Simon Atanasyan [Wed, 17 Jun 2015 22:27:54 +0000 (22:27 +0000)]
[Mips] Support R_MICROMIPS_GPREL16 / R_MICROMIPS_GPREL7_S2 relocations handling
llvm-svn: 239960
Simon Atanasyan [Wed, 17 Jun 2015 22:27:39 +0000 (22:27 +0000)]
[Mips] Support R_MIPS_16 relocation handling
llvm-svn: 239959
Alexey Samsonov [Wed, 17 Jun 2015 22:27:32 +0000 (22:27 +0000)]
Revert "[Sanitizers] Provide better diagnostic for sanitizers unsupported for target triple."
This reverts commit r239953, while I'm investigating assertion
failure from http://lab.llvm.org:8080/green/job/clang-stage1-cmake-RA-incremental_check/9994/
llvm-svn: 239958
Simon Atanasyan [Wed, 17 Jun 2015 22:27:27 +0000 (22:27 +0000)]
[Mips] Do not use functions to calculate trivial relocations
No functional changes.
llvm-svn: 239957
Alexey Samsonov [Wed, 17 Jun 2015 22:24:49 +0000 (22:24 +0000)]
[CMake] Remove redundant checks in TSan CMakeLists.txt
llvm-svn: 239956
Alexey Samsonov [Wed, 17 Jun 2015 22:24:44 +0000 (22:24 +0000)]
[CMake] LSan is not actually available on Darwin.
llvm-svn: 239955
Richard Smith [Wed, 17 Jun 2015 22:13:23 +0000 (22:13 +0000)]
[modules] Fix typo in default argument merging.
llvm-svn: 239954
Alexey Samsonov [Wed, 17 Jun 2015 22:07:28 +0000 (22:07 +0000)]
[Sanitizers] Provide better diagnostic for sanitizers unsupported for target triple.
Summary:
Introduce ToolChain::getSupportedSanitizers() that would return the set
of sanitizers available on given toolchain. By default, these are
sanitizers which don't necessarily require runtime support (i.e.
set from -fsanitize=undefined-trap).
Sanitizers (ASan, DFSan, TSan, MSan etc.) which cannot function
without runtime library are marked as supported only on platforms
for which we actually build these runtimes.
This would allow more fine-grained checks in the future: for instance,
we have to restrict availability of -fsanitize=vptr to Mac OS 10.9+
(PR23539)
Update test cases accrodingly: add tests for certain unsupported
configurations, remove test cases for -fsanitize=vptr + PS4
integration, as we don't build the runtime for PS4 at the moment.
Test Plan: regression test suite
Reviewers: pcc
Subscribers: cfe-commits, filcab, eugenis, thakis, kubabrecka, emaste, rsmith
Differential Revision: http://reviews.llvm.org/D10467
llvm-svn: 239953
Pete Cooper [Wed, 17 Jun 2015 22:01:28 +0000 (22:01 +0000)]
Devirtualize and pack MCFragment to reduce memory usage.
MCFragment didn't really need vtables. The majority of virtual methods were just getters and setters.
This removes the vtables and uses dispatch on the kind to do things like delete which needs to
get the appropriate class.
This reduces memory on the verify use list order test case by about 2MB out of 800MB.
Reviewed by Rafael EspĂndola
llvm-svn: 239952
Davide Italiano [Wed, 17 Jun 2015 21:50:51 +0000 (21:50 +0000)]
[ELF] Now that there's an API for ELF symbol types in LLVM, use it.
Common symbols will be handled in a separate patch because it seems
Hexagon redefines the notion of common symbol, which I'm not (yet)
very familiar with.
llvm-svn: 239951
Reid Kleckner [Wed, 17 Jun 2015 21:50:02 +0000 (21:50 +0000)]
Re-land "[X86] Cache variables that only depend on the subtarget"
Re-instates r239949 without accidentally flipping the sense of UseLEA.
llvm-svn: 239950
Reid Kleckner [Wed, 17 Jun 2015 21:35:02 +0000 (21:35 +0000)]
Revert "[X86] Cache variables that only depend on the subtarget"
This reverts commit r239948, tests seem to be failing.
llvm-svn: 239949
Reid Kleckner [Wed, 17 Jun 2015 21:31:17 +0000 (21:31 +0000)]
[X86] Cache variables that only depend on the subtarget
There is a one-to-one relationship between X86Subtarget and
X86FrameLowering, but every frame lowering method would previously pull
the subtarget off the MachineFunction and query some subtarget
properties.
