Kevin Enderby [Tue, 18 Oct 2016 17:54:17 +0000 (17:54 +0000)]
Next set of additional error checks for invalid Mach-O files for the
load commands that use the MachO::routines_command and
and MachO::routines_command_64 types but are not used in llvm
libObject code but used in llvm tool code.
This includes the LC_ROUTINES and LC_ROUTINES_64
load commands.
llvm-svn: 284504
Justin Lebar [Tue, 18 Oct 2016 17:50:39 +0000 (17:50 +0000)]
[ADT] Remove CachedHash<T>.
Nobody is using it.
Differential Revision: https://reviews.llvm.org/D25630
llvm-svn: 284503
Justin Lebar [Tue, 18 Oct 2016 17:50:36 +0000 (17:50 +0000)]
Use CachedHashStringRef instead of CachedHash<StringRef>.
Summary:
This uses one less word on 64-bit platforms, so should be a strict
improvement. This change also lets us get rid of llvm::CachedHash.
Reviewers: rafael, timshen
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D25646
llvm-svn: 284502
Sanjoy Das [Tue, 18 Oct 2016 17:45:16 +0000 (17:45 +0000)]
[SCEV] Make CompareValueComplexity a little bit smarter
This helps canonicalization in some cases.
Thanks to Pankaj Chawla for the investigation and the test case!
llvm-svn: 284501
Sanjoy Das [Tue, 18 Oct 2016 17:45:13 +0000 (17:45 +0000)]
[SCEV] Extract out a helper function; NFC
llvm-svn: 284500
Jonathan Peyton [Tue, 18 Oct 2016 17:39:06 +0000 (17:39 +0000)]
Fix OpenMP 4.0 library build
Patch by Andrey Churbanov
Differential Revision: https://reviews.llvm.org/D25505
llvm-svn: 284499
Simon Pilgrim [Tue, 18 Oct 2016 17:31:07 +0000 (17:31 +0000)]
[X86][SSE] Added vector ashr combine tests
This doesn't cover all combines in DAGCombiner::visitSRA yet, but identifies several cases where we fail to combine vectors (or non-splatted) vectors
llvm-svn: 284498
Teresa Johnson [Tue, 18 Oct 2016 17:17:37 +0000 (17:17 +0000)]
[docs] Increase minimum supported GCC version for building LLVM to 4.8
Summary:
The RFC proposal sent to increase the minimum required GCC version
to 4.8 received a lot of support. See the following thread:
http://lists.llvm.org/pipermail/llvm-dev/2016-October/105955.html,
This patch implements that by updating the docs. I believe the
references to libstdc++ 4.7 issues can be removed as well, please
let me know if that is not the case or if they should be updated
a different way.
Reviewers: rengolin
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D25683
llvm-svn: 284497
Michal Gorny [Tue, 18 Oct 2016 17:07:30 +0000 (17:07 +0000)]
[cmake] Update lit search to match the one in LLVM
Update the lit search logic to support all names supported in LLVM
(since r283029). The search order (i.e. PATHS vs HINTS) does no really
matter since the established path is not used, except for determining
whether lit is available.
Differential Revision: https://reviews.llvm.org/D23745
llvm-svn: 284496
Sanjay Patel [Tue, 18 Oct 2016 17:05:05 +0000 (17:05 +0000)]
[Target] remove TargetRecip class; move reciprocal estimate isel functionality to TargetLowering
This is a follow-up to D24816 - where we changed reciprocal estimates to be function attributes
rather than TargetOptions.
This patch is intended to be a structural, but not functional change. By moving all of the
TargetRecip functionality into TargetLowering, we can remove all of the reciprocal estimate
state, shield the callers from the string format implementation, and simplify/localize the
logic needed for a target to enable this.
If a function has a "reciprocal-estimates" attribute, those settings may override the target's
default reciprocal preferences for whatever operation and data type we're trying to optimize.
If there's no attribute string or specific setting for the op/type pair, just use the target
default settings.
As noted earlier, a better solution would be to move the reciprocal estimate settings to IR
instructions and SDNodes rather than function attributes, but that's a multi-step job that
requires infrastructure improvements. I intend to work on that, but it's not clear how long
it will take to get all the pieces in place.
Differential Revision: https://reviews.llvm.org/D25440
llvm-svn: 284495
Michal Gorny [Tue, 18 Oct 2016 16:54:59 +0000 (16:54 +0000)]
[solaris] Convert the support library to C++ to fix -std=c++11 build
Convert the Solaris xlocale.c compatibility library from plain C to C++
in order to fix the build failures caused by the addition of -std=c++11
to LIBCXX_COMPILE_FLAGS. The additional flag got propagated to the C
file, resulting in error with strict compilers.
Differential Revision: https://reviews.llvm.org/D25431
llvm-svn: 284494
Michal Gorny [Tue, 18 Oct 2016 16:54:54 +0000 (16:54 +0000)]
[solaris] Fix iswxdigit_l() support function prototype
Fix the iswxdigit_l() function prototype to take wint_t parameter
instead of incorrect wchar_t.
Differential Revision: https://reviews.llvm.org/D25431
llvm-svn: 284493
Michal Gorny [Tue, 18 Oct 2016 16:38:44 +0000 (16:38 +0000)]
Fix a compile error on musl-libc due to strerror_r() prototype
Function strerror_r() has different signatures in different
implementations of libc: glibc's version returns a char*, while BSDs
and musl return a int. libomp unconditionally assumes glibc on Linux
and thus fails to compile against musl-libc. This patch addresses this
issue.
