Semi-automated, just needed ~20 manual fixes:
$ find \( -iname \*.cpp -or -iname \*.h \) -exec perl -pe 's/(\.|->)load\(\)/$1loadRelaxed\(\)/g' -i \{\} +
$ find \( -iname \*.cpp -or -iname \*.h \) -exec perl -pe 's/(\.|->)store\(/$1storeRelaxed\(/g' -i \{\} +
It can be easily improved (e.g. for store check that there are no commas
after the opening parens). The most common offender is QLibrary::load,
and some code using std::atomic directly.
Change-Id: I07c38a3c8ed32c924ef4999e85c7e45cf48f0f6c
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Current tip-of-tree clang (after Clang 8 was branched) added an
intrinsic function __builtin_ia32_xgetbv, and added the following
define that provides _xgetbv:
#define _xgetbv(A) __builtin_ia32_xgetbv((long long)(A))
This fallback declaration of the _xgetbv function only is used
in case the Q_OS_WIN branch of the #if/#elif below is used, if
the #if (defined(Q_CC_GNU) && !defined(Q_CC_EMSCRIPTEN)) ||
defined(Q_CC_GHS) wasn't taken. I left out the
!defined(Q_CC_EMSCRIPTEN) part as I believe Q_OS_WIN and
Q_CC_EMSCRIPTEN are mutually exclusive.
Change-Id: I257fc4283ff9f0845df51ab764cf58acdf285c66
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This is the squashed diff from wip/webassembly to dev.
Done-with: Peng Wu <peng.wu@intopalo.com>
Done-with: Sami Enne <sami.enne@intopalo.com>
Done-with: Morten Johan Sørvig <morten.sorvig@qt.io>
Started-by: Andrew Knight <andrew.knight@intopalo.com>
Change-Id: I6562433c0a38d6ec49ab675e0f104f2665f3392d
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
... instead of being the bit numbers. This allows us to test more than
one feature at a time with qCpuHasFeature (see commit about the Haswell
architecture features). The drawback is that we won't be able to handle
more than 63 different CPU features, though we're likely quite far from
it (x86 currently has only 36 features).
Change-Id: Iff4151c519c144d580c4fffd153a0acbfd74c2c6
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
This reverts commit 6e1190053d.
We had to bring back qatomic_msvc.h for MSVC 2015, which does
not provide 64-bit atomics on 32-bit platforms.
Task-number: QTBUG-68719
Change-Id: Iea902cf0b01191717e0a640944771fcede7b5ff8
Reviewed-by: Liang Qi <liang.qi@qt.io>
This adds detection for: VAES, GFNI, AVX512VBMI2, AVX512VNNI,
AVX512BITALG, AVX512VPOPCNTDQ, AVX512_4NNIW, AVX512_4FMAPS. These
features were found in the "Intel® Architecture Instruction Set
Extensions and Future Features" manual, revision 30. This commit also
adds support for RDPID (already in the main manual) and the Control-flow
Enforcement Technology, which appears in a separate Intel paper.
This new support was done by adding a new generator script so we don't
have to maintain two tables in sync, one in qsimd.cpp with the feature
names, and the other in qsimd_p.h.
Since we now need a lot more bits, it's no longer worth keeping the two
halves of the qt_cpu_features variable mostly similar to the main two
CPUID results. This commit goes back to keeping things in order, like we
used to prior to commit 6a8251a89b (Qt 5.6)
At the time of this commit, GCC 8 has macros for AVX512VPOPCNTDQ,
AVX512_4NNIW, AVX512_4FMAPS, AVX512VBMI2 and GFNI.
Change-Id: I938b024e38bf4aac9154fffd14f7afae50faaa96
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
We had that as a concession for early toolchains that failed to include
the necessary libraries. They must now be up-to-date.
Change-Id: I938b024e38bf4aac9154fffd14f7a630ef160cd5
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
This merge also blacklists a flaky tst_QGL::clipTest test on
OpenSUSE 13.1.
Conflicts:
src/network/socket/qnativesocketengine_winrt.cpp
tests/auto/opengl/qgl/BLACKLIST
Task-number: QTBUG-53133
Change-Id: I14b431aa5a189b7dd1d3e2dfff767d15df20fde3
On OpenBSD, ffsll needs to be defined to compile. This is the
same change as in commit 725a9c2702
for NetBSD.
Change-Id: I3060caa10950a9419084a12de8c88a2f98b34d07
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Move the Q_ALWAYS_INLINE and forcing of __builtin_memcpy to the existing
functions.
Change-Id: Icaa7fb2a490246bda156ffff143c137e520eea79
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
On NetBSD, ffsll needs to be defined to compile. The fix adds the
according Q_OS_NETBSD define to the list of operating systems
needing the define.
Fix obtained from NetBSD port maintainer Kamil Rytarowski <n54@gmx.com>
via IRC.