Over time, these locals began to grow in complexity and it became
important to keep their names and meaning in sync across all of the
frame lowering methods, leading to duplication. We can eliminate that
duplication by computing them once in the constructor.
llvm-svn: 239948
David Majnemer [Wed, 17 Jun 2015 21:21:16 +0000 (21:21 +0000)]
[docs] Fix "WARNING: Title underline too short."
llvm-svn: 239947
Benjamin Kramer [Wed, 17 Jun 2015 21:08:22 +0000 (21:08 +0000)]
Add missing include.
llvm-svn: 239946
Rui Ueyama [Wed, 17 Jun 2015 21:01:56 +0000 (21:01 +0000)]
COFF: Simplify. NFC.
Executor is a convenience class to run an external command.
llvm-svn: 239945
Benjamin Kramer [Wed, 17 Jun 2015 20:55:30 +0000 (20:55 +0000)]
[Bitcode] Replace hand-coded little endian handling with Endian.h functions.
No functional change intended.
llvm-svn: 239944
Matt Arsenault [Wed, 17 Jun 2015 20:55:25 +0000 (20:55 +0000)]
AMDGPU: Change unreachable into reported error
llvm-svn: 239943
Sanjay Patel [Wed, 17 Jun 2015 20:54:46 +0000 (20:54 +0000)]
remove unnecessary casts; NFC
llvm-svn: 239942
David Majnemer [Wed, 17 Jun 2015 20:53:19 +0000 (20:53 +0000)]
Update clang to take into account the changes to personality fns
llvm-svn: 239941
David Majnemer [Wed, 17 Jun 2015 20:52:32 +0000 (20:52 +0000)]
Move the personality function from LandingPadInst to Function
The personality routine currently lives in the LandingPadInst.
This isn't desirable because:
- All LandingPadInsts in the same function must have the same
personality routine. This means that each LandingPadInst beyond the
first has an operand which produces no additional information.
- There is ongoing work to introduce EH IR constructs other than
LandingPadInst. Moving the personality routine off of any one
particular Instruction and onto the parent function seems a lot better
than have N different places a personality function can sneak onto an
exceptional function.
Differential Revision: http://reviews.llvm.org/D10429
llvm-svn: 239940
Ahmed Bougacha [Wed, 17 Jun 2015 20:44:32 +0000 (20:44 +0000)]
[CodeGenPrepare] Generalize inserted set from truncs to any inst.
It's been used before to avoid infinite loops caused by separate CGP
optimizations undoing one another. We found one more such issue
caused by r238054. To avoid it, generalize the "InsertedTruncs"
set to any inst, and use it to avoid touching those again.
llvm-svn: 239938
Rui Ueyama [Wed, 17 Jun 2015 20:40:43 +0000 (20:40 +0000)]
COFF: Create import library files.
On Windows, we have to create a .lib file for each .dll.
When linking against DLLs, the linker doesn't use the DLL files,
but instead read a list of dllexported symbols from corresponding
lib files.
A library file containing descriptors of a DLL is called an
import library file.
lib.exe has a feature to create an import library file from a
module-definition file. In this patch, we create a module-definition
file and pass that to lib.exe.
We eventually want to create an import library file by ourselves
to eliminate dependency to lib.exe. For now, we just use the MSVC
tool.
llvm-svn: 239937
Richard Smith [Wed, 17 Jun 2015 20:39:41 +0000 (20:39 +0000)]
[modules] If we merge a template, also track that its parameters are merged so
that we know when its default arguments should be visible.
llvm-svn: 239936
Colin LeMahieu [Wed, 17 Jun 2015 20:29:33 +0000 (20:29 +0000)]
[Hexagon] Adding a number of other tests for min/max instructions and loading i1s.
llvm-svn: 239935
Richard Smith [Wed, 17 Jun 2015 20:16:32 +0000 (20:16 +0000)]
[modules] Improve diagnostic for a template-id that's invalid because a default
argument is not visible.
llvm-svn: 239934
Rafael Espindola [Wed, 17 Jun 2015 20:08:20 +0000 (20:08 +0000)]
Move IsUsedInReloc from MCSymbolELF to MCSymbol.
There is a free bit is MCSymbol and MachO needs the same information.
llvm-svn: 239933
Ranjeet Singh [Wed, 17 Jun 2015 19:56:30 +0000 (19:56 +0000)]
[ARM] Replace hard coded metadata arguments in tests with a regex.