Differential Revision: https://reviews.llvm.org/D25071
llvm-svn: 284492
Simon Pilgrim [Tue, 18 Oct 2016 16:36:00 +0000 (16:36 +0000)]
[DAGCombiner] Add splatted vector support to (udiv x, (shl pow2, y)) -> x >>u (log2(pow2)+y)
llvm-svn: 284491
Adrian Prantl [Tue, 18 Oct 2016 16:24:43 +0000 (16:24 +0000)]
Disable fatal errors in the Verifier instantiated by bugpoint's crash
debugger.
When bugpoint hacks at a testcase it may at one point create illegal
debug info metadata that won't even pass the Verifier. A bugpoint
*driver* built with assertions should not assert on it, but reject the
malformed intermediate step and continue to do its job.
llvm-svn: 284490
Jonathan Peyton [Tue, 18 Oct 2016 16:20:55 +0000 (16:20 +0000)]
Mixed type atomic routines added for capture and update/capture reverse.
New mixed type atomic routines added for regular capture operations as well as
reverse update/capture operations. LHS - all integer and float types (no
complex so far), RHS - float16.
Patch by Olga Malysheva
Differential Revision: https://reviews.llvm.org/D25275
llvm-svn: 284489
Simon Pilgrim [Tue, 18 Oct 2016 15:45:37 +0000 (15:45 +0000)]
[X86][AVX512] Add mask/maskz writemask support to constant pool shuffle decode commentx
llvm-svn: 284488
Rafael Espindola [Tue, 18 Oct 2016 15:31:23 +0000 (15:31 +0000)]
Use reserve instead of grow.
The table was still being resized as grow doesn't account for the fact
that the table needs to remain 3/4 full.
llvm-svn: 284487
Erik Pilkington [Tue, 18 Oct 2016 15:26:43 +0000 (15:26 +0000)]
Revert r284265 "[Sema] Refactor context checking for availability diagnostics"
This has a bug in it, pointed out by Bob Wilson!
llvm-svn: 284486
Simon Dardis [Tue, 18 Oct 2016 15:17:17 +0000 (15:17 +0000)]
[mips][ias] Handle more complicated expressions for memory operands
This patch teaches ias for mips to handle expressions such as
(8*4)+(8*31)($sp). Such expression typically occur from the expansion
of multiple macro definitions.
This partially resolves PR/30383.
Thanks to Sean Bruno for reporting the issue!
Reviewers: zoran.jovanovic, vkalintiris
Differential Revision: https://reviews.llvm.org/D24667
llvm-svn: 284485
Todd Fiala [Tue, 18 Oct 2016 15:15:24 +0000 (15:15 +0000)]
xfail TestMiSyntax.py's test_lldbmi_output_grammar on macOS
Needs to be investigated. This is failing locally and on the
Xcode CI.
rdar://
28805064
llvm-svn: 284484
Simon Dardis [Tue, 18 Oct 2016 14:42:13 +0000 (14:42 +0000)]
[mips] Fix sync instruction definition
The 'sync' instruction for MIPS was defined in MIPS-II as taking no operands.
MIPS32 extended the define of 'sync' as taking an optional unsigned 5 bit
immediate.
This patch correct the definition of sync so that it is accepted with an
operand of 0 or no operand for MIPS-II to MIPS-V, and a 5 bit unsigned
immediate for MIPS32 and later revisions.
Additionally a clear error is given when the MIPS32 version of sync is
used when targeting pre MIPS32.
This partially resolves PR/30714.
Thanks to Daniel Sanders for reporting this issue!
Reveiwers: vkalintiris
Differential Revision: https://reviews.llvm.org/D25672
llvm-svn: 284483
Victor Leschuk [Tue, 18 Oct 2016 14:31:22 +0000 (14:31 +0000)]
DebugInfo: change alignment type from uint64_t to uint32_t to save space.
In futher patches we shall have alignment field added to DIVariable family
and switching from uint64_t to uint32_t will save 4 bytes per variable.
Differential Revision: https://reviews.llvm.org/D25620
llvm-svn: 284482
Simon Dardis [Tue, 18 Oct 2016 14:28:00 +0000 (14:28 +0000)]
[mips] Macro expansion for ld, sd for O32
ld and sd when assembled for the O32 ABI expand to a pair of 32 bit word loads
or stores using the specified source or destination register and the next
register.
This patch does not add support for the cases where the offset is greater than
a 16 bit signed immediate as that would lead to a wrong/misleading error
message as the assembler would report "instruction requires a CPU feature
not currently enabled" for ld & sd for MIPS64 when their offset is not a signed
16 bit number.
This fixes PR/29159.
Thanks to Sean Bruno for reporting this issue!
