBUG-100484
Change-Id: I11e0363a7acb3464476859d12ec7f94319d82be7
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
As a drive-by, did also minor refactorings/improvements.
Task-number: QTBUG-98434
Change-Id: I81964176ae2f07ea63674c96f47f9c6aa046854f
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Anton Kudryavtsev <antkudr@mail.ru>
If the QPromise is being destroyed, we should signal the associated
futures to stop waiting. No matter in which state the promise is, if
it's not finished, we should always cancel to avoid infinite waits.
This is also what docs state:
"The promise implicitly transitions to a canceled state on destruction
unless finish() is called beforehand by the user."
Fixes: QTBUG-101284
Pick-to: 6.3 6.2
Change-Id: I65ebfefe03b79b41afacda78a4f49938c54d8b37
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Codechecker complains about passing a forwarding reference to std::move,
which may unexpectedly cause lvalues to be moved. We should use
std::forward instead.
This amends cbf8fc0ac4
Change-Id: Ieba9e960b5dfb448e9cebabe9541967698995b8f
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Explicitly move the argument into the return value to silence:
qfuture_impl.h:153:20: warning: local variable 'arg' will be copied
despite being returned by name [-Wreturn-std-move]
return arg;
^~~
[...]
qfuture_impl.h:153:20: note: call 'std::move' explicitly to avoid copying
Change-Id: I49fab7c31efb7d809182c8d56e6a5ed8f05a5622
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Improve the docs for the recently added move operations.
Change-Id: Iae17f5289d30faab4aaaec14a27036f7a03129ab
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
There is no reason to allow relock() on a locked locker, or unlock()
or an unlocked one, just like we don't allow that on a plain mutex
to begin with. The docs already said that e.g. relock() locks an
_unlocked_ locker.
[ChangeLog][QtCore][QMutexLocker] QMutexLocker allowed relock() and
unlock() on an already closed (resp. open) locker object. These
semantics have always been undocumented and are now unsupported
(in both cases they yield undefined behavior.)
Change-Id: Id5f67beb5dc30d6435dae88a3085fba93ec7d96e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The class is similar to unique_lock in that it allows for unlocking
and relocking. Since the locked state is tracked by QMutexLocker itself,
it's trivial to make it movable.
[ChangeLog][QtCore][QMutexLocker] The class is now movable.
Change-Id: I534044f8024575e996c12efb2236761d493798a3
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Connecting to nullptr, or connecting to a non-signal PMF, would result
in a QFuture which would never finish. Catch these cases and handle
them.
Windows+MSVC for some reason fails the test. I can't entirely understand
why, so I've marked it as XFAIL, with QTBUG-101761 to track it.
Change-Id: I314980e7e9b7156d8cddd3b33d5cbf1d0bcd6116
Pick-to: 6.2 6.3
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
These semantics are deprecated and about to be changed.
Drive by, use nullptr instead of 0.
Change-Id: I213bf7d528d2fa312549f9fa49a5533198eddf88
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QObject's cache the binding status pointer to avoid TLS lookups.
However, when an object is moved to a different thread, we need to
update the cached pointer (as the original thread might stop and thus no
longer exist, and to correctly allow setting up bindings in the object's
thread).
Fix this by also storing the binding status in QThreadPrivate and
updating the object's binding status when moved. This does only work
when the thread is already running, though. If it is not running, we
instead treat the QThreadPrivate's status pointer as a pointer to a
vector of pending objects. Once the QThread has been started, we check
if there are pending objects, and update them at this point.
Pick-to: 6.2 6.3
Fixes: QTBUG-101177
Change-Id: I0490bbbdc1a17cb5f85044ad6eb2e1a8c759d4b7
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Rename isLocked in preparation for a future commit. Rename m
as well for consistency.
Change-Id: I1c8d040bca6825a698ec804ea142d208abacd5cc
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
By changing it to unique_ptr.
Pick-to: 6.2 6.3
Change-Id: I91abb69445b537d4c95983ae735341882352b29d
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
See script in qtbase/util/includeprivate for the rules.
Since these files are being touched anyway, I also ran the
updatecopyright.pl script too.
Change-Id: Ib056b47dde3341ef9a52ffff13ef677e471674b6
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
cf043a785a unified QtConcurrent::run() and
QtConcurrent::runWithPromise(). Remove the last mentions of
runWithPromise. Also did some minor cosmetic improvements.
Pick-to: 6.3
Change-Id: Ie3b39600978ccfa4a009d3ff68567a348dc7b166
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
2dea20e4b0 de-inlined the dtor to avoid
duplicating its vtable in multiple translation units, but the class is
used externally, so we need to export it now.
