Hal Finkel [Sat, 16 Dec 2017 02:42:18 +0000 (02:42 +0000)]
[PowerPC, AsmParser] Enable the mnemonic spell corrector
r307148 added an assembly mnemonic spelling correction support and enabled it
on ARM. This enables that support on PowerPC as well.
Patch by Dmitry Venikov, thanks!
Differential Revision: https://reviews.llvm.org/D40552
llvm-svn: 320911
Craig Topper [Sat, 16 Dec 2017 02:40:28 +0000 (02:40 +0000)]
[X86] Add 128 and 256-bit VPOPCNTDQ instructions. Adjust some tablegen classes LZCNT/POPCNT.
I think when this instruction was first published it was only for a Knights CPU and thus VLX version was missing.
llvm-svn: 320910
Vitaly Buka [Sat, 16 Dec 2017 02:40:20 +0000 (02:40 +0000)]
[LTO] Update tests for r320905
llvm-svn: 320909
Hal Finkel [Sat, 16 Dec 2017 02:23:22 +0000 (02:23 +0000)]
[VerifyDiagnosticConsumer] support -verify=<prefixes>
This mimics FileCheck's --check-prefixes option.
The default prefix is "expected". That is, "-verify" is equivalent to
"-verify=expected".
The goal is to permit exercising a single test suite source file with different
compiler options producing different sets of diagnostics. While cpp can be
combined with the existing -verify to accomplish the same goal, source is often
easier to maintain when it's not cluttered with preprocessor directives or
duplicate passages of code. For example, this patch also rewrites some existing
clang tests to demonstrate the benefit of this feature.
Patch by Joel E. Denny, thanks!
Differential Revision: https://reviews.llvm.org/D39694
llvm-svn: 320908
Vitaly Buka [Sat, 16 Dec 2017 02:12:35 +0000 (02:12 +0000)]
Remove trailing whitespace
llvm-svn: 320907
Sam Clegg [Sat, 16 Dec 2017 02:10:16 +0000 (02:10 +0000)]
[WebAssembly] Return ArrayRef's rather than const std::vector&
From working on lld I've learned this is generally the
preferred way for several reasons (e.g. more concise, improves
encapsulation).
Differential Revision: https://reviews.llvm.org/D41265
llvm-svn: 320906
Vitaly Buka [Sat, 16 Dec 2017 02:10:00 +0000 (02:10 +0000)]
[LTO] Make processing of combined module more consistent
Summary:
1. Use stream 0 only for combined module. Previously if combined module was not
processes ThinLTO used the stream for own output. However small changes in input,
could trigger combined module and shuffle outputs making life of llvm::LTO harder.
2. Always process combined module and write output to stream 0. Processing empty
combined module is cheap and allows llvm::LTO users to avoid implementing processing
which is already done in llvm::LTO.
Subscribers: mehdi_amini, inglorion, eraman, hiraditya
Differential Revision: https://reviews.llvm.org/D41267
llvm-svn: 320905
Hal Finkel [Sat, 16 Dec 2017 01:40:19 +0000 (01:40 +0000)]
[TextDiagnosticBuffer] Fix diagnostic note emission order
The frontend currently groups diagnostics from the command line according to
diagnostic level, but that places all notes last. Fix that by emitting such
diagnostics in the order they were generated.
Patch by Joel E. Denny, thanks!
Differential Revision: https://reviews.llvm.org/D40995
llvm-svn: 320904
Teresa Johnson [Sat, 16 Dec 2017 01:35:36 +0000 (01:35 +0000)]
Add another missing -enable-import-metadata to test
r320895 modified a test so that it needs -enable-import-metadata which
is false by default for NDEBUG, found another place that needs this
added.
llvm-svn: 320903
Vedant Kumar [Sat, 16 Dec 2017 01:28:25 +0000 (01:28 +0000)]
[CodeGen] Specialize mixed-sign mul-with-overflow (fix PR34920)
This patch introduces a specialized way to lower overflow-checked
multiplications with mixed-sign operands. This fixes link failures and
ICEs on code like this:
void mul(int64_t a, uint64_t b) {
int64_t res;
__builtin_mul_overflow(a, b, &res);
}
The generic checked-binop irgen would use a 65-bit multiplication
intrinsic here, which requires runtime support for _muloti4 (128-bit
multiplication), and therefore fails to link on i386. To get an ICE
on x86_64, change the example to use __int128_t / __uint128_t.
Adding runtime and backend support for 65-bit or 129-bit checked
multiplication on all of our supported targets is infeasible.
This patch solves the problem by using simpler, specialized irgen for
the mixed-sign case.
llvm.org/PR34920, rdar://
34963321
Testing: Apart from check-clang, I compared the output from this fairly
comprehensive test driver using unpatched & patched clangs:
https://gist.github.com/vedantk/
3eb9c88f82e5c32f2e590555b4af5081
Differential Revision: https://reviews.llvm.org/D41149
llvm-svn: 320902
Hal Finkel [Sat, 16 Dec 2017 01:26:25 +0000 (01:26 +0000)]
[SimplifyLibCalls] Inline calls to cabs when it's safe to do so
When unsafe algerbra is allowed calls to cabs(r) can be replaced by:
sqrt(creal(r)*creal(r) + cimag(r)*cimag(r))
Patch by Paul Walker, thanks!