Reviewers: vkalintiris, seanbruno, zoran.jovanovic
Differential Review: https://reviews.llvm.org/D24556
llvm-svn: 284481
Michael Zuckerman [Tue, 18 Oct 2016 13:52:39 +0000 (13:52 +0000)]
[x86][inline-asm][avx512] allow swapping of '{k<num>}' & '{z}' marks
Committing on behalf of Coby Tayree: After check-all and LGTM
Desc:
AVX512 allows dest operand to be followed by an op-mask register specifier ('{k<num>}', which in turn may be followed by a merging/zeroing specifier ('{z}')
Currently, the following forms are allowed:
{k<num>}
{k<num>}{z}
This patch allows the following forms:
{z}{k<num>}
and ignores the next form:
{z}
Justification would be quite simple - GCC
Differential Revision: http://reviews.llvm.org/D25013
llvm-svn: 284479
Simon Pilgrim [Tue, 18 Oct 2016 13:44:00 +0000 (13:44 +0000)]
Strip trailing whitespace (NFCI)
llvm-svn: 284478
Daniel Marjamaki [Tue, 18 Oct 2016 13:16:53 +0000 (13:16 +0000)]
alpha.core.UnreachableCode - don't warn about unreachable code inside macro
In macros, 'do {...} while (0)' is often used. Don't warn about the condition 0 when it is unreachable.
Differential Revision: https://reviews.llvm.org/D25606
llvm-svn: 284477
Simon Pilgrim [Tue, 18 Oct 2016 13:15:31 +0000 (13:15 +0000)]
Fix signed/unsigned comparison warnings
llvm-svn: 284476
Vasileios Kalintiris [Tue, 18 Oct 2016 13:05:42 +0000 (13:05 +0000)]
[mips][FastISel] Instantiate the MipsFastISel class only for targets that support FastISel.
Summary:
Instead of instantiating the MipsFastISel class and checking if the
target is supported in the overriden methods, we should perform that
check before creating the class. This allows us to enable FastISel *only*
for targets that truly support it, ie. MIPS32 to MIPS32R5.
Reviewers: sdardis
Subscribers: ehostunreach, llvm-commits
Differential Revision: https://reviews.llvm.org/D24824
llvm-svn: 284475
George Rimar [Tue, 18 Oct 2016 11:11:03 +0000 (11:11 +0000)]
[ELF] - Update testcases after llvm change.
llvm-svn: 284474
Artem Dergachev [Tue, 18 Oct 2016 11:06:28 +0000 (11:06 +0000)]
[analyzer] Add NumberObjectConversion checker.
When dealing with objects that represent numbers, such as Objective-C NSNumber,
the language provides little protection from accidentally interpreting
the value of a pointer to such object as the value of the number represented
by the object. Results of such mis-interpretation may be unexpected.
The checker attempts to fill this gap in cases when the code is obviously
incorrect.
With "Pedantic" option enabled, this checker enforces a coding style to
completely prevent errors of this kind (off by default).
Differential Revision: https://reviews.llvm.org/D22968
llvm-svn: 284473
Alex Lorenz [Tue, 18 Oct 2016 10:55:01 +0000 (10:55 +0000)]
[CodeCompletion] Add a block property setter completion result
This commit changes code completion results for Objective-C block properties:
clang now suggests an additional completion result that displays the block
property together with '=' and the block literal placeholder for the appropriate
readwrite block properties.
This commit uses a simple heuristic to determine when it's appropriate to
suggest a setter completion for block properties: the additional block setter
completion is provided iff the member access that's being completed is a
standalone statement.
rdar://
28481726
Differential Revision: https://reviews.llvm.org/D25520
llvm-svn: 284472
George Rimar [Tue, 18 Oct 2016 10:54:56 +0000 (10:54 +0000)]
[llvm-readobj] - Teach readobj to print PT_OPENBSD_RANDOMIZE/PT_OPENBSD_WXNEEDED headers.
These are OpenBSD specific program headers and
currently we support them in LLD.
Description of headers (just in case) available here:
http://man.openbsd.org/OpenBSD-current/man5/elf.5
OpenBSD commits were:
For PT_OPENBSD_RANDOMIZE:
https://github.com/openbsd/src/commit/
c494713c450d98da3f2e1451ee8c7fb675a7c461
For PT_OPENBSD_WXNEEDED:
https://github.com/openbsd/src/commit/
2a5a8fc7e30928c2cff57cfe5fb491c90d8478ad
Differential revision: https://reviews.llvm.org/D25616
llvm-svn: 284471
George Rimar [Tue, 18 Oct 2016 10:49:50 +0000 (10:49 +0000)]
[ELF] - Linkerscript: accept integer values for PHDRS types.
Both gold and ld accepts integers instead of named constants
for PHDRS.
Patch adds support for that.
Differential revision: https://reviews.llvm.org/D25549
llvm-svn: 284470
Pavel Labath [Tue, 18 Oct 2016 10:46:45 +0000 (10:46 +0000)]
Fixup r284466 - try to unbreak NetBSD
NetBSD does not have getopt as well - we need to apply the workaround there too.
FreeBSD seems to be fine though.
llvm-svn: 284469
Alex Lorenz [Tue, 18 Oct 2016 10:38:58 +0000 (10:38 +0000)]
[CodeCompletion][NFC] Extract a function that formats block placeholders.
This commit extracts a new function named `formatBlockPlaceholder` from
the function `FormatFunctionParameter` so that it can be reused in follow-up
commits that improve code completion for block property setters.
Differential Revision: https://reviews.llvm.org/D25519
llvm-svn: 284468
Alex Lorenz [Tue, 18 Oct 2016 10:35:27 +0000 (10:35 +0000)]
[CodeCompletion][NFC] Extract a function that looks for block decl type locs.