Differential Revision: http://reviews.llvm.org/D10507
llvm-svn: 239932
Rui Ueyama [Wed, 17 Jun 2015 19:28:01 +0000 (19:28 +0000)]
COFF: Fix a test which was failing with debug build.
llvm-svn: 239931
Rui Ueyama [Wed, 17 Jun 2015 19:19:25 +0000 (19:19 +0000)]
COFF: Support module-definition files.
Module-definition files (.def files) are yet another way to
specify parameters to the linker. You can write a list of dllexported
symbols in module-definition files instead of using /export command
line option. It also supports a few more directives.
The parser code is taken from lib/Driver/WinLinkModuleDef.cpp
with the following modifications.
- variable names are updated to comply with the LLVM coding style.
- Instead of returning parsing results as "directive" objects,
it updates Config object directly.
llvm-svn: 239929
Peter Collingbourne [Wed, 17 Jun 2015 19:08:05 +0000 (19:08 +0000)]
CodeGen: Factor out some of the bitset entry creation code. NFC.
llvm-svn: 239927
Eric Christopher [Wed, 17 Jun 2015 18:42:07 +0000 (18:42 +0000)]
Move xtest to its own file to match the gcc header organization.
llvm-svn: 239926
Eric Christopher [Wed, 17 Jun 2015 18:42:03 +0000 (18:42 +0000)]
Update comments on HLE, RTM, and ADX support for intrinsics.
llvm-svn: 239925
Pavel Labath [Wed, 17 Jun 2015 18:38:49 +0000 (18:38 +0000)]
[NativeProcessLinux] Use fast memory reads, if the system supports it
Summary:
Memory reads using the ptrace API need to be executed on a designated thread
and in 4-byte increments. The process_vm_read syscall has no such requirements
and it is about 50 times faster. This patch makes lldb-server use the faster
API if the target kernel supports it. Kernel support for this feature is
determined at runtime. Using process_vm_writev in the same manner is more
complicated since this syscall (unlike ptrace) respects page protection settings
and so it cannot be used to set a breakpoint, since code pages are typically
read-only. However, memory writes are not currently a performance bottleneck as
they happen much more rarely.
Test Plan: all tests continue to pass
Reviewers: ovyalov, vharron
Subscribers: tberghammer, lldb-commits
Differential Revision: http://reviews.llvm.org/D10488
llvm-svn: 239924
Peter Collingbourne [Wed, 17 Jun 2015 18:31:02 +0000 (18:31 +0000)]
LowerBitSets: Do not assign names to aliases of unnamed bitset element objects.
The restriction on unnamed aliases was removed in r239921. Mostly reverts
r239590, but we keep the test.
llvm-svn: 239923
Rafael Espindola [Wed, 17 Jun 2015 17:53:31 +0000 (17:53 +0000)]
Allow aliases to be unnamed.
If globals can be unnamed, there is no reason for aliases to be different.
The restriction was there since the original implementation in r36435. I
can only guess it was there because of the old bison parser for the old
alias syntax.
llvm-svn: 239921
Reid Kleckner [Wed, 17 Jun 2015 17:47:30 +0000 (17:47 +0000)]
[fixit] Use overwriteChangedFiles() to deal with Windows mapped files
Fixes one instance of PR17960.
llvm-svn: 239920
Rafael Espindola [Wed, 17 Jun 2015 17:33:37 +0000 (17:33 +0000)]
Use a range loop. NFC.
llvm-svn: 239919
Ed Maste [Wed, 17 Jun 2015 17:29:56 +0000 (17:29 +0000)]
Correct type in for loop to remove signedness warning
llvm-svn: 239918
Colin LeMahieu [Wed, 17 Jun 2015 17:19:05 +0000 (17:19 +0000)]
[Hexagon] Adding some compare tests, fixing existing XFAILed tests, and removing mcpu=hexagonv4 since that's the minimum version anyway.
llvm-svn: 239917
Sanjay Patel [Wed, 17 Jun 2015 16:34:48 +0000 (16:34 +0000)]
fix typos in comments; NFC
llvm-svn: 239916
Diego Novillo [Wed, 17 Jun 2015 16:28:22 +0000 (16:28 +0000)]
Add documentation for new backedge mass propagation in irregular loops.
Tweak test cases and rename headerIndexFor -> getHeaderIndex.
llvm-svn: 239915
Rafael Espindola [Wed, 17 Jun 2015 16:26:47 +0000 (16:26 +0000)]
Use named temporaries for directional labels.