Differential Revision: https://reviews.llvm.org/D40069
llvm-svn: 320901
Hal Finkel [Sat, 16 Dec 2017 01:12:50 +0000 (01:12 +0000)]
[LV] NFC patch for moving VP*Recipe class definitions from LoopVectorize.cpp to VPlan.h
This is a small step forward to move VPlan stuff to where it should belong (i.e., VPlan.*):
1. VP*Recipe classes in LoopVectorize.cpp are moved to VPlan.h.
2. Many of VP*Recipe::print() and execute() definitions are still left in
LoopVectorize.cpp since they refer to things declared in LoopVectorize.cpp. To
be moved to VPlan.cpp at a later time.
3. InterleaveGroup class is moved from anonymous namespace to llvm namespace.
Referencing it in anonymous namespace from VPlan.h ended up in warning.
Patch by Hideki Saito, thanks!
Differential Revision: https://reviews.llvm.org/D41045
llvm-svn: 320900
Teresa Johnson [Sat, 16 Dec 2017 01:00:48 +0000 (01:00 +0000)]
Add -enable-import-metadata to test
r320895 modified a test so that it needs -enable-import-metadata which
is false by default for NDEBUG.
llvm-svn: 320899
Craig Topper [Sat, 16 Dec 2017 00:33:16 +0000 (00:33 +0000)]
[X86] Add back the assert from r320830 that was reverted in r320850
Hopefully r320864 has fixed the offending case that failed the assert.
llvm-svn: 320898
Teresa Johnson [Sat, 16 Dec 2017 00:29:31 +0000 (00:29 +0000)]
Fix NDEBUG build problem in r320895
Fix incorrect placement of #endif causing NDEBUG build failures.
llvm-svn: 320897
Shoaib Meenai [Sat, 16 Dec 2017 00:23:24 +0000 (00:23 +0000)]
[COFF] Clean up debug option handling
/debug and /debug:dwarf are orthogonal. An object file can contain both
CodeView and DWARF debug info, so the combination of /debug:dwarf and
/debug should generate both DWARF and a PDB, rather than /debug:dwarf
always suppressing PDB creation.
/nopdb is now redundant and can be removed. /debug /nopdb was previously
used to support DWARF, but specifying /debug:dwarf is entirely
equivalent to that combination now.
Differential Revision: https://reviews.llvm.org/D41310
llvm-svn: 320896
Teresa Johnson [Sat, 16 Dec 2017 00:18:12 +0000 (00:18 +0000)]
[ThinLTO] Enable importing of aliases as copy of aliasee
Summary:
This implements a missing feature to allow importing of aliases, which
was previously disabled because alias cannot be available_externally.
We instead import an alias as a copy of its aliasee.
Some additional work was required in the IndexBitcodeWriter for the
distributed build case, to ensure that the aliasee has a value id
in the distributed index file (i.e. even when it is not being
imported directly).
This is a performance win in codes that have many aliases, e.g. C++
applications that have many constructor and destructor aliases.
Reviewers: pcc
Subscribers: mehdi_amini, inglorion, eraman, llvm-commits
Differential Revision: https://reviews.llvm.org/D40747
llvm-svn: 320895
Shoaib Meenai [Fri, 15 Dec 2017 23:52:46 +0000 (23:52 +0000)]
[COFF] Update an outdated comment. NFC
This comment dates from when LLD didn't produce actual PDBs, and is very
outdated now.
llvm-svn: 320894
David Blaikie [Fri, 15 Dec 2017 23:52:06 +0000 (23:52 +0000)]
Fix WebAssembly backend for some LLVM API changes
llvm-svn: 320893
Shoaib Meenai [Fri, 15 Dec 2017 23:51:14 +0000 (23:51 +0000)]
[COFF] Simplify hasArgs calls. NFC
We can just pass multiple options to hasArgs (which will check for any
of those options being present) instead of calling it multiple times.
llvm-svn: 320892
Davide Italiano [Fri, 15 Dec 2017 23:27:10 +0000 (23:27 +0000)]
[CMake] darwin-debug is an hard dependency for tests on macOS.
Fixes a few failured on the testsuite with CMake.
llvm-svn: 320891
Quentin Colombet [Fri, 15 Dec 2017 23:24:39 +0000 (23:24 +0000)]
[TableGen][GlobalISel] Make the different Matcher comparable
This opens refactoring opportunities in the match table now that we can
check that two predicates are the same.
NFC.
llvm-svn: 320890
Quentin Colombet [Fri, 15 Dec 2017 23:24:36 +0000 (23:24 +0000)]
[TableGen][GlobalISel] Fix unused variable warning in release mode
Introduced in r320887.
