Rafael Espindola [Sun, 24 Apr 2016 21:30:18 +0000 (21:30 +0000)]
Add a test for how we handle protected visibility.
llvm-svn: 267366
Saleem Abdulrasool [Sun, 24 Apr 2016 21:01:04 +0000 (21:01 +0000)]
unwind: remove unnecessary header
Availablity.h is not used within config.h. The locations which use the
availability infrastructure already include the necessary header(s). NFC.
llvm-svn: 267365
Saleem Abdulrasool [Sun, 24 Apr 2016 21:00:59 +0000 (21:00 +0000)]
unwind: unify _LIBUNWIND_ABORT
Rather than use the `__assert_rtn` on libSystem based targets and a local
`assert_rtn` function on others, expand the function definition into a macro
which will perform the writing to stderr and then abort. This unifies the
definition and behaviour across targets.
Ensure that we flush stderr prior to aborting.
llvm-svn: 267364
Ulrich Weigand [Sun, 24 Apr 2016 20:49:56 +0000 (20:49 +0000)]
Fix unwind failures when PC points beyond the end of a function
RegisterContextLLDB::InitializeNonZerothFrame already has code to attempt
to detect and handle the case where the PC points beyond the end of a
function, but there are certain cases where this doesn't work correctly.
In fact, there are *two* different places where this detection is attempted,
and the failure is in fact a result of an unfortunate interaction between
those two separate attempts.
First, the ResolveSymbolContextForAddress routine is called with the
resolve_tail_call_address flag set to true. This causes the routine
to internally accept a PC pointing beyond the end of a function, and
still resolving the PC to that function symbol.
Second, the InitializeNonZerothFrame routine itself maintains a
"decr_pc_and_recompute_addr_range" flag and, if that turns out to
be true, itself decrements the PC by one and searches again for
a symbol at that new PC value.
Both approaches correctly identify the symbol associated with the PC.
However, the problem is now that later on, we also need to find the
DWARF CFI record associated with the PC. This is done in the
RegisterContextLLDB::GetFullUnwindPlanForFrame routine, and uses
the "m_current_offset_backed_up_one" member variable.
However, that variable only actually contains the PC "backed up by
one" if the *second* approach above was taken. If the function was
already identified via the first approach above, that member variable
is *not* backed up by one but simply points to the original PC.
This in turn causes GetEHFrameUnwindPlan to not correctly identify
the DWARF CFI record associated with the PC.
Now, in many cases, if the first method had to back up the PC by one,
we *still* use the second method too, because of this piece of code:
// Or if we're in the middle of the stack (and not "above" an asynchronous event like sigtramp),
// and our "current" pc is the start of a function...
if (m_sym_ctx_valid
&& GetNextFrame()->m_frame_type != eTrapHandlerFrame
&& GetNextFrame()->m_frame_type != eDebuggerFrame
&& addr_range.GetBaseAddress().IsValid()
&& addr_range.GetBaseAddress().GetSection() == m_current_pc.GetSection()
&& addr_range.GetBaseAddress().GetOffset() == m_current_pc.GetOffset())
{
decr_pc_and_recompute_addr_range = true;
}
In many cases, when the PC is one beyond the end of the current function,
it will indeed then be exactly at the start of the next function. But this
is not always the case, e.g. if there happens to be alignment padding
between the end of one function and the start of the next.
In those cases, we may sucessfully look up the function symbol via
ResolveSymbolContextForAddress, but *not* set decr_pc_and_recompute_addr_range,
and therefore fail to find the correct DWARF CFI record.
A very simple fix for this problem is to just never use the first method.
Call ResolveSymbolContextForAddress with resolve_tail_call_address set
to false, which will cause it to fail if the PC is beyond the end of
the current function; or else, identify the next function if the PC
is also at the start of the next function. In either case, we will
then set the decr_pc_and_recompute_addr_range variable and back up the
PC anyway, but this time also find the correct DWARF CFI.
A related problem is that the ResolveSymbolContextForAddress sometimes
returns a "symbol" with empty name. This turns out to be an ELF section
symbol. Now, usually those get type eSymbolTypeInvalid. However, there
is code in ObjectFileELF::ParseSymbols that tries to change the type of
invalid symbols to eSymbolTypeCode or eSymbolTypeData if the symbol
lies within the code or data section.
Unfortunately, this check also hits the symbol for the code section
itself, which is then marked as eSymbolTypeCode. While the size of
the section symbol is 0 according to the ELF file, LLDB considers
this size invalid and attempts to figure out the "correct" size.
Depending on how this goes, we may end up with a symbol that overlays
part of the code section, even outside areas covered by real function
symbols.