Pick-to: 6.3
Fixes: QTBUG-100569
Change-Id: I7cf2abdfdeb59bced0631838fe329ba94ab8c73a
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Change ref() and deref() in qatomic_cxx11.h to use acq_rel ordered
semantics as documented instead of the stricter sequential consistency
that was previously used implicitly.
It was pointed out in the discussion on atomic reference counts that
acq_rel semantics should be sufficient for reference counts. It also
turned out that that is all that the definition of "Ordered" in the
documentation really guarantees.
Original-patch-by: Kevin Kofler <kevin.kofler@chello.at>
Change-Id: I6cd50b34bdc699ff8b7c8fe6d303b28d71494731
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
It is an important and useful part of the API contract that the
QFutureWatcher signals will emit complete information after
setFuture() on an already started QFuture. If that wasn't so, some
API clients would need two code paths to gather state and progress
information: one through the signals for not yet started QFutures,
and one through the state querying APIs for already started
QFutures.
The previous documentation was not precise enough to rely on, even
though the implemented behavior - cf.
QFutureInterfaceBasePrivate::connectOutputInterface() -
seems well thought out and implemented to rely on.
Pick-to: 6.3 6.2
Change-Id: Iaf3db37a851f07ac85881adaea5c83612a7fd843
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
If the signal passed to QtFuture::connect() takes multiple arguments,
we need to wrap the arguments in a std::tuple when reporting the result.
To detect this case we were checking if the result type of a QFuture
returned by QtFuture::connect() is a std::tuple, but this was not
correct: the result type could be a std::tuple also if the passed
signal takes a single std::tuple argument. Instead, check if the signal
takes more than one argument.
As a drive-by modified the tst_QFuture::signalConnect to use const
values for tuples used in multiple test-cases, to avoid repetition.
Fixes: QTBUG-100071
Pick-to: 6.2 6.3
Change-Id: I1ce39cf87028f36ef94a9d1a4423b0c51473afd4
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
The class is not intended for inheriting from it (see also
e502906305), so we can mark it as final
to explicitly forbid this. The tests were still using it as a base
class to clean the results during destruction, so fix them accordingly.
Task-number: QTBUG-99883
Pick-to: 6.3
Change-Id: I4a7ee3e2b462bd704e4b5a95ed733144805d6e5b
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
QPromise is now used in qfuture_impl.h, so we need to include it. Remove
qfuture.h include from qpromise.h, to avoid circular dependency.
As a drive-by, simplify a type-trait: is_convertible_v<T, S> is true
whenever is_same_v<T, S> is, so we don't need to instantiate both.
Fixes: QTBUG-100144
Pick-to: 6.3 6.2
Change-Id: Ic6df43d96d9d168cc44c2949e41c5e490f4c50ce
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
There's no class inheriting from ResultStoreBase (and likely won't be),
so the destructor was marked to be made non-virtual in Qt 7. For the
same reason, the internal members don't need to be protected, and the
class shouldn't have "Base" in its name. Add a note about it.
Task-number: QTBUG-99883
Change-Id: I00d7a96d99d2c326d29bd421235a15d68b4d4e5c
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
After QFuture continuations became non-copyable (see earlier commits),
we have to always use ContinuationWrapper to save the continuations
inside std::function, since it requires the callable to be copyable.
Optimize the wrapper, by storing the callable directly (instead of using
a ref-counted QSharedPointer) and introducing a fake copy-constructor
that makes sure that it's never called.
Pick-to: 6.3 6.2
Change-Id: I0ed5f90ad62ede3b5c6d6e56ef58eb6377122920
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This is required to ensure that the continuation attached to a
QFuture returned by QtFuture::when* methods is cleaned in the destructor
of the associated QPromise, so that it doesn't keep any ref-counted
copies to the shared data, thus preventing it from being deleted.
Task-number: QTBUG-99534
Pick-to: 6.3
Change-Id: If4e2929b2e638d6b48c95f0aef9dc886066cedbe
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Continuations were using QFutureInterface to create and return the
associated future to the user. Attaching a continuation to the returned
future could cause memory leaks (described in an earlier commit). Use a
QPromise when saving the continuation, to make sure that the attached
continuation is cleaned in the destructor of the associated QPromise, so
that it doesn't keep any ref-counted copies to the shared data, thus
preventing it from being deleted.