NFC.
llvm-svn: 320889
Paul Robinson [Fri, 15 Dec 2017 23:21:52 +0000 (23:21 +0000)]
Revert "Recommit "[DWARFv5] Dump an MD5 checksum in the line-table header.""
This reverts commit
0afef672f63f0e4e91938656bc73424a8c058bfc.
Still failing at runtime on bots.
llvm-svn: 320888
Quentin Colombet [Fri, 15 Dec 2017 23:07:42 +0000 (23:07 +0000)]
[TableGen][GlobalISel] Have the predicate directly know which data they are dealing with
Prior to this patch, a predicate wouldn't make sense outside of its
rule. Indeed, it was only during emitting a rule that a predicate would
be made aware of the IDs of the data it is checking. Because of that,
predicates could not be moved around or compared between each other.
NFC.
llvm-svn: 320887
Paul Robinson [Fri, 15 Dec 2017 22:57:17 +0000 (22:57 +0000)]
Recommit "[DWARFv5] Dump an MD5 checksum in the line-table header."
Adds missing support for DW_FORM_data16.
Update of r320852, fixing the unittest to use a hand-coded struct
instead of std::array to guarantee data layout.
Differential Revision: https://reviews.llvm.org/D41090
llvm-svn: 320886
Matthias Braun [Fri, 15 Dec 2017 22:53:33 +0000 (22:53 +0000)]
Fix unused variable in non-assert builds
llvm-svn: 320885
Matthias Braun [Fri, 15 Dec 2017 22:22:58 +0000 (22:22 +0000)]
MachineFunction: Return reference from getFunction(); NFC
The Function can never be nullptr so we can return a reference.
llvm-svn: 320884
Davide Italiano [Fri, 15 Dec 2017 22:22:51 +0000 (22:22 +0000)]
[MacOSX/Queues] Relax an overly aggressive assertion in a test.
"Default" is a valid QoS for a thread on older versions of macOS,
like the one installed in the bot.
Thanks to Jason Molenda for helping me figuring out the problem.
<rdar://problem/
28346273>
llvm-svn: 320883
Matthias Braun [Fri, 15 Dec 2017 22:22:46 +0000 (22:22 +0000)]
MachineFunction: Slight refactoring; NFC
Slight cleanup/refactor in preparation for upcoming commit.
llvm-svn: 320882
Matthias Braun [Fri, 15 Dec 2017 22:22:42 +0000 (22:22 +0000)]
MachineModuleInfo: Remove unused function; NFC
Remove the unused setModule() function; it would be dangerous if someone
actually used it as it wouldn't reset/recompute various other module
related data.
llvm-svn: 320881
Sam Clegg [Fri, 15 Dec 2017 22:17:15 +0000 (22:17 +0000)]
[WebAssembly] Don't include lazy symbols in import table
This bug was introduced in: https://reviews.llvm.org/D41304.
Add a test for this case.
Differential Revision: https://reviews.llvm.org/D41309
llvm-svn: 320872
Galina Kistanova [Fri, 15 Dec 2017 22:15:29 +0000 (22:15 +0000)]
Fixed the gcc 'enumeral and non-enumeral type in conditional expression [-Werror=extra]' warning introduced by r320750
llvm-svn: 320868
Krzysztof Parzyszek [Fri, 15 Dec 2017 21:34:05 +0000 (21:34 +0000)]
[Hexagon] Remove recursion in visitUsesOf, replace with use queue
This is primarily to reduce stack usage, but ordering the use queue
according to the position in the code (earlier instructions visited
before later ones) reduces the number of unnecessary bottoms due to
visiting instructions out of order, e.g.
%reg1 = copy %reg0
%reg2 = copy %reg0
%reg3 = and %reg1, %reg2
Here, reg3 should be known to be same as reg0-2, but if reg3 is
evaluated after reg1 is updated, but before reg2 is updated, the two
inputs to the and will appear different, causing reg3 to become
bottom.
llvm-svn: 320866
Krzysztof Parzyszek [Fri, 15 Dec 2017 21:23:12 +0000 (21:23 +0000)]
[Hexagon] Handle concat_vectors of all allowed HVX types
llvm-svn: 320865
Craig Topper [Fri, 15 Dec 2017 21:18:06 +0000 (21:18 +0000)]
[X86] Use AND32ri8 instead of AND64ri8 in Asan code in EmitCallAsanReport for 32-bit mode.
This seemed to work due to a quirk in the X86 MC encoder that didn't emit a REX byte that the AND64ri8 implies when in 32-bit mode. This made the encoding the same as AND32ri8. I tried to add an assert to catch the dropped REX prefix that caught this.
llvm-svn: 320864
Craig Topper [Fri, 15 Dec 2017 21:18:05 +0000 (21:18 +0000)]
[X86] In LowerVectorCTPOP use ISD::ZERO_EXTEND/ISD::TRUNCATE instead of the target specific nodes.