Therefore, if we call ResolveSymbolContextForAddress with PC pointing
beyond the end of a function, we may get this bogus section symbol.
This again means InitializeNonZerothFrame thinks we have a valid PC,
but then we don't find any unwind info for it.
The fix for this problem is me to simply always leave ELF section
symbols as type eSymbolTypeInvalid.
Differential Revision: http://reviews.llvm.org/D18975
llvm-svn: 267363
Simon Pilgrim [Sun, 24 Apr 2016 20:30:48 +0000 (20:30 +0000)]
[X86][AVX] Added PR24935 test case
llvm-svn: 267362
Saleem Abdulrasool [Sun, 24 Apr 2016 20:12:48 +0000 (20:12 +0000)]
ARM: fix __chkstk Frame Setup on WoA
This corrects the MI annotations for the stack adjustment following the __chkstk
invocation. We were marking the original SP usage as a Def rather than Kill.
The (new) assigned value is the definition, the original reference is killed.
Adjust the ISelLowering to mark Kills and FrameSetup as well.
This partially resolves PR27480.
llvm-svn: 267361
Simon Pilgrim [Sun, 24 Apr 2016 19:31:56 +0000 (19:31 +0000)]
Tweak comments to make it clear that these combines are for SSE scalar instructions.
llvm-svn: 267360
Simon Pilgrim [Sun, 24 Apr 2016 18:35:59 +0000 (18:35 +0000)]
[InstCombine][SSE] Reduce DIVSS/DIVSD to FDIV if only first element is required
As discussed on D19318, if we only demand the first element of a DIVSS/DIVSD intrinsic, then reduce to a FDIV call. This matches the existing FADD/FSUB/FMUL patterns.
llvm-svn: 267359
Davide Italiano [Sun, 24 Apr 2016 18:23:21 +0000 (18:23 +0000)]
[ELF] Reinstate 'else' which was previously removed.
It turns out it's actually needed.
llvm-svn: 267358
Simon Pilgrim [Sun, 24 Apr 2016 18:23:14 +0000 (18:23 +0000)]
[InstCombine][SSE] Demanded vector elements for scalar intrinsics (Part 2 of 2)
Split from D17490. This patch improves support for determining the demanded vector elements through SSE scalar intrinsics:
1 - demanded vector element support for unary and some extra binary scalar intrinsics (RCP/RSQRT/SQRT/FRCZ and ADD/CMP/DIV/ROUND).
2 - addss/addsd get simplified to a fadd call if we aren't interested in the pass through elements
3 - if we don't need the lowest element of a scalar operation then just use the first argument (the pass through elements) directly
We can add support for propagating demanded elements through any equivalent packed SSE intrinsics in a future patch (these wouldn't use the pass through patterns).
Differential Revision: http://reviews.llvm.org/D19318
llvm-svn: 267357
Simon Pilgrim [Sun, 24 Apr 2016 18:12:42 +0000 (18:12 +0000)]
[InstCombine][SSE] Demanded vector elements for scalar intrinsics (Part 1 of 2)
This patch improves support for determining the demanded vector elements through SSE scalar intrinsics:
1 - recognise that we only need the lowest element of the second input for binary scalar operations (and all the elements of the first input)
2 - recognise that the roundss/roundsd intrinsics use the lowest element of the second input and the remaining elements from the first input
Differential Revision: http://reviews.llvm.org/D17490
llvm-svn: 267356
Simon Pilgrim [Sun, 24 Apr 2016 17:57:27 +0000 (17:57 +0000)]
[InstCombine] Avoid updating argument demanded elements in separate passes.
As discussed on D17490, we should attempt to update an intrinsic's arguments demanded elements in one pass if we can.
llvm-svn: 267355
Nick Lewycky [Sun, 24 Apr 2016 17:55:57 +0000 (17:55 +0000)]
Fix typo in comment. NFC
llvm-svn: 267354
Nick Lewycky [Sun, 24 Apr 2016 17:55:41 +0000 (17:55 +0000)]
Remove emacs mode markers from .cpp files. NFC
.cpp files are unambiguously C++, you only need the mode markers on .h files.
llvm-svn: 267353
Simon Pilgrim [Sun, 24 Apr 2016 17:23:46 +0000 (17:23 +0000)]
[X86][InstCombine] Tidyup VPERMILVAR -> shufflevector conversion to helper function. NFCI.
llvm-svn: 267352
Simon Pilgrim [Sun, 24 Apr 2016 17:00:34 +0000 (17:00 +0000)]
[X86][InstCombine] Tidyup PSHUFB -> shufflevector conversion to helper function. NFCI.