Task-number: QTBUG-99534
Pick-to: 6.3 6.2
Change-Id: I52d5501292095d41d1e060b7dd140c8e5d01335c
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Capturing a QFuture in the continuations attached to it results in
memory leaks. QFuture's ref-counted data can only be deleted when the
last copy referencing the data gets deleted. The saved continuation
that keeps a copy of the future (as in case of the lambda capture) will
prevent the data from being deleted. So we need to manually clean the
continuation after it is run. But this doesn't solve the problem if the
continuation isn't run. In that case, clean the continuation in the
destructor of the associated QPromise.
To avoid similar leaks, internally we should always create futures via
QPromise, instead of the ref-counted QFutureInterface, so that the
continuation is always cleaned in the destructor. Currently QFuture
continuations and QtFuture::when* methods use QFutureInterface directly,
which will be fixed by the follow-up commits.
Fixes: QTBUG-99534
Pick-to: 6.3 6.2
Change-Id: Ic13e7dffd8cb25bd6b87e5416fe4d1a97af74c9b
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This works around mismatch in threads starting and restarting QThreads,
and is safe since we don't need to establish a binding, and objectName
access in QThreadPool is locked behind a mutex.
Pick-to: 6.3 6.2
Fixes: QTBUG-96718
Change-Id: Id3f75e4f8344796ca658899645219fe3373ddd6d
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
qSwap() is a monster that looks for ADL overloads of swap() and also
detects the noexcept of the wrapped swap() function, so it should only
be used when the argument type is unknown. In the vast majority of
cases, the type is known to be efficiently std::swap()able or to have
a member-swap. Call either of these.
For the common case of pointer types, circumvent the expensive trait
checks on std::swap() by providing a hand-rolled qt_ptr_swap()
template, the advantage being that it can be unconditionally noexcept,
removing all type traits instantiations. Don't document it, otherwise
we'd be unable to pick it to 6.2.
Effects on Clang -ftime-trace of a PCH'ed libQt6Gui.so build:
before:
**** Template sets that took longest to instantiate:
[...]
27766 ms: qSwap<$> (9073 times, avg 3 ms)
[...]
2806 ms: std::swap<$> (1229 times, avg 2 ms)
(30572ms)
after:
**** Template sets that took longest to instantiate:
[...]
5047 ms: qSwap<$> (641 times, avg 7 ms)
[...]
3371 ms: std::swap<$> (1376 times, avg 2 ms)
[qt_ptr_swap<$> does not appear in the top 400, so < 905ms]
(< 9323ms)
As a drive-by, remove superfluous inline keywords and template
ornaments.
Task-number: QTBUG-97601
Pick-to: 6.3 6.2
Change-Id: I88f9b4e3cbece268c4a1238b6d50e5712a1bab5a
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
QThreadPool allows method calls from any thread, but QObject does not
so copy objectName so we may use it locally under our own lock.
Pick-to: 6.3 6.2
Task-number: QTBUG-99775
Change-Id: Ib28910649f5d0f9ce698c7da495069635d608d03
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
This is a data race, as the thread accesses QObject::objectName on the
QThread instance while the thread owning the QThread might modify the
objectName.
Instead, make a copy in the QThreadPrivate that can be accessed safely.
Task-number: QTBUG-96718
Pick-to: 6.3 6.2 5.15
Change-Id: I10701551d498993ca5055daf161636bfb648840c
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
The Unix code stores an additional flag, ThreadPriorityResetFlag, in
the Policy enum, but ubsan does not approve:
qthread_unix.cpp:303:30: runtime error: load of value 2147483648, which is not a valid value for type 'Priority'
qthread_unix.cpp:304:75: runtime error: load of value 2147483648, which is not a valid value for type 'Priority'
Fix by making the variable of std::underlying_type_t<Priority>.
The masking and unmasking code can now be simplified, too.
In the Windows version, replace some switch targets with equivalent
ones to keep -Wswitch-like warnings, though I hasten to note that both
switches use a default case, so have anyway implicitly disabled said
warning.
Pick-to: 6.3 6.2 5.15
Change-Id: Ie4ea7d05e2928d2755ad12d36535197f85493191
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Extract function terminate_on_exception() that de-duplicates the
#ifdef'ery around the try/catch and the handling of the pthread
cancellation pseudo-exception.
Apart from de-duplicating complex code, it will also help suppressing
a ubsan false positive, which is why we're picking it all the way to
5.15.
Pick-to: 6.3 6.2 5.15
Change-Id: I99ad2c0618b8dc30801931df09400c6611d9f9e4
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The qgenericatomic.h occurrences are getting in the way of splitting
up qglobal.h, so inline the define.