The target independent nodes will get legalized to the target specific nodes by their own legalization process. Someday I'd like to stop using a target specific for zero extends and truncates of legal types so the less places we reference the target specific opcode the better.
llvm-svn: 320863
Craig Topper [Fri, 15 Dec 2017 20:57:18 +0000 (20:57 +0000)]
[X86] Remove unnecessary TODO.
When I wrote it I thought we were missing a potential optimization for KNL. But investigating further shows that for KNL we still do the optimal thing by widening to v4f32 and then using special isel patterns to widen again to zmm a register.
llvm-svn: 320862
Martin Storsjo [Fri, 15 Dec 2017 20:53:10 +0000 (20:53 +0000)]
[MinGW] Ignore the --no-seh flag
The COFF linker automatically sets the IMAGE_DLL_CHARACTERISTICS_NO_SEH
when suitable, similarly to link.exe.
Differential Revision: https://reviews.llvm.org/D41275
llvm-svn: 320861
Martin Storsjo [Fri, 15 Dec 2017 20:53:03 +0000 (20:53 +0000)]
[COFF] Set the IMAGE_DLL_CHARACTERISTICS_NO_SEH flag automatically
This seems to match how link.exe sets it.
Differential Revision: https://reviews.llvm.org/D41252
llvm-svn: 320860
Vitaly Buka [Fri, 15 Dec 2017 20:50:25 +0000 (20:50 +0000)]
[LTO] Remove unused RegularLTOState::HasModule
llvm-svn: 320859
Jun Bum Lim [Fri, 15 Dec 2017 20:33:24 +0000 (20:33 +0000)]
Re-commit : [LICM] Allow sinking when foldable in loop
This recommits r320823 reverted due to the test failure in sink-foldable.ll and
an unused variable. Added "REQUIRES: aarch64-registered-target" in the test
and removed unused variable.
Original commit message:
Continue trying to sink an instruction if its users in the loop is foldable.
This will allow the instruction to be folded in the loop by decoupling it from
the user outside of the loop.
Reviewers: hfinkel, majnemer, davidxl, efriedma, danielcdh, bmakam, mcrosier
Reviewed By: hfinkel
Subscribers: javed.absar, bmakam, mcrosier, llvm-commits
Differential Revision: https://reviews.llvm.org/D37076
llvm-svn: 320858
Paul Robinson [Fri, 15 Dec 2017 20:29:25 +0000 (20:29 +0000)]
Revert "[DWARFv5] Dump an MD5 checksum in the line-table header."
Unit test fails on some bots.
llvm-svn: 320857
Jake Ehrlich [Fri, 15 Dec 2017 20:17:55 +0000 (20:17 +0000)]
[llvm-objcopy] Reformat everything using clang-format -i
Overtime some non-clang formatted code has creeped into llvm-objcopy. This
patch fixes all of that.
Differential Revision: https://reviews.llvm.org/D41262
llvm-svn: 320856
Krzysztof Parzyszek [Fri, 15 Dec 2017 20:13:57 +0000 (20:13 +0000)]
[Hexagon] Fix operand-swapping PatFrag for atomic stores
PatFrag now has the atomicity information stored as bit fields. They
need to be copied to the new PatFrag.
llvm-svn: 320855
Alex Lorenz [Fri, 15 Dec 2017 20:07:53 +0000 (20:07 +0000)]
__is_target_environment: Check the environment after parsing it
This ensures that target triples with environment versions can still work with
__is_target_environment.
llvm-svn: 320854
Alex Lorenz [Fri, 15 Dec 2017 19:58:38 +0000 (19:58 +0000)]
__is_target_arch: Check the arch and subarch instead of the arch name
This ensures that when compiling for "arm64" __is_target_arch will succeed for
both "arm64" and "aarch64".
Thanks to Bob Wilson who pointed this out!
llvm-svn: 320853
Paul Robinson [Fri, 15 Dec 2017 19:52:34 +0000 (19:52 +0000)]
[DWARFv5] Dump an MD5 checksum in the line-table header.
Adds missing support for DW_FORM_data16.
Differential Revision: https://reviews.llvm.org/D41090
llvm-svn: 320852
Rui Ueyama [Fri, 15 Dec 2017 19:39:59 +0000 (19:39 +0000)]
Do not add .dynamic entries that don't have to be added early in the ctor. NFC.
We add dynamic section entries both in the ctor of the class and
DynamicSection::finalizeContents(). Some entries need to be added early
in the ctor because they add strings to .dynstr. Other entries were
intended to be added in finalizeContents(). However, some entries are
added in the ctor even though they don't add strings. This patch
fix the issue.
llvm-svn: 320851
Craig Topper [Fri, 15 Dec 2017 19:38:14 +0000 (19:38 +0000)]
[X86] Remove assert in X86MCCodeEmitter.cpp that was added in r320830.
It seems to be failing real code which is concerning, but we were silently getting away with it. I'll investigate further.
llvm-svn: 320850
Craig Topper [Fri, 15 Dec 2017 19:35:22 +0000 (19:35 +0000)]
[SelectionDAG][X86] Fix insert_vector_elt lowering for v32i1/v64i1 with non-constant index
Summary:
Currently we don't handle v32i1/v64i1 insert_vector_elt correctly as we fail to look at the number of elements closely and assume it can only be v16i1 or v8i1.