This commit extracts a new function named `findTypeLocationForBlockDecl` from
the function `FormatFunctionParameter` so that it can be reused in follow-up
commits that improve code completion for block property setters.
Differential Revision: https://reviews.llvm.org/D25519
llvm-svn: 284467
Pavel Labath [Tue, 18 Oct 2016 10:26:57 +0000 (10:26 +0000)]
[cmake] Make dependencies of lldb libraries private, take 2
Summary:
The dependencies of our libraries (only liblldb, really) we marked as public, which caused all
their dependencies to be repeated when linking any executables to them. This is a problem because
then all the .a files could end up being linked twice, once to liblldb and once
again to to the executable linking against liblldb (lldb, lldb-mi). As it turns out,
our build actually depends on this behavior:
- on windows, lldb does not have getopt, so it pulls it from inside liblldb, even
though getopt is not a part of the exported interface of liblldb (maybe some of
the bsd variants have this problem as well)
- lldb-mi uses llvm, which again is not exported by liblldb
This change does not actually fix these problems (that is going to be a hard
one), but it does make them explicit by moving this magic from add_lldb_library
to the places the executable targets are defined. That way, I can link the
additional .a files only on targets that really need it, and the other targets
can build cleanly and make sure we don't regress further. It also fixes the
LLVM_LINK_LLVM_DYLIB build on linux.
Reviewers: zturner, beanz
Subscribers: ki.stfu, lldb-commits, mgorny
Differential Revision: https://reviews.llvm.org/D25680
llvm-svn: 284466
John Brawn [Tue, 18 Oct 2016 10:10:53 +0000 (10:10 +0000)]
[SCEV] More accurate calculation of max backedge count of some less-than loops
In loops that look something like
i = n;
do {
...
} while(i++ < n+k);
where k is a constant, the maximum backedge count is k (in fact the backedge
count will be either 0 or k, depending on whether n+k wraps). More generally
for LHS < RHS if RHS-(LHS of first comparison) is a constant then the loop will
iterate either 0 or that constant number of times.
This allows for more loop unrolling with the recent upper bound loop unrolling
changes, and I'm working on a patch that will let loop unrolling additionally
make use of the loop being executed either 0 or k times (we need to retain the
loop comparison only on the first unrolled iteration).
Differential Revision: https://reviews.llvm.org/D25607
llvm-svn: 284465
Igor Kudrin [Tue, 18 Oct 2016 09:42:03 +0000 (09:42 +0000)]
[libclang] Add missing cursor kinds to python bindings.
Differential Revision: https://reviews.llvm.org/D25673
llvm-svn: 284464
Igor Kudrin [Tue, 18 Oct 2016 09:30:33 +0000 (09:30 +0000)]
[libclang] Fix a failure in a test for python bindings on CursorKind.OVERLOAD_CANDIDATE.
The test fails because the value does not lay in any existing group.
Differential Revision: https://reviews.llvm.org/D25470
llvm-svn: 284463
Renato Golin [Tue, 18 Oct 2016 09:30:18 +0000 (09:30 +0000)]
Revert "Resubmit "Add support for advanced number formatting.""
This reverts commits 284436 and 284437 because they still break AArch64 bots:
Value of: format_number(-10, IntegerStyle::Integer, 1)
Actual: "-0"
Expected: "-10"
llvm-svn: 284462
Simon Pilgrim [Tue, 18 Oct 2016 09:29:13 +0000 (09:29 +0000)]
[X86][SSE] Added extra (mul x, (1 << c)) -> x << c style vector tests
vXi64 will benefit more from lowering to shifts than multiplies
llvm-svn: 284461
Javed Absar [Tue, 18 Oct 2016 09:08:54 +0000 (09:08 +0000)]
[ARM] Assign cost of scaling for Cortex-R52
This patch assigns cost of the scaling used in addressing for Cortex-R52.
On Cortex-R52 a negated register offset takes longer than a non-negated
register offset, in a register-offset addressing mode.
Differential Revision: http://reviews.llvm.org/D25670
Reviewer: jmolloy
llvm-svn: 284460
Simon Pilgrim [Tue, 18 Oct 2016 07:42:15 +0000 (07:42 +0000)]
[X86][SSE] Add lowering to cvttpd2dq/cvttps2dq for sitofp v2f64/2f32 to 2i32
As discussed on PR28461 we currently miss the chance to lower "fptosi <2 x double> %arg to <2 x i32>" to cvttpd2dq due to its use of illegal types.
This patch adds support for fptosi to 2i32 from both 2f64 and 2f32.
It also recognises that cvttpd2dq zeroes the upper 64-bits of the xmm result (similar to D23797) - we still don't do this for the cvttpd2dq/cvttps2dq intrinsics - this can be done in a future patch.
Differential Revision: https://reviews.llvm.org/D23808
llvm-svn: 284459
Richard Smith [Tue, 18 Oct 2016 07:13:55 +0000 (07:13 +0000)]
[c++1z] Include "noexcept" in builtin function types where appropriate. Fixes
an assertion failure looking up a matching ::operator delete for
__builtin_operator_delete.
llvm-svn: 284458
Richard Smith [Tue, 18 Oct 2016 06:47:03 +0000 (06:47 +0000)]
[c++1z] Use canonical expression equivalence to determine whether two different
dependent noexcept specifications result in the same canonical function type.