Task-number: QTBUG-99313
Change-Id: I0d6e3553c5ed4ac9e426bef2fe07d2672d73c671
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Instead of returning the total number of logical processors in the
system, which we may not be allowed to wholly run on, use the affinity
set that the calling thread is allowed to use.
Implemented for Linux and FreeBSD only, with up to 4x the number of
processors than CPU_SETSIZE has as default (that would be 4096 logical
processors on Linux, 1024 on FreeBSD). Implementation for Windows is
possible, but the API there is always limited to 64, so I'm unsure if it
is correct. Darwin (macOS) does not have this capability.
Testing:
$ ./tst_qthread idealThreadCount | grep QDEBUG
QDEBUG : tst_QThread::idealThreadCount() Ideal thread count: 8
$ taskset 3 ./tst_qthread idealThreadCount | grep QDEBUG
QDEBUG : tst_QThread::idealThreadCount() Ideal thread count: 2
[ChangeLog][QtCore][QThread] idealThreadCount() will now return the
number of logical processors that the current process (thread) has
assigned in its affinity set, instead of the total number of processors
in the system. These two numbers can be different if the process is
launched by the parent with a different affinity set, with tools like
Linux's taskset(1) or schedtool(1). This is currently implemented for
Linux and FreeBSD.
Change-Id: I2cffe62afda945079b63fffd16bd086f64f5f314
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Filter out the last argument of type QPrivateSignal from the signal's
arguments passed to QtFuture::connect().
Pick-to: 6.2
Fixes: QTBUG-92501
Change-Id: Idcd6baba1f01fcc94fa64b1c7030a629d01ed7a1
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
They're not atomic anyway. But it does make it possible to use
QAtomicInteger in both bootstrapped and non-bootstrapped code.
Change-Id: I2cffe62afda945079b63fffd16bcc67e082d92ae
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
If one destroys a running QThread, so far the behavior has been to crash
(à la std::thread) -- assuming the thread hasn't already signalled that
it has finished. This behavior is hostile to solutions such as using
QThread::create(), which always require a wait() before destroying the
thread object.
We can use the opportunity to change the behavior without breaking any
valid code. Instead of crashing, inside QThread's destructor we can ask
the new thread to quit, and then join it (à la std::jthread). This
simplifies the implementation of long-living runnables and the code that
manages them.
Deploying this solution for the whole QThread class may not be entirely
painless. While no correct code would work differently with the proposed
changes, incorrect code that deletes a running thread would no longer
crash "loudly" -- instead, it might deadlock "quietly", have memory
corruptions, etc.
Hence I'm limiting this approach to only the threads created by
QThread::create(), at least for the time being. This also side-steps
perhaps the biggest problem of generalizing the approach, which is that
placing such interrupt+join logic into~QThread's destructor would cause
it to be run _after_ a QThread subclass' own destructor has run,
destroying the subclass' data members too early. This might create
an antipattern if one chooses to subclass QThread. With create(), a
subclass in question exists, and it indeed has NSDMs, but it's entirely
under our control (in fact, I'm placing the logic just in its dtor).
[ChangeLog][QtCore][QThread] Destroying a QThread object created by
QThread::create() while the thread that it manages is still running will
now automatically ask that thread to quit, and will wait until the
thread has finished. Before, this resulted in a program crash. See the
documentation of QThread::~QThread() for more details.
Change-Id: Ib268b13da422e277ee3ed6f6c7b2ecc8cea5750c
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
std::function, which is used to store the type-erased continuation
lambdas, requires the passed callable to be copy-constructible. This
makes impossible to use move-only callables with continuations/handlers.
In particular, it makes impossible passing lambdas that are capturing
move-only objects. The workaround is to store the continuation lambda
inside a wrapper for the callable, which stores the move-only lambda in
a QSharedPtr and can be stored in std::function, since it's copyable.
Pick-to: 6.2
Fixes: QTBUG-98493
Change-Id: I8b7a22fcf68dc132b3c533216a7a1665e9f9fb0a
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
It is not immediately clear from the signature that the time parameter's
unit is miliseconds.
Pick-to: 6.2
Change-Id: Ifee3c6410b6b2352e75571cc53578a660aabb32d
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Also fix an unrelated minor issue in QtFuture::whenAll docs.
Task-number: QTBUG-86714
Change-Id: I45f06b17db0508be10215e6d260ef76ede3077fb
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
[ChangeLog][QtCore] Added QtFuture::whenAll() and QtFuture::whenAny()
functions, returning a QFuture that becomes ready when all or any of the
supplied futures complete.
Task-number: QTBUG-86714
Change-Id: I2bb7dbb4cdc4f79a7a4fd494142df6a0f93a2b39
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This reverts commit 1283ee3245.