We also can't type legalize v64i1 insert_vector_elt correctly on KNL due to the type not being byte addressable as required by the legalizing through memory accesses path requires.
For the first issue, the patch now tries to pick a 512-bit register with the correct number of elements and promotes to that.
For the second issue, we now extend the vector to a byte addressable type, do the stores to memory, load the two halves, and then truncate the halves back to the original type. Technically since we changed the type, we may not need two loads, but actually checking that is more work and for the v64i1 case we do need them.
Reviewers: RKSimon, delena, spatel, zvi
Reviewed By: RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D40942
llvm-svn: 320849
Sean Fertile [Fri, 15 Dec 2017 19:29:12 +0000 (19:29 +0000)]
[Memcpy Loop Lowering] Insert loop BB inbetween the split BB.
The original memcpy expansion inserted the loop basic block inbetween
the 2 new basic blocks created by splitting the original block the memcpy
call was in. This commit makes the new memcpy expansion do the same to keep the
layout of the IR matching between the old and new implementations.
Differential Review: https://reviews.llvm.org/D41197
llvm-svn: 320848
Sam Clegg [Fri, 15 Dec 2017 19:23:49 +0000 (19:23 +0000)]
[WebAssembly] Base imports on Symtab. NFC.
Since imports are undefined symbols we know we can
find all of them my looking at the symbol table alone.
(i.e. imports cannot be have local binding).
This will be strictly faster and also allows us
to to remove a method from Symbol class
Differential Revision: https://reviews.llvm.org/D41304
llvm-svn: 320847
Craig Topper [Fri, 15 Dec 2017 19:01:51 +0000 (19:01 +0000)]
[X86] Add 'Requires<[In64BitMode]>' to a bunch of instructions that only have memory and immediate operands.
The asm parser wasn't preventing these from being accepted in 32-bit mode. Instructions that use a GR64 register are protected by the parser rejecting the register in 32-bit mode.
llvm-svn: 320846
Craig Topper [Fri, 15 Dec 2017 19:01:50 +0000 (19:01 +0000)]
[X86] Change BNDLDX to use anymem instead of i64mem for itsmemory operand.
This instruction doesn't access memory. It juse use a similar looking memory encoding. Don't require Intel syntax to put "qword ptr" in front of it.
llvm-svn: 320845
Craig Topper [Fri, 15 Dec 2017 19:01:49 +0000 (19:01 +0000)]
[X86] Remove the 'Requires' In64BitMode/Not64BitMode from the LWP instructions.
These aren't doing anything due to a top level "let Predicates =". I think the GR32/GR64 register class protects these anyway.
llvm-svn: 320844
Craig Topper [Fri, 15 Dec 2017 19:01:48 +0000 (19:01 +0000)]
[X86] Remove the 'Requires<[In64BitMode]>' from SHSTK instructions.
This has no effect due to a top level "let Predicates =" around the instructions. But its also not required because the GR64 usage in the instruction guarantees it can never match.
llvm-svn: 320843
Sanjay Patel [Fri, 15 Dec 2017 18:54:29 +0000 (18:54 +0000)]
[TargetLibraryInfo] fix documentation comment; NFC
llvm-svn: 320842
Julie Hockett [Fri, 15 Dec 2017 18:54:28 +0000 (18:54 +0000)]
[clang-tidy] Adding Fuchsia checker for virtual inheritance
Adds a check to the Fuchsia module to warn if classes are defined
with virtual inheritance.
See https://fuchsia.googlesource.com/zircon/+/master/docs/cxx.md for
reference.
Differential Revision: https://reviews.llvm.org/D40813
llvm-svn: 320841
Sanjay Patel [Fri, 15 Dec 2017 18:34:45 +0000 (18:34 +0000)]
[CodeGen] fix documentation comments; NFC
llvm-svn: 320840
Evandro Menezes [Fri, 15 Dec 2017 18:26:54 +0000 (18:26 +0000)]
[AArch64] Fix typo in the ASIMD instruction optimization pass
Fix typo in the representative instruction replacement.
Also, fix formatting and reword some comments.
llvm-svn: 320839
Sanjay Patel [Fri, 15 Dec 2017 18:25:13 +0000 (18:25 +0000)]
fix typo in comment and remove inaccurate comment; NFC
llvm-svn: 320838
Andrew V. Tischenko [Fri, 15 Dec 2017 18:13:05 +0000 (18:13 +0000)]
Fix for bug PR35549 - Repeated schedule comments.