Change-Id: I966a7b3fba43fb56e72f19f6b7f7cacc19c3c6cf
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
15b42af111 changed the qt_cpu_features
variable to be an array and never fixed this #if branch of the code.
Change-Id: Id69569111e7d4e619e22ffff144cf930f86f478e
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
This change partially reverts 1bfc7f68 about QT_HAS_BUILTIN define
and undef in src/corelib/tools/qsimd_p.h.
This change is also squashed with "Fall back to c++11 standard
compiler flag for host builds" which is done by Peter Seiderer.
Conflicts:
mkspecs/features/default_post.prf
src/3rdparty/sqlite/0001-Fixing-the-SQLite3-build-for-WEC2013-again.patch
src/3rdparty/sqlite/sqlite3.c
src/corelib/tools/qsimd_p.h
src/gui/kernel/qevent.cpp
src/gui/kernel/qwindowsysteminterface.cpp
src/gui/kernel/qwindowsysteminterface_p.h
src/plugins/bearer/blackberry/blackberry.pro
src/plugins/platforms/cocoa/qcocoasystemsettings.mm
src/plugins/platformthemes/gtk2/gtk2.pro
src/plugins/styles/bb10style/bb10style.pro
src/sql/drivers/sqlite2/qsql_sqlite2.cpp
tools/configure/configureapp.cpp
Task-number: QTBUG-51644
Done-with: Peter Seiderer <ps.report@gmx.net>
Change-Id: I6100d6ace31b2e8d41a95f0b5d5ebf8f1fd88b44
Found by UBSan:
src/corelib/tools/qstring.cpp:587:42: runtime error: load of misaligned address 0x2acbf4b7551b for type 'const long long int', which requires 8 byte alignment
src/corelib/json/qjson_p.h:405:30: runtime error: store to misaligned address 0x0000019b1e52 for type 'quint64', which requires 8 byte alignment
src/corelib/tools/qhash.cpp:116:27: runtime error: load of misaligned address 0x2b8f9ce80e85 for type 'const qlonglong', which requires 8 byte alignment
src/corelib/tools/qhash.cpp:133:26: runtime error: load of misaligned address 0x2b8f9ce80e8d for type 'const ushort', which requires 2 byte alignment
Fix by memcpy()ing into a local variable. Wrap this trick in
template functions in qsimd_p.h. These are marked as always-
inline and use __builtin_memcpy() where available in an
attempt to avoid the memcpy() function call overhead in debug
builds.
While this looks prohibitively expensive, from the pov of the
C++ abstract machine, it is 100% equivalent, except for the
absence of undefined behavior. In one case, the cast produces
a local temporary which is then copied into the function, and
in the other case, that local variable comes from return value
of qUnalignedLoad().
Consequently, GCC compiles these two versions into identical
assembler code (only verfied for ucstrncmp, but there's no
reason to believe that it wouldn't hold for the other cases,
too).
Task-number: QTBUG-51651
Change-Id: Ia50d4a1d7580b6f803e0895c9f3d89c7da37840c
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
iOS cannot do runtime feature detection, and querying the CPU is only
allowed in kernel mode (or beyond), so we have to decide the features
at compile time, in which case we might as well use the fallback code
path that uses the built in __ARM_* macros to point out which features
are supported, instead of hard-coding the features for iOS.
Change-Id: Ie507c0d8e962a7bdab16508c8b8122645276512e
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Change-Id: Ifcad547caf2d2a7ad7aa1ccb4fbed08810905cee
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Erik Verbruggen <erik.verbruggen@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
From Qt 5.7 -> tools & applications are lisenced under GPL v3 with some
exceptions, see
http://blog.qt.io/blog/2016/01/13/new-agreement-with-the-kde-free-qt-foundation/
Updated license headers to use new GPL-EXCEPT header instead of LGPL21 one
(in those files which will be under GPL 3 with exceptions)
Change-Id: I42a473ddc97101492a60b9287d90979d9eb35ae1
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
From Qt 5.7 -> LGPL v2.1 isn't an option anymore, see
http://blog.qt.io/blog/2016/01/13/new-agreement-with-the-kde-free-qt-foundation/
Updated license headers to use new LGPL header instead of LGPL21 one
(in those files which will be under LGPL v3)
Change-Id: I046ec3e47b1876cd7b4b0353a576b352e3a946d9
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
cpuid and cpuidex are available as intrinsics, but the GNU-style
assembly is needed for xgetbv.
Change-Id: Ib9f280ac6b69b7ffb9c39289b52fa4af5e2de9ba
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
If, after checking a condition, we issue a qFatal()
or a qCritical(), by definition that check is
unlikely to be true.
Tell the compiler so it can move the error handling
code out of the normal code path to increase the
effective icache size.
Moved conditional code around where possible so that
we could always use Q_UNLIKELY, instead of having to
revert to Q_LIKELY here and there.