We now have wrappers around std::mutex and std::condition_variable
that fall back to QMutex and QWaitCondition on the broken Integrity
toolchain. Use them.
Change-Id: I881aa931167b845b489713048b57ccc5f79d4237
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Mention that the futures returned by continuations will stay
uninitialized, until the corresponding continuation/handler starts
executing and do some general improvments to make the docs more
readable.
Task-number: QTBUG-97582
Pick-to: 6.2
Change-Id: I141ff1630b22ec7a856a457a41a69efec980d44b
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This change allows canceling the chain of continuations attached to a
future through canceling the future itself at any point of execution of
the chain.
[ChangeLog][QtCore][Important Behavior Changes] The chain of
continuations attached to a future now can be cancelled through
cancelling the future itself at any point of the execution of the chain,
as it was documented. Previously canceling the future would cancel the
chain only if it was done before the chain starts executing, otherwise
the cancellation would be ignored. Now the part of the chain that wasn't
started at the moment of cancellation will be canceled.
Task-number: QTBUG-97582
Change-Id: I4c3b3c68e34d3a044243ac9a7a9ed3c38b7cb02e
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
src/corelib/kernel/qmetatype.cpp:1605: (qdoc) warning: Command '\snippet (//! [[implicit]])' failed at end of file 'qmetatyp>
src/corelib/kernel/qmetatype.cpp:1615: (qdoc) warning: Command '\snippet (//! [[member]])' failed at end of file 'qmetatype/>
src/corelib/kernel/qmetatype.cpp:1626: (qdoc) warning: Command '\snippet (//! [[memberOk]])' failed at end of file 'qmetatyp>
src/corelib/kernel/qmetatype.cpp:1639: (qdoc) warning: Command '\snippet (//! [[unaryfunc]])' failed at end of file 'qmetaty>
src/corelib/text/qbytearraymatcher.cpp:233: (qdoc) warning: No such parameter 'view' in QByteArrayMatcher::indexIn()
src/corelib/time/qdatetime.cpp:1854: (qdoc) warning: Can't link to 'QLocaleie:pmText()'
src/corelib/thread/qsemaphore.cpp:494: (qdoc) warning: Undocumented return value (hint: use 'return' or 'returns' in the text
src/corelib/thread/qsemaphore.cpp:505: (qdoc) warning: Undocumented parameter 'timeout' in QSemaphore::try_acquire_for()
src/corelib/thread/qsemaphore.cpp:505: (qdoc) warning: Undocumented return value (hint: use 'return' or 'returns' in the text
src/corelib/thread/qsemaphore.cpp:516: (qdoc) warning: Undocumented parameter 'tp' in QSemaphore::try_acquire_until()
src/corelib/thread/qsemaphore.cpp:516: (qdoc) warning: Undocumented return value (hint: use 'return' or 'returns' in the text
Change-Id: Ib612c69525ec7542f2ad3dd9a07e89f266718fd8
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
Unify cancel and finish in QPromise destructor in a single call. This
saves us one extra mutex lock and atomic state change.
Task-number: QTBUG-84977
Change-Id: Iac06302c39a2863008b27325fcf6792d4f58c8ae
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Unfortunately we can't depend on the C++11, 14 or 17 Standard
Library, which is not properly and completely implemented
everywhere. All Library features above C++98 must be checked with
their corresponding __cpp_lib macro before use. This does not
apply to the C++17 Core Language.
qwaitcondition_p.h:144:20: error: 'mutex' in namespace 'std' does not name a type
144 | using mutex = std::mutex;
| ^~~~~
qwaitcondition_p.h:59:1: note: 'std::mutex' is defined in header '<mutex>'; did you forget to '#include <mutex>'?
58 | #include <condition_variable>
+++ |+#include <mutex>
59 | #include <mutex>
qwaitcondition_p.h:145:33: error: 'condition_variable' in namespace 'std' does not name a type
145 | using condition_variable = std::condition_variable;
| ^~~~~~~~~~~~~~~~~~
qwaitcondition_p.h:59:1: note: 'std::condition_variable' is defined in header '<condition_variable>'; did you forget to '#include <condition_variable>'?
58 | #include <condition_variable>
+++ |+#include <condition_variable>
59 | #include <mutex>
Pick-to: 6.2
Change-Id: I2bbf422288924c198645fffd16a9249e5330136a
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Add startOnReservedThread that specifically releases a reserved thread
and uses it atomically for a given task. This can make a positive
number of reserved threads work.