Differential Revision: https://reviews.llvm.org/D40960
llvm-svn: 320837
Jun Bum Lim [Fri, 15 Dec 2017 18:12:49 +0000 (18:12 +0000)]
Revert "Re-commit : [LICM] Allow sinking when foldable in loop"
This reverts commit r320833.
llvm-svn: 320836
Sanjay Patel [Fri, 15 Dec 2017 18:09:33 +0000 (18:09 +0000)]
[CodeGen] fix documentation comments; NFC
llvm-svn: 320835
Jun Bum Lim [Fri, 15 Dec 2017 17:58:59 +0000 (17:58 +0000)]
Re-commit : [LICM] Allow sinking when foldable in loop
This recommit r320823 after fixing a test failure.
Original commit message:
Continue trying to sink an instruction if its users in the loop is foldable.
This will allow the instruction to be folded in the loop by decoupling it from
the user outside of the loop.
Reviewers: hfinkel, majnemer, davidxl, efriedma, danielcdh, bmakam, mcrosier
Reviewed By: hfinkel
Subscribers: javed.absar, bmakam, mcrosier, llvm-commits
Differential Revision: https://reviews.llvm.org/D37076
llvm-svn: 320833
Michael Trent [Fri, 15 Dec 2017 17:57:40 +0000 (17:57 +0000)]
Updated llvm-objdump to display local relocations in Mach-O binaries
Summary:
llvm-objdump's Mach-O parser was updated in r306037 to display external
relocations for MH_KEXT_BUNDLE file types. This change extends the Macho-O
parser to display local relocations for MH_PRELOAD files. When used with
the -macho option relocations will be displayed in a historical format.
All tests are passing for llvm, clang, and lld. llvm-objdump builds without
compiler warnings.
rdar://
35778019
Reviewers: enderby
Reviewed By: enderby
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D41199
llvm-svn: 320832
Filipe Cabecinhas [Fri, 15 Dec 2017 17:30:50 +0000 (17:30 +0000)]
[ubsan-minimal] Add a path for non-POSIX (and bare-metal) use of the library
Summary:
Hook on -DKERNEL_USE (which is also used in lib/builtins) to not import
strlen and not rely on write() being implemented with the stderr on fd 2.
With this, the only requirements to use this library are:
- "Good enough" std::atomic<void*> and std::atomic<int>
- abort() being implemented
- ubsan_message(const char*) being implemented
Reviewers: eugenis
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D39791
llvm-svn: 320831
Craig Topper [Fri, 15 Dec 2017 17:22:58 +0000 (17:22 +0000)]
[X86] Fix XSAVE64 and similar instructions to not be allowed by the assembler in 32-bit mode.
There was a top level "let Predicates =" in the .td file that was overriding the Requires on each instruction.
I've added an assert to the code emitter to catch more cases like this. I'm sure this isn't the only place where the right predicates aren't being applied. This assert already found that we don't block btq/btsq/btrq in 32-bit mode.
llvm-svn: 320830
Erich Keane [Fri, 15 Dec 2017 16:37:14 +0000 (16:37 +0000)]
Remove "FunctionName -" from docs on FunctionDecl(NFC)
Removed the repetative usage of the operator name on the
documentation for FunctionDecl. Also reflowed some of the
comments since this changes the 80 character rule.
llvm-svn: 320829
Jun Bum Lim [Fri, 15 Dec 2017 16:35:09 +0000 (16:35 +0000)]
Revert "[LICM] Allow sinking when foldable in loop"
This reverts commit r320823.
llvm-svn: 320828
Francis Visoiu Mistrih [Fri, 15 Dec 2017 16:33:45 +0000 (16:33 +0000)]
[CodeGen] Print stack object references as %(fixed-)stack.0 in both MIR and debug output
Work towards the unification of MIR and debug output by printing
`%stack.0` instead of `<fi#0>`, and `%fixed-stack.0` instead of
`<fi#-4>` (supposing there are 4 fixed stack objects).
Only debug syntax is affected.
Differential Revision: https://reviews.llvm.org/D41027
llvm-svn: 320827
Alexey Bataev [Fri, 15 Dec 2017 16:28:31 +0000 (16:28 +0000)]
[OPENMP] Codegen `declare simd` for function declarations.
Previously the attributes were emitted only for function definitions.
Patch adds emission of the attributes for function declarations.
llvm-svn: 320826
Eugene Leviant [Fri, 15 Dec 2017 16:27:33 +0000 (16:27 +0000)]
[ThinLTO] Disallow multiple prevailing defs
https://reviews.llvm.org/D41291
llvm-svn: 320825
Craig Topper [Fri, 15 Dec 2017 16:22:20 +0000 (16:22 +0000)]
[X86] Widen (v2i32 (fp_to_uint v2f64)) to (v8i32 (fp_to_uint v8f64)) during legalization if we have AVX512F, but not VLX. NFC
Previously we widened it using isel patterns.
llvm-svn: 320824
Jun Bum Lim [Fri, 15 Dec 2017 16:09:54 +0000 (16:09 +0000)]
[LICM] Allow sinking when foldable in loop
Summary:
Continue trying to sink an instruction if its users in the loop is foldable.
This will allow the instruction to be folded in the loop by decoupling it from
the user outside of the loop.