Directional labels can show up in symbol tables (and we have a llvm-mc test for
that). Given that, we need to make sure they are named.
With that out of the way, use setUseNamesOnTempLabels in llvm-mc so that it
too benefits from the memory saving.
llvm-svn: 239914
Benjamin Kramer [Wed, 17 Jun 2015 16:02:56 +0000 (16:02 +0000)]
[ArchiveWriter] Use EndianStream. No functional change intended.
llvm-svn: 239913
Jonathan Peyton [Wed, 17 Jun 2015 15:43:34 +0000 (15:43 +0000)]
Have CMake error out on Windows when user wants OMPT support
Currently, OMPT support requires the weak attribute which isn't supported
on Windows. This patch has CMake error out when LIBOMP_OMPT_SUPPORT=true
and the users is building on Windows.
http://lists.cs.uiuc.edu/pipermail/openmp-dev/2015-June/000692.html
Patch by Jonas Hahnfeld
llvm-svn: 239912
Benjamin Kramer [Wed, 17 Jun 2015 15:14:35 +0000 (15:14 +0000)]
[MC/Dwarf] Encode DW_CFA_advance_loc in target endianess.
This matches GNU as output.
llvm-svn: 239911
Toma Tabacu [Wed, 17 Jun 2015 14:31:51 +0000 (14:31 +0000)]
[mips] [IAS] Add support for expanding LASym with a source register operand.
Reviewers: dsanders
Reviewed By: dsanders
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D9348
llvm-svn: 239910
Daniel Marjamaki [Wed, 17 Jun 2015 14:19:35 +0000 (14:19 +0000)]
clang-tidy: Add checker that warn when macro argument with side effects is repeated in the macro
llvm-svn: 239909
Adhemerval Zanella [Wed, 17 Jun 2015 14:00:12 +0000 (14:00 +0000)]
[ELF/x86_64] Fix initial-exec TLS access
Current approach for initial-exec in ELF/x86_64 is to create a GOT entry
and change the relocation to R_X86_64_PC32 to be handled as a GOT offfset.
However there are two issues with this approach: 1. the R_X86_64_PC32 is
not really required since the GOT relocation will be handle dynamically and
2. the TLS symbols are not being exported externally and then correct
realocation are not being applied.
This patch fixes the R_X86_64_GOTTPOFF handling by just emitting a
R_X86_64_TPOFF64 dynamically one; it also sets R_X86_64_TPOFF64 to be
handled by runtime one. For second part, the patches uses a similar
strategy used for aarch64, by reimplementing buildDynamicSymbolTable
from X86_64ExecutableWriter and adding the TLS symbols in the dynamic
symbol table.
Some tests had to be adjusted due the now missing R_X86_64_PC32 relocation.
With this test the simple testcase:
* t1.c:
__thread int t0;
__thread int t1;
__thread int t2;
__thread int t3;
* t0.c:
extern __thread int t0;
extern __thread int t1;
extern __thread int t2;
extern __thread int t3;
__thread int t4;
__thread int t5;
__thread int t6;
__thread int t7;
int main ()
{
t0 = 1;
t1 = 2;
t2 = 3;
t3 = 4;
t4 = 5;
t5 = 6;
t6 = 7;
t7 = 8;
printf ("%i %i %i %i\n", t0, t1, t2, t3);
printf ("%i %i %i %i\n", t4, t5, t6, t7);
return 0;
}
Shows correct output for x86_64.
llvm-svn: 239908
James Y Knight [Wed, 17 Jun 2015 13:53:12 +0000 (13:53 +0000)]
Tweak wording of alignment static_assert messages.
llvm-svn: 239907
Adhemerval Zanella [Wed, 17 Jun 2015 13:46:07 +0000 (13:46 +0000)]
[ELF] Fix wrong TBSS size
This patch fixes the wrong .tbss segment size generated for cases where
multiple modules have non initialized threads variables. For instance:
* t0.c
__thread int x0;
__thread int x1;
__thread int x2;
extern __thread int e0;
extern __thread int e1;
extern __thread int e2;
extern __thread int e3;
int foo0 ()
{
return x0;
}
int main ()
{
return x0;
}
* t1.c
__thread int e0;
__thread int e1;
__thread int e2;
__thread int e3;
lld is generating (for aarch64):
[14] .tbss NOBITS
0000000000401000 00001000
0000000000000010 0000000000000000 WAT 0 0 4
Where is just taking in consideration the largest tbss segment, not all
from all objects. ld generates a correct output:
[17] .tbss NOBITS
0000000000410dec 00000dec
000000000000001c 0000000000000000 WAT 0 0 4
This issue is at 'lib/ReaderWriter/ELF/SegmentChunks.cpp' where
Segment<ELFT>::assignVirtualAddress is setting wrong slice values, not taking care
of although tbss segments file size does noy play role in other segment virtual
address placement, its size should still be considered.
llvm-svn: 239906
Toma Tabacu [Wed, 17 Jun 2015 13:20:24 +0000 (13:20 +0000)]
[mips] [IAS] Add support for the B{L,G}{T,E}(U) branch pseudo-instructions.