In some cases, simplified the expressions newly wrapped
in Q_UNLIKELY as a drive-by.
Change-Id: I67537d62b04bc6977d69254690c5ebbdf98bfd6d
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The ffsll function is missing with that SDK, but Clang and GCC support
it as the builtin.
Change-Id: I9f4d900898814b569c1affff140e15fb59fb45d8
Reviewed-by: Samuel Gaist <samuel.gaist@edeltech.ch>
Reviewed-by: Jake Petroules <jake.petroules@theqtcompany.com>
The other left shifts in this file already have the Q_UINT64_C wrapper.
This one was missed.
Change-Id: I42e7ef1a481840699a8dffff140d758ac370c402
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Detection for most of them is free because we're loading the entire
registers anyway. The only exception is AVX512VBMI, which is in a new
register we hadn't yet read from.
I've also added the new GCC names so they can be used with
QT_FUNCTION_TARGET. The only two exceptions are "movbe" and "popcnt",
which are extremely restricted in use and we are not likely to have code
dedicated to using them.
Change-Id: Ib306f8f647014b399b87ffff13f1d8fd29e58be0
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Instead of trying to detect one bit and set another, let's just use the
bits from the x86 CPUID instruction on x86. This makes use of the full
64-bit space now.
Since MSVC doesn't like enums bigger than 32-bit, we have to store the
bit number instead of the actual bit value in the constant. For that
reason, I also renamed the constants, to catch anyone who was using them
directly, instead of through qCpuHasFeature.
Change-Id: Ib306f8f647014b399b87ffff13f1d587692d827a
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
I'm going to need the extra bits for x86.
Change-Id: Ib306f8f647014b399b87ffff13f1d3d23e138518
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
It does. When CPUID was introduced in late 486, it already supported
leaf 1. That means all Pentium-class systems (including Intel's new
Quark systems) have CPUID and they have at least leaf 1.
Change-Id: Ib306f8f647014b399b87ffff13f1d5d2530d9ddd
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
There's no __HLE__ macro and there won't be, since the HLE prefix can be
run on older CPUs. There's no need for runtime detection.
Change-Id: Ib306f8f647014b399b87ffff13f1daba0e654b02
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
If you're running Qt under valgrind, it stands to reason that you know
what you're doing.
In particular, Valgrind does support AVX and AVX2 instructions, but some
versions seem to be missing the necessary CPUID bits.
Change-Id: I9a75ad8521ae4e5cbbe5ffff13d1940f6fa5c4f4
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: David Faure <david.faure@kdab.com>
Bionic on x86 does not have an implementation of ffs(),
for some reason. The declaration is available in
strings.h, but causes a linker error. The toolchain has
a builtin version though, which can be used instead.
Change-Id: Iaf4c27542c1aabe88713842303251db44ae6ba86
Reviewed-by: Jani Heikkinen <jani.heikkinen@theqtcompany.com>
Reviewed-by: Christian Stromme <christian.stromme@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Qt copyrights are now in The Qt Company, so we could update the source
code headers accordingly. In the same go we should also fix the links to
point to qt.io.
Outdated header.LGPL removed (use header.LGPL21 instead)
Old header.LGPL3 renamed to header.LGPL3-COMM to match actual licensing
combination. New header.LGPL-COMM taken in the use file which were
using old header.LGPL3 (src/plugins/platforms/android/extract.cpp)
Added new header.LGPL3 containing Commercial + LGPLv3 + GPLv2 license
combination
Change-Id: I6f49b819a8a20cc4f88b794a8f6726d975e8ffbe
Reviewed-by: Matti Paaso <matti.paaso@theqtcompany.com>
The sequence of (__GNUC__ * 100 + __GNUC_MINOR__) was used in quite a
few places. Simplify it to make the code more readable.
This follows the change done for Clang, which was quite necessary since
Apple's version of Clang has different build numbers.
Change-Id: I886271a5a5f21ae59485ecf8d140527723345a46
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
This code hasn't been tested for at least 4 years. It's not maintained
and probably doesn't work.
Change-Id: I4b9a5179e34111b400914f91caa6b741b69771bb
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
The #undef in qcompilerdetection.h was missing.
And apparently we can detect Neon since Windows Mobile 6 too.
Change-Id: I38a5f71b2704a29a706183e39f43db3a78a729db
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Giuseppe D'Angelo
Liang Qi
Jesus Fernandez
Martin Storsjö
Thiago Macieira
Morten Johan Sørvig
Ralf Nolden
Friedemann Kleint
Liang Qi
Marc Mutz
Erik Verbruggen
Tor Arne Vestbø
Jani Heikkinen
Tobias Koenig
Rolland Dudemaine
Frederik Gladhorn
Bjoern Breitmeyer
Eskil Abrahamsen Blomfeldt
Matti Paaso