"AVX512MIC" (Many Integrated Cores) is the set of AVX-512 features found
on the Intel Xeon Phi coprocessors (codename "Knights Landing"), which
is an unlikely architecture for Qt to run on.
The two profiles with VL came from study of early GCC code and are no
longer applicable. GCC source code now shows both VBMI and IFMA as part
of the -march=cannonlake feature set.
Change-Id: Iff4151c519c144d580c4fffd153a0f268919fe2c
Reviewed-by: Lars Knoll <lars.knoll@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>
Since the x86_simd/main.cpp file already has all the source for each and
every test anyway, just reuse it.
Change-Id: I938b024e38bf4aac9154fffd14f779f450827fb9
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
This has two main benefits:
1) introduces a qmake CONFIG we can use in .pro/.pri/.prf files
2) removes the need to keep an up-to-date list of which compilers
support the feature
The test is implemented as trying to compile every single SIMD test we
currently have, but without passing the -mXXX option. The reason for
trying all of them is that some people may have modified their mkspecs
to add -mXXX options or -march=XXX, which could enable the particular
feature we tried, resulting in a false positive outcome.
Change-Id: I938b024e38bf4aac9154fffd14f7784dc8d1f020
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
It seems the compiler supports /arch:AVX512 and /arch:AVX512F but none
of the other switches (and neither are documented). And when you pass
those, you also get Conflict Detection (CD), Double & Quad (DQ), Byte &
Word (BW) and Vector Length (VL), which matches the ICC switch
"-xCORE-AVX512". Unlike ICC, there doesn't seem to be an option to
enable only the common part of AVX-512.
Support for Intel Xeon Phi's current features (Exponential &
Reciprocation and Prefetch) and future ones (IFMA, VBMI, 4FMAPS, 4VNNI
and VPOPCNTDQ) seems to be missing altogether.
See https://blogs.msdn.microsoft.com/vcblog/2017/07/11/microsoft-visual-studio-2017-supports-intel-avx-512/
Change-Id: I98105cd9616b8097957db680d73eb1f86e487e6d
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Convert QSysInfo/QOperatingSystemVersion to __builtin_available where
required or possible, or to QOperatingSystemVersion where
__builtin_available cannot be used and is not needed (such as negated
conditions, which are not supported by that construct).
Change-Id: I83c0e7e777605b99ff4d24598bfcccf22126fdda
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The use as in the code:
futimesat(fd, NULL, &tv)
is not documented to work. The file descriptor should be a directory's
one, not an open file (though the Linux source code seems to handle that
case). This call was done as a fallback to futimes, so it's very
unlikely a system would have futimesat and not futimes.
Both the Linux and the FreeBSD man pages say it's deprecated anyway.
Change-Id: I8d96dea9955d4c749b99fffd14cd94068dc7668a
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Availability of D3D11_QUERY_DATA_TIMESTAMP_DISJOINT depends on the used
MinGW version so that the check for MINGW is not sufficient. The newly
added configure test can be used for every toolset.
Task-number: QTBUG-57916
Change-Id: Ia9cb48f3e673841101a93cbc8ea23aff9547f639
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
only few tests remain, and many of these were mis-classified anyway.
Change-Id: Ic3bc96928a0c79fe77b9ec10e6508d4822f18df2
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
We're adding a lot of unnecessary files that end up later as cargo-cult,
for at most a handful of lines. So instead move the testcases directly
into the .json file.
The following sources were not inlined, because multiple tests share
them, and the inlining infra does not support that (yet):
- avx512
- openssl
- gnu-libiconv/sun-libiconv (there is also a command line option to
select the exact variant, which makes it hard/impossible to properly
coalesce the library sources)
The following sources were not inlined because of "complications":
- verifyspec contains a lengthy function in the project file
- stl contains lots of code in the source file
- xlocalescanprint includes a private header from the source tree via a
relative path, which we can't do, as the test's physical location is
variable.