Summary:
This does not include support for the immediate variants of these pseudo-instructions.
Fixes llvm.org/PR20968.
Reviewers: dsanders
Reviewed By: dsanders
Subscribers: seanbruno, emaste, llvm-commits
Differential Revision: http://reviews.llvm.org/D8537
llvm-svn: 239905
Alexander Kornienko [Wed, 17 Jun 2015 13:11:37 +0000 (13:11 +0000)]
[clang-tidy] Move user-defined matches to unnamed namespaces to prevent ODR violations.
llvm-svn: 239904
Daniel Jasper [Wed, 17 Jun 2015 13:08:06 +0000 (13:08 +0000)]
clang-format: clang-format (NFC)
llvm-svn: 239903
Alexander Kornienko [Wed, 17 Jun 2015 12:49:00 +0000 (12:49 +0000)]
Qualify all types used in AST matcher macros.
Summary:
Qualify all types used in AST matcher macros. This makes it possible to
put AST matchers in user code into a namespace other than clang::ast_matchers
and this way prevent ODR violations that could happen when a matcher with the
same name is defined in multiple translation units. Updated comments
accordingly.
Reviewers: djasper, klimek
Reviewed By: djasper, klimek
Subscribers: klimek, cfe-commits
Differential Revision: http://reviews.llvm.org/D10501
llvm-svn: 239902
Toma Tabacu [Wed, 17 Jun 2015 12:30:37 +0000 (12:30 +0000)]
[mips] [IAS] Fix LA with relative label operands.
Summary:
Call MCSymbolRefExpr::create() with a MCSymbol* argument, not with a StringRef
of the Symbol's name, in order to avoid creating invalid temporary symbols for
relative labels (e.g. {$,.L}tmp00, {$,.L}tmp10 etc.).
Reviewers: dsanders
Reviewed By: dsanders
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D10498
llvm-svn: 239901
Daniel Jasper [Wed, 17 Jun 2015 12:23:15 +0000 (12:23 +0000)]
clang-format: Don't generate unnecessary replacements for \r\n line endings.
Patch by Mathieu Champlon. Thank you.
llvm-svn: 239900
Toma Tabacu [Wed, 17 Jun 2015 11:46:37 +0000 (11:46 +0000)]
[mips] [IAS] Add test for SW with relative label operands. NFC.
Reviewers: dsanders
Reviewed By: dsanders
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D10497
llvm-svn: 239899
Alexander Kornienko [Wed, 17 Jun 2015 11:31:18 +0000 (11:31 +0000)]
Remove empty directories left after r239657
llvm-svn: 239898
Toma Tabacu [Wed, 17 Jun 2015 10:43:45 +0000 (10:43 +0000)]
[mips] [IAS] Fix LW with relative label operands.
Summary:
Previously, MCSymbolRefExpr::create() was called with a StringRef of the symbol
name, which it would then search for in the Symbols StringMap (from MCContext).
However, relative labels (which are temporary symbols) are apparently not stored
in the Symbols StringMap, so we end up creating a new {$,.L}tmp symbol
({$,.L}tmp00, {$,.L}tmp10 etc.) each time we create an MCSymbolRefExpr by
passing in the symbol name as a StringRef.
Fortunately, there is a version of MCSymbolRefExpr::create() which takes an
MCSymbol* and we already have an MCSymbol* at that point, so we can just pass
that in instead of the StringRef.
I also removed the local StringRef calls to MCSymbolRefExpr::create() from
expandMemInst(), as those cases can be handled by evaluateRelocExpr() anyway.