Reviewers: hfinkel, majnemer, davidxl, efriedma, danielcdh, bmakam, mcrosier
Reviewed By: hfinkel
Subscribers: javed.absar, bmakam, mcrosier, llvm-commits
Differential Revision: https://reviews.llvm.org/D37076
llvm-svn: 320823
Sam Parker [Fri, 15 Dec 2017 15:30:39 +0000 (15:30 +0000)]
[ARM] Some DAG combine tests
Add some more and and shift load combine tests.
llvm-svn: 320822
Peter Smith [Fri, 15 Dec 2017 15:30:00 +0000 (15:30 +0000)]
[ELF] Reduce size of ELF file produced by test
We only need to exceed 128 Megabytes to provoke the generation of a range
extension thunk. This brings the file size down to just over 128 Megabytes.
llvm-svn: 320821
Pavel Labath [Fri, 15 Dec 2017 15:19:45 +0000 (15:19 +0000)]
llgs-tests: Add support for "exit" stop-reply packets
Summary:
This makes StopReply class abstract, so that we can represent different
types of stop replies such as StopReplyStop and StopReplyExit (there
should also be a StopReplySignal, but I don't need that right now so I
haven't implemented it yet).
This prepares the ground for a new test I'm writing.
Reviewers: eugene, zturner
Subscribers: lldb-commits
Differential Revision: https://reviews.llvm.org/D41067
llvm-svn: 320820
Francis Visoiu Mistrih [Fri, 15 Dec 2017 15:17:18 +0000 (15:17 +0000)]
[MIR] Add support for missing CFI directives
The following CFI directives are suported by MC but not by MIR:
* .cfi_rel_offset
* .cfi_adjust_cfa_offset
* .cfi_escape
* .cfi_remember_state
* .cfi_restore_state
* .cfi_undefined
* .cfi_register
* .cfi_window_save
Add support for printing, parsing and update tests.
Differential Revision: https://reviews.llvm.org/D41230
llvm-svn: 320819
Rafael Espindola [Fri, 15 Dec 2017 14:56:01 +0000 (14:56 +0000)]
Delete a really large test output.
llvm-svn: 320818
Rafael Espindola [Fri, 15 Dec 2017 14:52:40 +0000 (14:52 +0000)]
Handle a VersymIndex of 0 as an error.
I noticed that the continue this patch deletes was not tested. Trying
to add a test I realized that we never put a VER_NDX_LOCAL symbol in
the dynamic symbol table. There doesn't seem to be any reason for a
linker to use VER_NDX_LOCAL for a defined shared symbol.
llvm-svn: 320817
Pavel Labath [Fri, 15 Dec 2017 14:39:12 +0000 (14:39 +0000)]
Fix 32-bit builds broken by 320813
cast to size_t to avoid narrowing error.
llvm-svn: 320816
Simon Pilgrim [Fri, 15 Dec 2017 14:37:28 +0000 (14:37 +0000)]
[X86] Add RTM schedule tests
llvm-svn: 320815
Haicheng Wu [Fri, 15 Dec 2017 14:34:41 +0000 (14:34 +0000)]
[InlineCost] Find repeated loads in the callee
SROA analysis of InlineCost can figure out that some stores can be removed
after inlining and then the repeated loads clobbered by these stores are also
free. This patch finds these clobbered loads and adjust the inline cost
accordingly.
Differential Revision: https://reviews.llvm.org/D33946
llvm-svn: 320814
Pavel Labath [Fri, 15 Dec 2017 14:23:58 +0000 (14:23 +0000)]
ObjectFileELF: Add support for compressed sections
Summary:
We use the llvm decompressor to decompress SHF_COMPRESSED sections. This enables
us to read data from debug info sections, which are sometimes compressed,
particuarly in the split-dwarf case. This functionality is only available if
llvm is compiled with zlib support.
Reviewers: clayborg, zturner
Subscribers: emaste, mgorny, aprantl, lldb-commits
Differential Revision: https://reviews.llvm.org/D40616
llvm-svn: 320813
Simon Pilgrim [Fri, 15 Dec 2017 14:22:15 +0000 (14:22 +0000)]
[X86] Add MWAITX/MONITORX schedule tests
llvm-svn: 320812
Nemanja Ivanovic [Fri, 15 Dec 2017 14:17:45 +0000 (14:17 +0000)]
Fix the second build bot break introduced by r320791.
llvm-svn: 320811
Simon Pilgrim [Fri, 15 Dec 2017 14:02:35 +0000 (14:02 +0000)]
[X86] Add XOP schedule tests
llvm-svn: 320810
Pavel Labath [Fri, 15 Dec 2017 13:56:22 +0000 (13:56 +0000)]
llgs-tests: Make addition of new tests easier
Summary:
Adding a new test would require one to duplicate a significant part of
the existing test that we have. This attempts to reduce that by moving
some part of that code to the test fixture. The StandardStartupTest
fixture automatically starts up the server and connects it to the
client. I also add a more low-level TestBase fixture, which allows one
to start up the client and server in a custom way (I am going to need
this for the test I am writing).