- corewlan uses objective c++, which the inline system doesn't support
reduce_relocs and reduce_exports now create libraries with main(), which
is weird enough, but doesn't hurt.
Done-with: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Change-Id: Ic3a088f9f08a4fd7ae91fffd14ce8a262021cca0
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
testing the non-xkb xcb parts is unnecessary, as we already did that
separately.
Change-Id: I452cc746315117a0169f0e0c764fe7e0229437e9
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
just use 'which' and be done with it. the script was rather arcane, and
worked around deficiencies of cygwin (no longer relevant) and solaris
(assumed to be somewhat sane meanwhile).
Change-Id: I2e11ea3c87ac06a85604ac8d58d8fee95eae2e15
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
remove useless define and ifdef/error, and check for a more appropriate
function.
Change-Id: I1a1622cc157c49ebb6787068ade7b33e45e228ca
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
these would be caught by the central verifyspec test, if we even got
that far.
Change-Id: I3eda80c4614b94f869d907f0a40166f4914ea692
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
- #error hack for MIPSpro compiler from unix/clock-*
- irix exclusions from unix/*iconv
- hpux defines from unix/{getifaddrs,ipv6ifname} (the obsolete
hpuxi-g++-64 spec would add them automatically anyway)
Change-Id: Ib227e5626c0e8c8f6faebf76c312d77955f80b92
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
amends b4525b3407, which gives no indication for why this was done.
judging by other tests, it wasn't cargo-culted. i presume it was just
about disabling debug-and-release, and the debug choice was arbitrary.
the central system nowadays does the same, just that it uses release.
amends 1533bfc5fc, which certainly _was_ cargo-culted from the former.
as a side effect, this removes some other CONFIG manipulations which are
handled centrally or are wholly ineffective nowadays.
Change-Id: Ib9af2837925a2f19af05506e798a26d363635735
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
in the gles tests, use #ifdef __APPLE__ directly instead of defining
BUILD_ON_MAC. this is consistent with the desktop gl test, less magic,
and corresponds with the later usage (which does #ifdef Q_OS_MAC).
amends f3d82a89.
in the desktop gl test, clean out the dead Q_OS_MAC define from the
project file (added in 45dc5852 concurrently to the more direct fix in
864815ef).
Change-Id: Ieebad3c7e87218b887f61485c331b93eadfcf406
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
It's a Linux-specific call that was added to the kernel in pre
historical times (before Git). Sqlite3 uses mremap(2) but it has its own
checking. Nothing else in Qt uses this.
Looks like the last user was the QPF font engine, removed in commit
d7e424ee66 almost four years ago. And
that's considering that the QPF font engine wasn't in use since Qt 5.0
because QWS was no more...
Change-Id: Idaa189413f404cffb1eafffd14ceee7488514c1d
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
This patch-set implements a new QSslSocket backend based on OpenSSL 1.1.
1. General.
The code in this patch was organized to achieve these (somewhat contradicting)
objectives:
- keep the new code free of #if-ery, as far as possible;
- make it easy to clean away dead code when we're eventually able to retire
out-dated OpenSSL versions;
- reduce the amount of code duplication.
If changes in some file/component were insignificant (~5 one-liners per file),
we still use pp-checks like: #if QT_CONFIG(opensslv11) ... #else ... #endif -
the logic is simple and it's still easy to clean the code if we remove the legacy
back-end. Where it saved #if-ery, we also introduced 'forward-compatible'
macros implementing equivalents of 1.1 functions using older OpenSSL.
In case some class contains a lot of version-specific ifdefs (particularly where
nested #if-ery was complex) we choose to split code into: "pre11" h/cpp files,
"shared" h/cpp files (they preserve their original names, e.g qsslsocket_openssl.cpp)
and "11" h/cpp files. If in future we remove the legacy back-end, "pre11" should be
removed; "shared" and "11" parts - merged.