Reviewers: dsanders
Reviewed By: dsanders
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D9938
llvm-svn: 239897
NAKAMURA Takumi [Wed, 17 Jun 2015 10:40:51 +0000 (10:40 +0000)]
Fix clang/test/Headers/x86intrin-2.c. _readfsbase_u32() is provided in not i686 but x86-64.
llvm-svn: 239896
Alexey Bataev [Wed, 17 Jun 2015 09:51:18 +0000 (09:51 +0000)]
[OPENMP] Fixed failed test for reduction clause in simd constructs.
llvm-svn: 239895
Daniel Jasper [Wed, 17 Jun 2015 09:44:07 +0000 (09:44 +0000)]
clang-format: [JS] Don't put top-level typescript enums on a single line.
This makes this consistent with non-typescript enums.
Also shuffle the language-dependent stuff in mustBreakBefore to a
single location.
Patch initiated by Martin Probst.
llvm-svn: 239894
Daniel Jasper [Wed, 17 Jun 2015 09:44:02 +0000 (09:44 +0000)]
clang-format: [JS] Fix typescript enum formatting.
Patch by Martin Probst.
Before:
enum {
A,
B
} var x = 1;
After:
enum {
A,
B
}
var x = 1;
llvm-svn: 239893
Daniel Jasper [Wed, 17 Jun 2015 09:43:56 +0000 (09:43 +0000)]
clang-format: NFC. Add a function to test whether an annotated line
starts with a given sequence of tokens. Only the one-token version is
used yet, but other usages are coming up momentarily.
llvm-svn: 239892
Chandler Carruth [Wed, 17 Jun 2015 08:29:32 +0000 (08:29 +0000)]
[PM/AA] Update Polly for r239886 which removed UnknownSize from
AliasAnalysis by getting it from its newly canonical home of
MemoryLocation::UnknownSize.
llvm-svn: 239890
Alexey Bataev [Wed, 17 Jun 2015 07:45:51 +0000 (07:45 +0000)]
[OPENMP] Code reformatting for omp simd codegen, NFC.
llvm-svn: 239889
Igor Breger [Wed, 17 Jun 2015 07:23:57 +0000 (07:23 +0000)]
AVX-512: cvtusi2ss/d intrinsics.
Change builtin function name and signature ( add third parameter - rounding mode ).
Added tests for intrinsics.
Differential Revision: http://reviews.llvm.org/D10473
llvm-svn: 239888
Chandler Carruth [Wed, 17 Jun 2015 07:21:41 +0000 (07:21 +0000)]
[PM/AA] Suffix lots of member variables that directly use enumeration
names for counts with the word 'Count' to make them less ambiguous.
This will be an actual error if we use unscoped enums for any of these,
and generally this seems much clearer to read.
Also, use clang-format to normalize the formatting of this code which
seems to have been needlessly odd.
No functionality changed here.
llvm-svn: 239887
Chandler Carruth [Wed, 17 Jun 2015 07:21:38 +0000 (07:21 +0000)]
[PM/AA] Remove the UnknownSize static member from AliasAnalysis.
This is now living in MemoryLocation, which is what it pertains to. It
is also an enum there rather than a static data member which is left
never defined.
llvm-svn: 239886
Chandler Carruth [Wed, 17 Jun 2015 07:18:54 +0000 (07:18 +0000)]
[PM/AA] Remove the Location typedef from the AliasAnalysis class now
that it is its own entity in the form of MemoryLocation, and update all
the callers.
This is an entirely mechanical change. References to "Location" within
AA subclases become "MemoryLocation", and elsewhere
"AliasAnalysis::Location" becomes "MemoryLocation". Hope that helps
out-of-tree folks update.
llvm-svn: 239885
Chandler Carruth [Wed, 17 Jun 2015 07:12:40 +0000 (07:12 +0000)]
[PM/AA] Split the location computation out of getArgLocation so the
virtual interface on AliasAnalysis only deals with ModRef information.
This interface was both computing memory locations by using TLI and
other tricks to estimate the size of memory referenced by an operand,
and computing ModRef information through similar investigations. This
change narrows the scope of the virtual interface on AliasAnalysis
slightly.
Note that all of this code could live in BasicAA, and be done with
a single investigation of the argument, if it weren't for the fact that
the generic code in AliasAnalysis::getModRefBehavior for a callsite
calls into the virtual aspect of (now) getArgModRefInfo. But this
patch's arrangement seems a not terrible way to go for now.
The other interesting wrinkle is how we could reasonably extend LLVM
with support for custom memory location sizes and mod/ref behavior for
library routines. After discussions with Hal on the review, the
conclusion is that this would be best done by fleshing out the much
desired support for extensions to TLI, and support these types of
queries in that interface where we would likely be doing other library
API recognition and analysis.