We still use non-canonical hashing when deduplicating type sugar so that
diagnostics will point to the right place.
llvm-svn: 284457
Dean Michael Berris [Tue, 18 Oct 2016 05:54:15 +0000 (05:54 +0000)]
[XRay] Support for for tail calls for ARM no-Thumb
This patch adds simplified support for tail calls on ARM with XRay instrumentation.
Known issue: compiled with generic flags: `-O3 -g -fxray-instrument -Wall
-std=c++14 -ffunction-sections -fdata-sections` (this list doesn't include my
specific flags like --target=armv7-linux-gnueabihf etc.), the following program
#include <cstdio>
#include <cassert>
#include <xray/xray_interface.h>
[[clang::xray_always_instrument]] void __attribute__ ((noinline)) fC() {
std::printf("In fC()\n");
}
[[clang::xray_always_instrument]] void __attribute__ ((noinline)) fB() {
std::printf("In fB()\n");
fC();
}
[[clang::xray_always_instrument]] void __attribute__ ((noinline)) fA() {
std::printf("In fA()\n");
fB();
}
// Avoid infinite recursion in case the logging function is instrumented (so calls logging
// function again).
[[clang::xray_never_instrument]] void simplyPrint(int32_t functionId, XRayEntryType xret)
{
printf("XRay: functionId=%d type=%d.\n", int(functionId), int(xret));
}
int main(int argc, char* argv[]) {
__xray_set_handler(simplyPrint);
printf("Patching...\n");
__xray_patch();
fA();
printf("Unpatching...\n");
__xray_unpatch();
fA();
return 0;
}
gives the following output:
Patching...
XRay: functionId=3 type=0.
In fA()
XRay: functionId=3 type=1.
XRay: functionId=2 type=0.
In fB()
XRay: functionId=2 type=1.
XRay: functionId=1 type=0.
XRay: functionId=1 type=1.
In fC()
Unpatching...
In fA()
In fB()
In fC()
So for function fC() the exit sled seems to be called too much before function
exit: before printing In fC().
Debugging shows that the above happens because printf from fC is also called as
a tail call. So first the exit sled of fC is executed, and only then printf is
jumped into. So it seems we can't do anything about this with the current
approach (i.e. within the simplification described in
https://reviews.llvm.org/D23988 ).
Differential Revision: https://reviews.llvm.org/D25030
llvm-svn: 284456
Craig Topper [Tue, 18 Oct 2016 05:44:04 +0000 (05:44 +0000)]
[AVX-512] Add test case to check shuffle decoding for masked vpermilps for r284450.
This is harder to do for vpermilpd as shuffle combining turns the constant vector into an immediate since all vpermilpd's inputs with constant vector can also be encoded with the immediate form.
llvm-svn: 284455
Justin Bogner [Tue, 18 Oct 2016 05:17:23 +0000 (05:17 +0000)]
Object: Add a missing return in ObjectFile::createObjectFile
When Error was threaded through these APIs back in r265606 the
"return" was missed here, which triggers a warning if/when I add
LLVM_NODISCARD to the Error type.
llvm-svn: 284454
Craig Topper [Tue, 18 Oct 2016 04:48:33 +0000 (04:48 +0000)]
[X86] Fix DecodeVPERMVMask to handle cases where the constant pool entry has a different type than the shuffle itself.
This is especially important for 32-bit targets with 64-bit shuffle elements.
llvm-svn: 284453
Andrew Trick [Tue, 18 Oct 2016 04:17:44 +0000 (04:17 +0000)]
Improve tablegen gen-subtarget diagnostics for missing machine models.
-debug-only=subtarget-emitter prints a lot of machine model diagnostics.
This prunes the output so that the "No machine model for XXX on processor YYY"
only appears when there is definitely no machine model for that opcode.
Previously it was printing that error even if the opcode was covered by
a more general scheduling class.
<rdar://problem/
15919845> [TableGen][CodeGenSchedule] Debug output does not help spotting the missing scheduling classes
llvm-svn: 284452
Craig Topper [Tue, 18 Oct 2016 04:00:32 +0000 (04:00 +0000)]
[AVX-512] Fix DecodeVPERMV3Mask to handle cases where the constant pool entry has a different type than the shuffle itself.
Summary: This is especially important for 32-bit targets with 64-bit shuffle elements.This is similar to how PSHUFB and VPERMIL handle the same problem.
Reviewers: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D25666
llvm-svn: 284451
Craig Topper [Tue, 18 Oct 2016 03:36:52 +0000 (03:36 +0000)]
[AVX-512] Add support for decoding shuffle mask from constant pool for masked VPERMILPS/PD.
llvm-svn: 284450
Jim Ingham [Tue, 18 Oct 2016 01:52:32 +0000 (01:52 +0000)]
Remove a debug print statement.
llvm-svn: 284448
Jim Ingham [Tue, 18 Oct 2016 01:43:22 +0000 (01:43 +0000)]
More testsuite xfail markings cleanup.