2. Configuration.
We introduced a new feature 'opensslv11' which complements the pre-existing
'openssl' and 'openssl-linked' features. The 'opensslv11' feature is enabled
by a simple test which either compiles successfully or ends in a compilation
error, depending on a value of the OPENSSL_VERSION_NUMBER constant. If the
feature was enabled, we also append an additional compilation flag
-DOPENSSL_API_COMPAT=0x10100000L to make sure our new code does not contain
deprecated structures, function calls, macro-invocations from OpenSSL < 1.1.
Change-Id: I2064efbe9685def5d2bb2233a66f7581954fb74a
Reviewed-by: André Klitzing <aklitzing@gmail.com>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
The getentropy function, first found in OpenBSD, is present in glibc
since version 2.25 and Bionic since Android 6.0 and NDK r11. It uses the
Linux 3.17 getrandom system call. Unlike glibc's getrandom() wrapper,
the glibc implementation of getentropy() function is not a POSIX thread
cancellation point, so we prefer to use that even though we have to
break the reading into 256-byte blocks.
The big advantage is that these functions work even in the absence of a
/dev/urandom device node, in addition to a few cycles shaved off by not
having to open a file descriptor and close it at exit. What's more, the
glibc implementation blocks until entropy is available on early boot, so
we don't have to worry about a failure mode. The Bionic implementation
will fall back by itself to /dev/urandom and, failing that, gathering
entropy from elsewhere in the system in a way it cannot fail either.
uClibc has a wrapper to getrandom(2) but no getentropy(3). MUSL has
neither.
Change-Id: Ia53158e207a94bf49489fffd14c8cee1b968a619
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Follow the usual pattern:
Add a config test and automatic include of GLES3/gl32.h if there
is a GLES 3.2 capable header+lib at build time.
Then, regardless of this being enabled, expose all new 3.2 API
functions in QOpenGLExtraFunctions and resolve them dynamically
at run time.
This way 3.2 functions will be available when deployed to a 3.2
capable system (or OpenGL 3/4.x with the functions in question
available) regardless of what was present in the sysroot at build
time.
Change-Id: Ia52551f3178591e1e56ceac8e45d89c6b13f4927
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
GCC didn't support it until version 5 or 6, so add configure tests for
both <random> and <sys/auxv.h>. Normally I'd say "upgrade", but this is
too low-level and important a feature.
There's a good chance that all our supported compilers have <random>
anyway. As for <sys/auxv.h>, it's present on Glibc, Bionic and MUSL, but
I don't see it in uClibc (AT_RANDOM is a Linux-specific feature).
Change-Id: Ia3e896da908f42939148fffd14c5b2af491f7a77
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
This was originally enabled in the mkspecs for 64-bit QNX 7.0.0
but that broke when the qtConfig change was made. It looks like
qtConfig shouldn't be used in the platform mkspecs. I suspect
the stack-protector changes were left out of the 32-bit mkspecs
so that 6.6.0 builds wouldn't be affected.
Ignore the stack-protector/stack-protector-all possibility since
it isn't possible to access it without a command line option.
Specifying both options doesn't even make sense since
stack-protector-all encompasses stack-protector.
For now, leave out command line control of this feature.
Task-number: QTBUG-59644
Change-Id: I99323216be5b592dd2c3bef6d22da195764a6e65
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
The instruction is "RDRAND", but the feature name, according to GCC, is
RDRND, so I had to change some macros in qsimd_p.h.
Change-Id: Icd0e0d4b27cb4e5eb892fffd14b5166779137e63
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
With this setup clang cannot use c++1z yet.
Fixed with clang 5.0.