Reviewers: eugene, zturner
Subscribers: lldb-commits, mgorny
Differential Revision: https://reviews.llvm.org/D41066
llvm-svn: 320809
Ilya Biryukov [Fri, 15 Dec 2017 13:20:13 +0000 (13:20 +0000)]
Added a separate install target for compilert-rt-headers
Summary:
This patch adds a new install target `install-compilert-rt-headers`,
that is similar to `install-clang-headers`.
It allows to install the headers without installing all of
compiler-rt.
Reviewers: alekseyshl, beanz
Reviewed By: beanz
Subscribers: smeenai, beanz, mgorny, #sanitizers, kcc, llvm-commits
Differential Revision: https://reviews.llvm.org/D41244
llvm-svn: 320808
Eric Liu [Fri, 15 Dec 2017 12:25:02 +0000 (12:25 +0000)]
[clangd] Build in-memory index on symbols in files.
Reviewers: sammccall
Reviewed By: sammccall
Subscribers: klimek, mgorny, ilya-biryukov, cfe-commits
Differential Revision: https://reviews.llvm.org/D41276
llvm-svn: 320807
Nemanja Ivanovic [Fri, 15 Dec 2017 11:47:48 +0000 (11:47 +0000)]
Fix code causing fallthrough warnings in the PPC back end.
llvm-svn: 320806
Simon Pilgrim [Fri, 15 Dec 2017 11:32:31 +0000 (11:32 +0000)]
[X86] Add AVX512 VPOPCNTDQ schedule tests
Demonstrates how to perform full coverage avx512 schedule tests
llvm-svn: 320805
Ilya Biryukov [Fri, 15 Dec 2017 11:27:51 +0000 (11:27 +0000)]
[clang] Add PPCallbacks list to preprocessor when building a preacompiled preamble.
Summary:
Revision D38639 needs this commit in order to properly make open
definition calls on include statements work.
Patch by William Enright.
Reviewers: malaperle, krasimir, bkramer, ilya-biryukov
Reviewed By: malaperle, ilya-biryukov
Subscribers: cfe-commits, arphaman, ilya-biryukov
Differential Revision: https://reviews.llvm.org/D39375
llvm-svn: 320804
Peter Smith [Fri, 15 Dec 2017 11:09:41 +0000 (11:09 +0000)]
[ELF] Remove Duplicate .ARM.exidx sections
The ARM.exidx section contains a table of 8-byte entries with the first
word of each entry an offset to the function it describes and the second
word instructions for unwinding if an exception is thrown from that
function. The SHF_LINK_ORDER processing will order the table in ascending
order of the functions described by the exception table entries. As the
address range of an exception table entry is terminated by the next table
entry, it is possible to merge consecutive table entries that have
identical unwind instructions.
For this implementation we define a table entry to be identical if:
- Both entries are the special EXIDX_CANTUNWIND.
- Both entries have the same inline unwind instructions.
We do not attempt to establish if table entries that are references to
.ARM.extab sections are identical.
This implementation works at a granularity of a single .ARM.exidx
InputSection. If all entries in the InputSection are identical to the
previous table entry we can remove the InputSection. A more sophisticated
but more complex implementation would rewrite InputSection contents so that
duplicates within a .ARM.exidx InputSection can be merged.
Differential Revision: https://reviews.llvm.org/D40967
llvm-svn: 320803
Peter Smith [Fri, 15 Dec 2017 11:02:50 +0000 (11:02 +0000)]
[ELF] Add missing test from r320800
Forgot to svn add the additional test for the commit. Adding with this one.
llvm-svn: 320802
Peter Smith [Fri, 15 Dec 2017 10:47:44 +0000 (10:47 +0000)]
[ELF] Fix buildbot warning error.
Forward declare InputSectionDescription as struct not class.
llvm-svn: 320801
Peter Smith [Fri, 15 Dec 2017 10:32:34 +0000 (10:32 +0000)]
[ELF] Complete implementation of --fix-cortex-a53-843419
This patch provides the mechanism to fix instances of the instruction
sequence that may trigger the cortex-a53 843419 erratum. The fix is
provided by an alternative instruction sequence to remove one of the
erratum conditions. To reach this alternative instruction sequence we
replace the original instruction with a branch to the alternative
sequence. The alternative sequence is responsible for branching back to
the original.
As there is only erratum to fix the implementation is specific to
AArch64 and the specific erratum conditions. It should be generalizable
to other targets and erratum if needed.
Differential Revision: https://reviews.llvm.org/D36749
llvm-svn: 320800
Alex Bradbury [Fri, 15 Dec 2017 10:20:51 +0000 (10:20 +0000)]
[RISCV] Change shift amount operand of RVC shift instructions to uimmlog2xlennonzero
c.slli/c.srli/c.srai allow a 5-bit shift in RV32C and a 6-bit shift in RV64C.
This patch adds uimmlog2xlennonzero to reflect this constraint as well as
tests.
Differential Revision: https://reviews.llvm.org/D41216
Patch by Shiva Chen.
llvm-svn: 320799