llvm-svn: 284446
Dominic Chen [Tue, 18 Oct 2016 01:15:19 +0000 (01:15 +0000)]
[analyzer] Update alpha and potential checker documentation, esp. alpha.valist
Summary:
Move alpha.valist from potential to alpha since it was implemented in D15227
Cleanup some HTML comments, add a missing link
Reviewers: jordan_rose, zaks.anna
Subscribers: cfe-commits, xazax.hun
Differential Revision: https://reviews.llvm.org/D25663
llvm-svn: 284445
Chris Bieneman [Tue, 18 Oct 2016 00:50:39 +0000 (00:50 +0000)]
[CMake] Add missing dependency on intrinsics_gen
COFF/InputFiles.h includes LTOModule.h which transitively relies on Attributes.inc.
llvm-svn: 284444
Chris Bieneman [Tue, 18 Oct 2016 00:50:20 +0000 (00:50 +0000)]
[CMake] Add a few default passthrough variables for bootstrap builds
This just passes through a few missing CMake variables for multi-stage builds.
llvm-svn: 284443
Vedant Kumar [Tue, 18 Oct 2016 00:23:27 +0000 (00:23 +0000)]
[Basic] unique_ptr-ify SourceManager::MacroArgsCacheMap (NFC)
Differential Revision: https://reviews.llvm.org/D25711
llvm-svn: 284442
Mandeep Singh Grang [Tue, 18 Oct 2016 00:11:19 +0000 (00:11 +0000)]
Fix differences in codegen between Linux and Windows toolchains
Summary:
There are differences in codegen between Linux and Windows due to:
1. Using std::sort which uses quicksort which is a non-stable sort.
2. Iterating over Set data structure where the iteration order is
non deterministic.
Reviewers: arsenm, grosbach, junbuml, zinob, MatzeB
Subscribers: MatzeB, wdng
Differential Revision: https://reviews.llvm.org/D25695
llvm-svn: 284441
Vedant Kumar [Tue, 18 Oct 2016 00:02:28 +0000 (00:02 +0000)]
[profile] Mark lprofCurFilename as COMPILER_RT_WEAK
This makes __llvm_profile_set_filename() work across dylib boundaries on
Darwin.
This functionality was originally meant to work on all platforms, but
was moved to a Linux-only directory with r272404. The root cause of the
test failure on Darwin was that lprofCurFilename was not marked weak.
Each dylib maintained its own copy of the variable due to the two-level
namespace.
Tested with check-profile (on Darwin). I don't expect this to regress
other platforms.
Differential Revision: https://reviews.llvm.org/D25707
llvm-svn: 284440
Jim Ingham [Mon, 17 Oct 2016 23:59:41 +0000 (23:59 +0000)]
Fix a crash in expressions with fixits in the dummy target.
In the expression command, if the target is NULL, you have to use the dummy
target.
<rdar://problem/
28811687>
llvm-svn: 284439
Zachary Turner [Mon, 17 Oct 2016 23:08:47 +0000 (23:08 +0000)]
Rename HexStyle -> HexFormatStyle, and remove a constexpr.
This should fix the remaining broken builds.
llvm-svn: 284437
Zachary Turner [Mon, 17 Oct 2016 22:49:24 +0000 (22:49 +0000)]
Resubmit "Add support for advanced number formatting."
This resubmits commits 284425 and r284428, which were reverted
in r284429 due to some infinite recursion caused by an incorrect
selection of function overloads. Reproduced the failure on Linux
using GCC 4.8.4, and confirmed that with the new patch the tests
path on GCC as well as MSVC. So hopefully this fixes everything.
llvm-svn: 284436
Konstantin Zhuravlyov [Mon, 17 Oct 2016 22:40:15 +0000 (22:40 +0000)]
[AMDGPU] Mark .note section SHF_ALLOC so lld creates a segment for it
Differential Revision: https://reviews.llvm.org/D25694
llvm-svn: 284435
Justin Lebar [Mon, 17 Oct 2016 22:24:36 +0000 (22:24 +0000)]
[ADT] Move CachedHashString to its own header in ADT, and rename to CachedHashStringRef.
Summary:
Reclaiming the name 'CachedHashString' will let us add a type with that
name that owns its value.
Reviewers: timshen
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D25644
llvm-svn: 284434
Justin Lebar [Mon, 17 Oct 2016 22:24:32 +0000 (22:24 +0000)]
[ADT] Add an initializer_list constructor to {Small,}DenseSet.
Reviewers: timshen
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D25629
llvm-svn: 284433
Justin Lebar [Mon, 17 Oct 2016 22:24:28 +0000 (22:24 +0000)]
[ADT] Add SmallDenseSet.
Summary: This matches SmallDenseMap.
Reviewers: timshen
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D25628
llvm-svn: 284432
Kevin Enderby [Mon, 17 Oct 2016 22:09:25 +0000 (22:09 +0000)]
Next set of additional error checks for invalid Mach-O files for the
load commands that use the MachO::sub_framework_command,
MachO::sub_umbrella_command, MachO::sub_library_command
and MachO::sub_client_command types but are not used in llvm
libObject code but used in llvm tool code.
This includes the LC_SUB_FRAMEWORK, LC_SUB_UMBRELLA,
LC_SUB_LIBRARY and LC_SUB_CLIENT load commands.
llvm-svn: 284431
Petr Hosek [Mon, 17 Oct 2016 22:02:53 +0000 (22:02 +0000)]
[Driver] Use stem rather than filename for executable name
When comparing the linker name in Fuchsia driver, use stem rather
than filename to get the name of the linker becase on Windows, the
filename will have an extension.