llvm-svn: 267351
Simon Pilgrim [Sun, 24 Apr 2016 16:49:53 +0000 (16:49 +0000)]
[X86][SSE] getTargetShuffleMaskIndices - dropped (unused) UNDEF handling
We aren't currently making use of this in any successful mask decode and its actually incorrect as it inserts the wrong number of SM_SentinelUndef mask elements.
llvm-svn: 267350
Simon Pilgrim [Sun, 24 Apr 2016 16:33:35 +0000 (16:33 +0000)]
[X86][SSE] Use range loop. NFCI.
llvm-svn: 267349
Craig Topper [Sun, 24 Apr 2016 16:30:51 +0000 (16:30 +0000)]
[Lanai] Use EVT::getEVTString() to print a type as a string instead of an enum encoding value.
llvm-svn: 267348
Simon Pilgrim [Sun, 24 Apr 2016 15:45:06 +0000 (15:45 +0000)]
[X86][SSE] Added SSSE3/AVX/AVX2 BITREVERSE tests
Codegen is pretty bad at the moment but could use PSHUFB quite efficiently
llvm-svn: 267347
Simon Pilgrim [Sun, 24 Apr 2016 15:05:04 +0000 (15:05 +0000)]
[X86][XOP] Fixed VPPERM permute op decoding (PR27472).
Fixed issue with VPPERM target shuffle mask decoding that was incorrectly masking off the 3-bit permute op with a 2-bit mask.
llvm-svn: 267346
Duncan P. N. Exon Smith [Sun, 24 Apr 2016 15:04:28 +0000 (15:04 +0000)]
BitcodeReader: Delay metadata parsing until reading a function body
There's hardly any functionality change here. Instead of calling
materializeMetadata on the first call to materialize(GlobalValue*), wait
until the first one that's actually going to do something. Noticed by
inspection; I don't have a concrete case where this makes a difference.
Added an assertion in materializeMetadata to be sure this (or a future
change) doesn't delay materializeMetadata after function-level metadata.
llvm-svn: 267345
Teresa Johnson [Sun, 24 Apr 2016 14:57:11 +0000 (14:57 +0000)]
[ThinLTO] Remove GlobalValueInfo class from index
Summary:
Remove the GlobalValueInfo and change the ModuleSummaryIndex to directly
reference summary objects. The info structure was there to support lazy
parsing of the combined index summary objects, which is no longer
needed and not supported.
Reviewers: joker.eph
Subscribers: joker.eph, llvm-commits
Differential Revision: http://reviews.llvm.org/D19462
llvm-svn: 267344
Simon Pilgrim [Sun, 24 Apr 2016 14:53:54 +0000 (14:53 +0000)]
[X86][SSE] Improved support for decoding target shuffle masks through bitcasts
Reused the ability to split constants of a type wider than the shuffle mask to work with masks generated from scalar constants transfered to xmm.
This fixes an issue preventing PSHUFB target shuffle masks decoding rematerialized scalar constants and also exposes the XOP VPPERM bug described in PR27472.
llvm-svn: 267343
Duncan P. N. Exon Smith [Sun, 24 Apr 2016 14:25:37 +0000 (14:25 +0000)]
ModuleSummaryIndex: Avoid enum bitfields for MSVC portability
Enum bitfields have crazy portability issues with MSVC. Use unsigned
instead of LinkageTypes here in the ModuleSummaryIndex to address
Takumi's concerns from r267335.
llvm-svn: 267342
Duncan P. N. Exon Smith [Sun, 24 Apr 2016 14:13:17 +0000 (14:13 +0000)]
Revert "Declare GlobalValue::LinkageTypes based on unsigned."
This reverts commit r267335. The build has been broken for hours
because of it:
http://lab.llvm.org:8080/green/job/clang-stage1-cmake-RA-incremental_build/23352/
The correct fix is avoid using any enum in a bitfield.
llvm-svn: 267341
Marcin Koscielnicki [Sun, 24 Apr 2016 13:57:49 +0000 (13:57 +0000)]
[SystemZ] [SSP] Add support for LOAD_STACK_GUARD.
This fixes PR22248 on s390x. The previous attempt at this was D19101,
which was before LOAD_STACK_GUARD existed. Compared to the previous
version, this always emits a rather ugly block of 4 instructions, involving
a thread pointer load that can't be shared with other potential users.
However, this is necessary for SSP - spilling the guard value (or thread
pointer used to load it) is counter to the goal, since it could be
overwritten along with the frame it protects.