Change-Id: I4bd1dced24bb46fcb365f12cbc9c7905dc66cdf1
Reviewed-by: David Faure <david.faure@kdab.com>
This patch removes most of the checks that are made using C++20
__cpp_* macros for features available in C++17 and earlier.
Library feature check macros (__cpp_lib_*) are unaffected.
Change-Id: I557b2bd0d4ff09b13837555e9880eb28e0355f64
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Don't bother overwaiting in waitForDone(), if it was done at one point
after it was called we can return true. And do not stop threads recently
awakened by a startThread call as they have tasks to do.
Make allowing at least one thread regardless of reservation more
standard instead of hacked in certain places.
Pick-to: 6.2
Change-Id: I304bcdc5822f440d5e72fc33ba2aa1678c9ba0d0
Reviewed-by: David Faure <david.faure@kdab.com>
Unix QThread implementation stores pthread_t as a system thread ID
when the thread is created, but never resets the system ID when
those threads are destroyed. Some implementations may reuse the
same thread IDs for new threads, and this may cause QThread::wait()
to erroneously complain that "Thread tried to wait on itself".
This patch sets the system thread ID to nullptr when the thread is
about to exit and be destroyed by the system.
A regression test is added to tst_qthread.
Fixes: QTBUG-96846
Pick-to: 5.15 6.2
Change-Id: I0850425dd0e09af50e59c9038e7e662a2a624beb
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The thread ID is already initialized inside QThread::start() while
the thread lock is taken. This is completed before the attempted
initialization in QThreadPrivate::start() because it tries to take
the same lock.
Task-number: QTBUG-96846
Pick-to: 5.15 6.2
Change-Id: Ic9588f3e2e2f3c2180afbed8ec01155b33043eb3
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The thread build variant is not implemented; the no-thread one simply
returns the current thread. As this has no utility at best and
misleading at worst, remove those functions.
Change-Id: I93ee45ae7d55e3a7ad1f03168b14149a5d4c9253
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The idea is to prevent silly mistakes such as
QMutexLocker(mutex);
doSomething();
where the locker is constructed and destroyed immediately. Compilers
don't normally warn in these cases (as the constructor/destructor
pairs involved do have side effects), but we can mark the type as
[[nodiscard]] to encourage warnings.
There is another couple of classes for which this would make sense
(notably, the R/W lockers), but unfortunately those are exported
classes, and GCC has a bug where one can't mix two different attribute
syntaxes on the same entity [1], so I'm skipping those.
[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102399
Change-Id: I75a2443dc71e6b80613b8edd52a04d3379355728
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Ensure that expired threads have actually finished before attempting
to restart them. Calling start() on a thread that is not yet finished
does nothing.
Add a regression test into tst_qthreadpool that attempts to trigger
reuse of expired threads and verifies that all submitted tasks
execute.
Fixes: QTBUG-72872
Pick-to: 6.2
Change-Id: I2109b628b8a4e91491115dc56aebf3eb249646b5
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
std::chrono values are passed by const reference.
Still warnings from undocumented parameters, but rephrasing
the documentation doesn't make it better in this case, so perhaps
qdoc needs a way to suppress the warning. Adding an \omit block
where the parameters or return values are mentioned doesn't help.
Change-Id: I7d495d73d8367d9d90dd33a4880ac7c978382d19
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
riscv32 fails to build because __NR_futex is not defined on this
architecture:
In file included from thread/qmutex_linux.cpp:45,
from thread/qmutex.cpp:804:
thread/qfutex_p.h: In function 'int QtLinuxFutex::_q_futex(int*, int, int, quintptr, int*, int)':
thread/qfutex_p.h:116:30: error: '__NR_futex' was not declared in this scope; did you mean '_q_futex'?
116 | int result = syscall(__NR_futex, addr, op | FUTEX_PRIVATE_FLAG, val, val2, addr2, val3);
| ^~~~~~~~~~
| _q_futex
Pick-to: 6.1 6.2
Fixes: QTBUG-96067
Change-Id: Ib6a9bcc496f37e69ac39362cb0a021fccaf311f5
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Use the new event dispatcher for all non-GUI threads,
nn practice for the main thread when using QCoreApplication,
and when calling QThread::exec().
Change-Id: I9184d52532e06da7e6a87ee27c7d53e0d15e693a
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Replace phrases like "future has been running", "parent" with more
precise descriptions.
Pick-to: 6.1 6.2
Fixes: QTBUG-95273
Change-Id: Ibd5a464007d41cc437da49ba250b9ea0a46078c6
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
We will not gain anything if we have to do multiple function calls to
obtain the thread id. Therefore we introduce a macro to signal that we
have a fast implementation of currentThreadId, and only use the function
if it is defined.