Differential Revision: https://reviews.llvm.org/D25700
llvm-svn: 284430
Zachary Turner [Mon, 17 Oct 2016 21:25:41 +0000 (21:25 +0000)]
Revert formatting changes.
This reverts r288425 and r284428 as they are causing test crashes
on some systems.
llvm-svn: 284429
Zachary Turner [Mon, 17 Oct 2016 21:14:27 +0000 (21:14 +0000)]
Try to fix build after invalid pointer conversion.
llvm-svn: 284428
Sanjay Patel [Mon, 17 Oct 2016 21:08:39 +0000 (21:08 +0000)]
remove FIXME comment (fixed with r284424); NFC
llvm-svn: 284427
Eli Friedman [Mon, 17 Oct 2016 21:04:26 +0000 (21:04 +0000)]
Handle multi-dimensional invariant load.
If the address of a load depends on another load, make sure to emit
the loads in the right order.
llvm-svn: 284426
Zachary Turner [Mon, 17 Oct 2016 20:57:45 +0000 (20:57 +0000)]
[Support] Add support for "advanced" number formatting.
raw_ostream has not afforded a lot of flexibility in terms of
how to format numbers when outputting. Wrap this all up into
a set of low level helper functions that can be used to output
numbers with arbitrary precision, alignment, format, etc and
then update raw_ostream to use these functions.
This will be useful for upcoming improvements to llvm's string
formatting libraries, but are still useful independently.
Differential Revision: https://reviews.llvm.org/D25497
llvm-svn: 284425
Sanjay Patel [Mon, 17 Oct 2016 20:41:39 +0000 (20:41 +0000)]
[DAG] use isConstOrConstSplat in ComputeNumSignBits to optimize SRA
The scalar version of this pattern was noted in:
https://reviews.llvm.org/D25485
and fixed with:
https://reviews.llvm.org/rL284395
More refactoring of the constant/splat helpers is needed and will happen in follow-up patches.
Differential Revision: https://reviews.llvm.org/D25685
llvm-svn: 284424
Adrian Prantl [Mon, 17 Oct 2016 20:37:56 +0000 (20:37 +0000)]
Explicitly pass an isysroot to avoid the SDKROOT overriding the deployment target.
This fixes the green dragon builders after r284416.
llvm-svn: 284423
Walter Erquinigo [Mon, 17 Oct 2016 20:28:19 +0000 (20:28 +0000)]
[lldb] Read modules from memory when a local copy is not available
Summary:
When the local lldb doesn't have access to a copy of the modules in the target, e.g. winphone, with this change now we read these modules from memory.
There are mainly 2 changes:
1. create pecoff object files from memory
2. read from memory when the local file is not available
Reviewers: sas, fjricci, zturner
Subscribers: #lldb
Tags: #lldb
Differential Revision: https://reviews.llvm.org/D24284
llvm-svn: 284422
Sanjay Patel [Mon, 17 Oct 2016 20:26:46 +0000 (20:26 +0000)]
[DAG] make isConstOrConstSplat and isConstOrConstSplatFP more accessible; NFC
As noted in:
https://reviews.llvm.org/D25685
This is the next-to-smallest step needed to enable the ComputeNumSignBits fix in that patch.
In a minor attempt to keep some structure, we're pulling the FP helper over along with its
integer sibling, but clearly we can and should do more refactoring of the similar helper
functions in DAGCombiner and SelectionDAG to simplify and not duplicate functionality.
llvm-svn: 284421
Adrian Prantl [Mon, 17 Oct 2016 20:14:23 +0000 (20:14 +0000)]
Improve the CHECK lines in debug-options.c by separating out the check
for debug info kind and dwarf version.
llvm-svn: 284420
Davide Italiano [Mon, 17 Oct 2016 20:06:32 +0000 (20:06 +0000)]
[Coverage] Update test after r284418.
We now strip coverage metadata if debug info are not present.
llvm-svn: 284419
Davide Italiano [Mon, 17 Oct 2016 20:05:35 +0000 (20:05 +0000)]
[opt] Strip coverage if debug info is not present.
If -coverage is passed, but -g is not, clang populates the PassManager
pipeline with StripSymbols(debugOnly = true).
The stripSymbol pass therefore scans the list of named metadata,
drops !llvm.dbg.cu, but leaves !llvm.gcov and !0 (the compileUnit MD)
around. The verifier runs, and finds out that there's a CU not listed
in !llvm.dbg.cu (as it was previously dropped) -> crash.
When we strip debug info, so, check if there's coverage data,
and strip it as well, in order to avoid pending metadata left around.
Differential Revision: https://reviews.llvm.org/D25689
llvm-svn: 284418
Adrian Prantl [Mon, 17 Oct 2016 19:46:26 +0000 (19:46 +0000)]
Update testcase for r284416.
llvm-svn: 284417
Adrian Prantl [Mon, 17 Oct 2016 19:36:18 +0000 (19:36 +0000)]
Driver/Darwin: Set the DWARF version based on the deployment target.
System utilities such as atos only support DWARF 4 on OS X 10.11+ and
iOS 9+. We thus want to enable DWARF 4 only if the deployment target
has a recent enough operating system version and use DWARF 2 for older
systems.