In file included from /data/sources/qt/qt5/qtbase/src/corelib/codecs/qtextcodec.cpp:53:
In file included from ../../../include/QtCore/5.9.1/QtCore/private/qcoreglobaldata_p.h:1:
In file included from ../../../include/QtCore/5.9.1/QtCore/private/../../../../../../../../../sources/qt/qt5/qtbase/src/corelib/kernel/qcoreglobaldata_p.h:55:
In file included from ../../../include/QtCore/qmap.h:1:
In file included from ../../../include/QtCore/../../../../../../sources/qt/qt5/qtbase/src/corelib/tools/qmap.h:52:
In file included from /usr/bin/../lib64/gcc/x86_64-pc-linux-gnu/7.1.1/../../../../include/c++/7.1.1/map:60:
In file included from /usr/bin/../lib64/gcc/x86_64-pc-linux-gnu/7.1.1/../../../../include/c++/7.1.1/bits/stl_tree.h:72:
In file included from /usr/bin/../lib64/gcc/x86_64-pc-linux-gnu/7.1.1/../../../../include/c++/7.1.1/bits/node_handle.h:39:
/usr/bin/../lib64/gcc/x86_64-pc-linux-gnu/7.1.1/../../../../include/c++/7.1.1/optional:1032:27: error: use of class template 'optional' requires template arguments
Change-Id: Ib4cd8a9f5791a6e6cae4e6d61dfec3ad50dd63ab
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This change adds USB mouse handling support for INTEGRITY
Change-Id: I8a2a51c8c3578898e90dd5bbb01f6aed6c64e2a4
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Timo Aarnipuro <timo.aarnipuro@qt.io>
Reviewed-by: Nikola Velinov <nvelinov@ghs.com>
Reviewed-by: Rolland Dudemaine <rolland@ghs.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Tero Alamaki <tero.alamaki@qt.io>
We can't depend on QT_HAS_INCLUDE for such an important functionality in
QtQml, so detect at configure time.
alloca() is not a POSIX function (it apparently first appeared in
Version 32V AT&T UNIX), so the actual header that defines it varies from
system to system. Clearly, if alloca.h exists, that's the one, so we try
it first. On most other systems that don't define it, it's in stdlib.h.
The only exception is Windows, where it's actually defined in malloc.h.
Task-number: QTBUG-59700
Started-by: Thiago Macieira <thiago.macieira@intel.com>
Change-Id: Icd0e0d4b27cb4e5eb892fffd14b4b2b389a4684e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
This commit revives the old native QPixmap and QPaintEngine
implementations that were present in Qt4. The backing store supports
regular raster windows in this commit. Support for render-to-texture
widgets and OpenGL compositing will be added in a follow-up commit.
Change-Id: I80a9c4f0c42a6f68f571dfee930d95000d5dd950
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
MSVC, Apple's Clang and Clang prior to 3.9 do not recognize _cvtss_sh
and _cvtsh_ss. So expand the operation to use directly the packed
intrinsics.
Change-Id: I27b55fdf514247549455fffd14b2046fd638593d
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Test failed on QNX 7, even though alloca is available. On QNX7,
it's a macro that expands to a line with NULL, but without define
for it.
alloca.cpp:44:5: error: 'NULL' was not declared in this scope
Task-number: QTBUG-59700
Change-Id: I3631d139990020a3adbab8b72e49929b6e721e80
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The configure-time detection (cxx11default) isn't enough if the compiler
can be changed. This is especially necessary if Qt is compiled with a
compiler that defaults to >= C++11 (e.g., GCC 6) and then the user
selects a compiler another compiler (e.g., Clang) via -spec option. In
that case, we'd miss adding the -std=c++11 or -std=gnu++11 option to the
command-line, causing the compilation to fail.
As a nice side-effect, even moc without moc_predefs.h will now get the
__cplusplus setting.
Task-number: QTBUG-58321
Change-Id: I74966ed02f674a7295f8fffd14a8be35da9640e1
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
A typo caused the test to never detect the system wide PCRE.
Task-number: QTBUG-59226
Change-Id: I42ada99aac240455d11b53d2ab59d712d8f811ff
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This is needed to encode the correct ABI into the generated qml caches.
It is identical with the return value of QSysInfo::buildAbi().
Change-Id: I2d581b22326da4220f412ab4f517156f4ba31897
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