Pick-to: 6.2
Change-Id: I3347489ea91992896bb753b796ae26e391c2c99c
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
When inserting items into the result store, a ResultItem is created,
which stores a pointer to the results list and their size. If the size
of the ResultItem is set to 0, it means that a single result is stored.
In case of trying to report results via an empty list, the size is 0, so
result store treats it as a single result.
Added checks before storing the results to make sure that the result
list isn't empty. Note that empty lists are allowed in some cases for
the filter mode, because ResultStoreBase::addResults() knows how to
handle those cases correctly.
Task-number: QTBUG-80957
Pick-to: 5.15 6.1 6.2
Change-Id: I399af4c3eef6adf82fea5df031fe9a9075006b1f
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
All the getters of QFutureWatcher are consistent with the getters of the
corresponding QFuture, except for the isFinished() method, which returns
'true' only after the finished() signal is delivered. This behavior
might be unintuitive for the users. In particular, isFinished() returns
'false', even if it's called immediately after waitForFinished().
[ChangeLog][QtCore][QFutureWatcher][Important Behavior Changes] The
QFutureWatcher::isFinished() method now indicates if the related
QFuture is finished, instead of indicating if the finished() signal was
delivered. This makes it consistent with the future that is being
watched.
Fixes: QTBUG-91048
Pick-to: 6.1 6.2
Change-Id: I6ae9b882b23e06198a82c95b026491bd480b3bf0
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
QMetaType now implicitly knows how to convert any QFuture<T> to
QFuture<void> without needing to manually register a converter
function.
QtWebChannel will make use of this to transparently support QFuture<T>
return types.
[ChangeLog][QtCore][QMetaType] QMetaType now supports converting any
QFuture<T> to QFuture<void>.
Task-number: QTBUG-92903
Change-Id: Ied7e71be37c346cc3d2c274ffb0d91a6821ab4d4
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
For generic interop with other parts of Qt, we need to be able to
extract the type-erased internal storage of QFuture<T>. In particular,
QtWebChannel needs this to transparently support QFuture<T> as a method
return type.
Task-number: QTBUG-92903
Change-Id: I763f054656b8810d58720262d364baf42c47eb37
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Revert a part of d3638e18d3 because the
change was not needed
Task-number: QTBUG-94407
Pick-to: 6.2
Change-Id: I7cf0969d6ecc4078d38b4241ffe7d24273a6d412
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
... and C++20 std::counting_semaphore API compatibility.
[ChangeLog][QtCore][QSemaphore] tryAcquire() now optionally takes a
<chrono> duration as timeout, not just int milliseconds.
[ChangeLog][QtCore][QSemaphore] Added try_acquire{,_for,_until}() for
C++20 std::counting_semaphore compatibility.
Change-Id: I34b6b4bf57a54745d4b97349903d090c4995338a
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
QMutexLocker is declared in qmutex.h, which is already included.
Pick-to: 6.2
Task-number: QTBUG-94407
Change-Id: I78e0630c27ef7f85feae68e98f8cdcbe4e142cd8
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Copying a QOrderedMutexLocker is questionable, and would currenly easily
lead to UB. Therefore we delete the copy ctor and copy assignment
operator, and implement well-behaving move operators.
In addition, provide an explicit dismiss method for cases where we don't
want the locker to unlock the mutexes, as they have been manually
unlocked (this could have been implemented previoulsy by using the copy
assignment operator).
Pick-to: 6.2 6.1 5.15
Change-Id: If2a888710e1c74277b28fd3e2939ab26fff0c7ae
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
qtbase\src\corelib\thread\qfutureinterface_p.h(180): error C2039: 'setException': is not a member of 'QtPrivate::ExceptionStore'
qtbase\src\corelib\thread\qexception.h(109): note: see declaration of 'QtPrivate::ExceptionStore'
Change-Id: Id746b87af3f88d91c473f6a3d986dc0f022dfa2f
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Some of the data members related to progress reporting (min, max and
text) aren't used when user doesn't want manual progress reporting, so
the data for them can be allocated on demand, when the user explicitly
sets them. Note, that we still need to always create other related data
(current value and progress timer), since in the non-manual mode
progress is still reported by incrementing the current value each time
a new result is reported.
Task-number: QTBUG-92045
Change-Id: I1e5bd17de2613a6ea72ccff0029812f67686708b
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
QFuture doesn't need both at the same time, calling QFuture::result(s)
either returns a result or throws an exception. Store result and
exception stores in a union, to reduce the memory.