<rdar://problem/
28766743>
llvm-svn: 284416
Dehao Chen [Mon, 17 Oct 2016 19:28:44 +0000 (19:28 +0000)]
Ignore debug info when making optimization decisions in SimplifyCFG.
Summary: Debug info should *not* affect code generation. This patch properly handles debug info to make sure the generated code are the same with or without debug info.
Reviewers: davidxl, mzolotukhin, jmolloy
Subscribers: aprantl, llvm-commits
Differential Revision: https://reviews.llvm.org/D25286
llvm-svn: 284415
Rafael Espindola [Mon, 17 Oct 2016 19:25:42 +0000 (19:25 +0000)]
Delete dead code.
llvm-svn: 284414
Rafael Espindola [Mon, 17 Oct 2016 19:25:08 +0000 (19:25 +0000)]
Avoid using getComdatSymbolTable.
This is not particularly efficient, but does avoid exposing Comdat*
out of LTO.h.
I will send a patch for review with a more efficient interface that
should map nicely to a bitcode symbol table.
llvm-svn: 284413
Mehdi Amini [Mon, 17 Oct 2016 19:23:19 +0000 (19:23 +0000)]
[doc] use double `` to prevent html output of merging double dash
llvm-svn: 284412
Michael LeMay [Mon, 17 Oct 2016 19:09:19 +0000 (19:09 +0000)]
Test commit.
llvm-svn: 284411
Walter Erquinigo [Mon, 17 Oct 2016 18:56:18 +0000 (18:56 +0000)]
Handle relocations to thumb functions when dynamic linking COFF modules
Summary:
This adds the necessary logic to support relocations to thumb functions in the COFF dynamic linker.
The jumps to function addresses are mostly blx, which requires the ISA selection bit when jumping to a thumb function.
Note: I'm determining if the relocation requires the ISA bit when creating the relocation entries and not when resolving the relocation. I have to do that because I need the ObjectFile and the actual Symbol, which are available only when creating the entries. It would require a gross refactor if I do it otherwise, but I'm okay with doing it if you think it's better.
Reviewers: peter.smith, compnerd
Subscribers: rengolin, sas
Differential Revision: https://reviews.llvm.org/D25151
llvm-svn: 284410
Rafael Espindola [Mon, 17 Oct 2016 18:51:27 +0000 (18:51 +0000)]
Update for llvm change.
llvm-svn: 284409
Rafael Espindola [Mon, 17 Oct 2016 18:51:02 +0000 (18:51 +0000)]
Return a StringRef instead of a Comdat*.
This is a small step in making this interface compatible with an
bitcode symbol table.
llvm-svn: 284408
Justin Bogner [Mon, 17 Oct 2016 18:40:35 +0000 (18:40 +0000)]
ELF: Implement skip() as (void)next()
Thanks to Rafael for pointing out the simplification.
llvm-svn: 284407
Tim Northover [Mon, 17 Oct 2016 18:36:53 +0000 (18:36 +0000)]
GlobalISel: support wider range of load/store sizes in AArch64.
llvm-svn: 284406
Justin Bogner [Mon, 17 Oct 2016 18:22:03 +0000 (18:22 +0000)]
unittests: Specify types in a bunch of unittest EXPECT's
The EXPECT and ASSERT macros in gtest don't do the usual arithmetic
conversions. Specify types in several of them to fix -Werror.
llvm-svn: 284405
Peter Smith [Mon, 17 Oct 2016 18:12:24 +0000 (18:12 +0000)]
[ELF] Support for R_ARM_TARGET2 relocation
The R_ARM_TARGET2 relocation is used in ARM exception tables to encode
a data dependency that will only be dereferenced by code in the
run-time support library. In a similar way to R_ARM_TARGET1 the
handling of the relocation is target specific, it maps to one of
R_ARM_ABS32, R_ARM_REL32 or R_ARM_GOT_PREL. The choice depends on the
run-time library. R_ARM_GOT_PREL is used for linux and BSD,
R_ARM_ABS32 and R_ARM_REL32 are used for bare-metal.
The command line option --target2=<target> can be used to select the
relocation used for R_ARM_TARGET2. The default is R_ARM_GOT_PREL.
Differential revision: https://reviews.llvm.org/D25684
llvm-svn: 284404
Michal Gorny [Mon, 17 Oct 2016 18:07:15 +0000 (18:07 +0000)]
[Driver] Use VFS to perform all distribution checks
Use the VFS provided by D.getVFS() for all distribution checks,
including those performing read of the release file. Requested
by @bruno on D24954.
Differential Revision: https://reviews.llvm.org/D25641
llvm-svn: 284403
Krzysztof Parzyszek [Mon, 17 Oct 2016 18:04:05 +0000 (18:04 +0000)]
Add a dummy file in each subdirectory in test/Driver/Inputs/hexagon_tree
Git does not store empty subdirectories (while SVN does). Git clone of
the clang repository did not create the fake Hexagon installation tree
used for testing the driver. This only became evident after a change
in the Hexagon toolchain that started checking for existence of certain
directories.
llvm-svn: 284402
Tim Northover [Mon, 17 Oct 2016 18:00:27 +0000 (18:00 +0000)]
Hexagon: add dummy files to test dir so git keeps them.
Should fix hexagon-elf-toolchain.c tests on Git.
llvm-svn: 284401