Differential Revision: http://reviews.llvm.org/D10259
llvm-svn: 239884
Eric Christopher [Wed, 17 Jun 2015 07:09:32 +0000 (07:09 +0000)]
Update the intel intrinsic headers to use the target attribute support.
This involved removing the conditional inclusion and replacing them
with target attributes matching the original conditional inclusion
and checks. The testcase update removes the macro checks for each
file and replaces them with usage of the __target__ attribute, e.g.:
int __attribute__((__target__(("sse3")))) foo(int a) {
_mm_mwait(0, 0);
return 4;
}
This usage does require the enclosing function have the requisite
__target__ attribute for inlining and code generation - also for
any macro intrinsic uses in the enclosing function. There's no change
for existing uses of the intrinsic headers.
llvm-svn: 239883
Eric Christopher [Wed, 17 Jun 2015 07:09:20 +0000 (07:09 +0000)]
Use a define for per-file function attributes for the Intel intrinsic headers.
This is a precursor to changing them to use the new target attribute
code.
llvm-svn: 239882
Alexey Bataev [Wed, 17 Jun 2015 06:21:39 +0000 (06:21 +0000)]
[OPENMP] Supported reduction clause in omp simd construct.
The following code is generated for reduction clause within 'omp simd' loop construct:
#pragma omp simd reduction(op:var)
for (...)
<body>
alloca priv_var
priv_var = <initial reduction value>;
<loop_start>:
<body> // references to original 'var' are replaced by 'priv_var'
<loop_end>:
var op= priv_var;
llvm-svn: 239881
Matthias Braun [Wed, 17 Jun 2015 04:02:32 +0000 (04:02 +0000)]
Revert "AArch64: Use CMP;CCMP sequences for and/or/setcc trees."
The patch triggers a miscompile on SPEC 2006 403.gcc with the (ref)
200.i and scilab.i inputs. I opened PR23866 to track analysis of this.
This reverts commit r238793.
llvm-svn: 239880
Saleem Abdulrasool [Wed, 17 Jun 2015 03:54:21 +0000 (03:54 +0000)]
parser: wordsmith diagnostic message
Address post-commit commit about the wording of the warning.
llvm-svn: 239879
Saleem Abdulrasool [Wed, 17 Jun 2015 03:54:19 +0000 (03:54 +0000)]
Basic: tweak whitespace in Attr.td
Separate two class definitions that had been merged into a single line. NFC.
llvm-svn: 239878
Rafael Espindola [Wed, 17 Jun 2015 03:13:26 +0000 (03:13 +0000)]
Try to fix the MSVC build.
llvm-svn: 239877
Colin LeMahieu [Wed, 17 Jun 2015 03:06:16 +0000 (03:06 +0000)]
[Hexagon] Adding MC ELF streamer and updating addend relocation test which shows correct ELF symbol.
llvm-svn: 239876
Sean Silva [Wed, 17 Jun 2015 02:21:35 +0000 (02:21 +0000)]
[.gitignore] ignore vim swap files harder
This matches the patterns for vim swap files in llvm/.gitignore
llvm-svn: 239875
Enrico Granata [Wed, 17 Jun 2015 02:11:48 +0000 (02:11 +0000)]
Revert r239873 - I actually want to think some more about this
llvm-svn: 239874
Enrico Granata [Wed, 17 Jun 2015 02:06:24 +0000 (02:06 +0000)]
Fix an issue where the oneliner printing of variables would ignore custom formatting
Because vector types use their formats in special ways (i.e. children get generated based on them), this change by itself would cause a regression in printing vector types with some custom formats
Work around that issue by special casing vector types out of this format-passdown mode. I believe there is a more general feature to be designed in this space, but until I see more cases of interest, I am going to leave this as a special case
Fixes rdar://
20810062
llvm-svn: 239873
James Y Knight [Wed, 17 Jun 2015 01:21:20 +0000 (01:21 +0000)]
Fix alignment issues in LLVM.
Adds static_asserts to ensure alignment of concatenated objects is
correct, and fixes them where they are not.
Also changes the definition of AlignOf to use constexpr, except on
MSVC, to avoid enum comparison warnings from GCC.
(There's not too much of this in llvm itself, most of the fun is in
clang).
This seems to make LLVM actually work without Bus Error on 32bit
sparc.
Differential Revision: http://reviews.llvm.org/D10271
llvm-svn: 239872
Rafael Espindola [Wed, 17 Jun 2015 01:15:47 +0000 (01:15 +0000)]
Handle forward referenced function when streaming bitcode.