Also added a note for making the ResultStoreBase destructor non-virtual
in Qt 7.
Task-number: QTBUG-92045
Change-Id: I7f0ac03804d19cc67c1a1466c7a1365219768a14
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Add a noexcept from a doc signature that was missing it.
Change-Id: I30b429a6ec4d182551ca68767d1c8aa61c9e4c09
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
refT()/derefT() can be marked noexcept; we don't care about not
overflowing the refcounter as a precondition. (This is BC, as no
compiler mangles noexcept.) This in turn allows to mark a constructor
calling refT() as noexcept.
Driveby: mark also the same functions to not be `const` in Qt 7.
They clearly are meant to modify *this, and constness only works
because of the unmanaged (raw) d-pointer.
Change-Id: I8d7d365640ff2e1cedc0a234c9abccdfc95ba6e3
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
std::exception_ptr is a reference-counted "smart pointer", so we
shouldn't copy it around freely. Unfortunately QFutureInterface
has exported functions taking it by value, so we can't just change
the signatures and keep BC. Simply prepare the code for Qt 7.
Change-Id: Ic5aae6a095c8c842872a40db440c99d2dfe371f1
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
QFuture shouldn't be passed by value: it's a reference counted
class ("shared future").
Change-Id: I3890062c45f9a6470b278223b88d482f57bf68a0
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Get rid of a reduntant initialization; a data member gets default
initialized anyways.
Change-Id: I85ad4c9a00b3066ccd993dfc2652a3f0164a860e
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
In random order:
* QFutureInterfaceBase has a d-pointer, and copy semantics. So,
"naturally" extend it to support move semantics as well. (These get
used by QPromise/QFuture, as they both hold a QFutureInterface*
as a data member). The only addition needed is a check for a null
d-pointer in the destructor.
Drive by, reorganize the code for the copies, and use copy-and-swap
instead of the hand-rolled solution. Also, add a free swap()
overload, and mark the existing one as candidate for inlining
in Qt 7 (doesn't need to be an exported function).
* QFutureInterface inherits QFutureInterfaceBase, again with value
semantics. To be honest, I'm not sure why QFutureInterfaceBase is
polymorphic -- could be a design mistake, as polymorphic classes
don't mix with value semantics. Anyways, reorganize the code for
copies, apply copy-and-swap, and add move semantics. This requires
adding a check into derefT().
* Finally, QPromise was already move-only, but had broken move
semantics: the move constructor was not noexcept (!) and it actually
allocated memory (!!!). Fix that one (can be defaulted now), and
streamline the move assignment via the proper macro.
Drive by, fix the signature of the constructor from QFutureInterface
(take const-ref, not plain ref -- it's eventually copied, so it
can keep the const), and add another internal constructor from
rvalue QFutureInterface that moves from it.
Change-Id: I9d61a9dd4d45f34942d8f34416baa118c0307390
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
The check is over-arching. is_move_constructible is sufficient;
we don't have to support "ridiculous" types that are copiable but
have deleted move operations (such types are fundamentally broken).
This is in line with the Move* (legacy) named requirements.
Change-Id: Idc7116b39013501b9be39628a4e7afd35fe15530
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
Either the mix of futex and atomic, or the mix of 32-bit futex and
64-bit atomic doesn't work. In any case, the existing code leads to
bad behavior.
Pick-to: 6.1 5.15
Fixes: QTBUG-92188
Change-Id: Icc6ba28d6e2465c373d00e84f4da2b92c037e797
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The documentation for QCoreApplication::exit() even refers to it as
such. Instead of refactoring the documentation to explain that it isn't,
just make it a slot and do the same for the other classes.
Complements 2e6c37fe51.
[ChangeLog][QtCore][QEventLoop] exit() is now a slot, like quit().
[ChangeLog][QtCore][QThread] exit() is now a slot, like quit().
Change-Id: Ic42004c9bf71440eb433fffd167f4a1b89bcac80
Reviewed-by: David Faure <david.faure@kdab.com>
Giuseppe D'Angelo
Marc Mutz
Thiago Macieira
Sona Kurazyan
Volker Hilsheimer
Fabian Kosmale
Mårten Nordheim
Andreas Hartmetz
Allan Sandfeld Jensen
Marc Mutz
Alexey Edelev
Venugopal Shivashankar
Jonas Kvinge
Alexander Neumann
Ievgenii Meshcheriakov
Fabrice Fontaine
Morten Johan Sørvig
Nico Vertriest
Arno Rehn
Tasuku Suzuki
Nodir Temirkhodjaev
Paul Wicking