Without this the included unit test would assert in
assert(BasicBlockFwdRefs.empty() && "Unresolved blockaddress fwd references");
llvm-svn: 239871
Rafael Espindola [Wed, 17 Jun 2015 00:40:56 +0000 (00:40 +0000)]
Handle MaterializeAll in getLazyBitcodeModuleImpl. NFC.
This just handles both cases in the same place.
Extracted from a patch by Karl Schimpf.
llvm-svn: 239870
Rui Ueyama [Wed, 17 Jun 2015 00:16:33 +0000 (00:16 +0000)]
COFF: Support creating DLLs.
DLL files are in the same format as executables but they have export tables.
The format of the export table is described in PE/COFF spec section 5.3.
A new class, EdataContents, takes care of creating chunks for export tables.
What we need to do is to parse command line flags for dllexports, and then
instantiate the class to create chunks. For the writer, export table chunks
are opaque data -- it just add chunks to .edata section.
llvm-svn: 239869
Rui Ueyama [Tue, 16 Jun 2015 23:51:58 +0000 (23:51 +0000)]
COFF: Fix tests.
I was accidentally testing not -flavor link2 but -flavor link.
llvm-svn: 239868
Rafael Espindola [Tue, 16 Jun 2015 23:29:49 +0000 (23:29 +0000)]
Use std::unique_ptr to manage the DataStreamer in bitcode parsing.
We were already deleting it, this just makes it explicit.
llvm-svn: 239867
Rafael Espindola [Tue, 16 Jun 2015 23:22:02 +0000 (23:22 +0000)]
Rename and improve emitSectionOffset.
Different object formats represent references from dwarf in different ways.
ELF uses a relocation to the referenced point (except for .dwo) and
COFF/MachO use the offset of the referenced point inside its section.
This patch renames emitSectionOffset because
* It doesn't produce an offset on ELF.
* It changes behavior depending on how DWARF is represented, so adding
dwarf to its name is probably a good thing.
The patch also adds an option to force the use of offsets.That avoids
funny looking code like
if (!UseOffsets)
Asm->emitSectionOffset....
It was correct, but read as if the ! was inverted.
llvm-svn: 239866
Enrico Granata [Tue, 16 Jun 2015 23:20:12 +0000 (23:20 +0000)]
Add a ThreadSafe adapter over llvm::DenseSet
llvm-svn: 239865
Rui Ueyama [Tue, 16 Jun 2015 23:13:00 +0000 (23:13 +0000)]
COFF: Add miscellaneous boolean flags.
llvm-svn: 239864
Greg Clayton [Tue, 16 Jun 2015 23:09:37 +0000 (23:09 +0000)]
Clear the "thread-format" setting after mucking with it.
llvm-svn: 239863
Tyler Nowicki [Tue, 16 Jun 2015 22:59:45 +0000 (22:59 +0000)]
Refactor RecurrenceInstDesc
Moved RecurrenceInstDesc into RecurrenceDescriptor to simplify the namespaces.
llvm-svn: 239862
Dawn Perchik [Tue, 16 Jun 2015 22:53:27 +0000 (22:53 +0000)]
Fix comments (MI).
llvm-svn: 239861
Sanjay Patel [Tue, 16 Jun 2015 22:37:50 +0000 (22:37 +0000)]
Add some tests based on PR21711
These were originally added in r227242,
but that patch was reverted because it
caused a failure on AArch64.
llvm-svn: 239860
Rafael Espindola [Tue, 16 Jun 2015 22:32:44 +0000 (22:32 +0000)]
Update for llvm api change.
llvm-svn: 239859
Rafael Espindola [Tue, 16 Jun 2015 22:27:55 +0000 (22:27 +0000)]
Return a unique_ptr from getLazyBitcodeModule and parseBitcodeFile. NFC.
llvm-svn: 239858
Richard Smith [Tue, 16 Jun 2015 21:57:05 +0000 (21:57 +0000)]
[modules] Fix merging of default template arguments onto friend templates.
Previously we'd complain about redefinition of default arguments when we
instantiated a class with a friend template that inherits its default argument,
because we propagate the default template arguemnt onto the friend when we
reload the AST.
llvm-svn: 239857
Simon Atanasyan [Tue, 16 Jun 2015 21:47:43 +0000 (21:47 +0000)]
[llvm-readobj] Print MIPS .reginfo section content
llvm-svn: 239856