Differential Revision: http://reviews.llvm.org/D19363
llvm-svn: 267340
Simon Pilgrim [Sun, 24 Apr 2016 13:45:30 +0000 (13:45 +0000)]
[X86][SSE] Demonstrate issue with decoding shuffle masks that have been lowered as rematerialized constants on scalar unit
Found whilst investigating PR27472
llvm-svn: 267339
Aaron Ballman [Sun, 24 Apr 2016 13:30:21 +0000 (13:30 +0000)]
Improve diagnostic checking for va_start to also warn on other instances of undefined behavior, such as a parameter declared with the register keyword in C, or a parameter of a type that undergoes default argument promotion.
This helps cover some more of the CERT secure coding rule EXP58-CPP. Pass an object of the correct type to va_start (https://www.securecoding.cert.org/confluence/display/cplusplus/EXP58-CPP.+Pass+an+object+of+the+correct+type+to+va_start).
llvm-svn: 267338
Aaron Ballman [Sun, 24 Apr 2016 13:03:20 +0000 (13:03 +0000)]
Silence two C4806 warnings ('|': unsafe operation: no value of type 'bool' promoted to type 'const unsigned int' can equal the given constant). The fact that they trigger with this code seems like it may be a bug, but the warning itself is still generally useful enough to retain it for now.
llvm-svn: 267337
Johannes Doerfert [Sun, 24 Apr 2016 12:31:02 +0000 (12:31 +0000)]
Remove unused iterators [NFC]
llvm-svn: 267336
NAKAMURA Takumi [Sun, 24 Apr 2016 10:11:45 +0000 (10:11 +0000)]
Declare GlobalValue::LinkageTypes based on unsigned.
Or, "LinkageTypes Linkage : 4;" might be sign-extended on msc.
llvm-svn: 267335
NAKAMURA Takumi [Sun, 24 Apr 2016 08:02:00 +0000 (08:02 +0000)]
llvm/test/tools/gold/X86/thinlto.ll: Possible fix corresponding to r267318.
llvm-svn: 267334
Davide Italiano [Sun, 24 Apr 2016 07:19:32 +0000 (07:19 +0000)]
[ELF] Simplify. Remove unneeded else. NFC.
llvm-svn: 267333
Duncan P. N. Exon Smith [Sun, 24 Apr 2016 06:52:01 +0000 (06:52 +0000)]
BitcodeReader: Fix some holes in upgrade from r267296
Add tests for some missing cases to bitcode upgrade in r267296.
- DICompositeType with an 'elements:' field, which will cause it to be
involved in a cycle after the upgrade.
- A DIDerivedType that references a class in 'extraData:'.
I updated test/Bitcode/dityperefs-3.8.ll with the missing cases and
regenerated test/Bitcode/dityperefs-3.8.ll.bc.
llvm-svn: 267332
Craig Topper [Sun, 24 Apr 2016 06:27:39 +0000 (06:27 +0000)]
[X86] Merge LowerCTLZ and LowerCTLZ_ZERO_UNDEF into a single function that branches internally for the one difference, allowing the rest of the code to be common. NFC
llvm-svn: 267331
Craig Topper [Sun, 24 Apr 2016 06:27:35 +0000 (06:27 +0000)]
[X86] Node need to check if AVX512 is supported when lowering vector CTLZ. The CTLZ operation is only Custom for vectors if AVX512 is enabled so if a vector gets here AVX512 is implied. NFC
llvm-svn: 267330
Mehdi Amini [Sun, 24 Apr 2016 05:31:43 +0000 (05:31 +0000)]
Add "hasSection" flag in the Summary
Reviewers: tejohnson
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D19405
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 267329
Gerolf Hoflehner [Sun, 24 Apr 2016 05:14:01 +0000 (05:14 +0000)]
[MachineCombiner] Support for floating-point FMA on ARM64 (re-commit r267098)
The original patch caused crashes because it could derefence a null pointer
for SelectionDAGTargetInfo for targets that do not define it.
Evaluates fmul+fadd -> fmadd combines and similar code sequences in the
machine combiner. It adds support for float and double similar to the existing
integer implementation. The key features are:
- DAGCombiner checks whether it should combine greedily or let the machine
combiner do the evaluation. This is only supported on ARM64.
- It gives preference to throughput over latency: the heuristic used is
to combine always in loops. The targets decides whether the machine
combiner should optimize for throughput or latency.
- Supports for fmadd, f(n)msub, fmla, fmls patterns
- On by default at O3 ffast-math
llvm-svn: 267328
Simon Atanasyan [Sun, 24 Apr 2016 05:08:43 +0000 (05:08 +0000)]
[ELF][MIPS] Add test case to check creation of GOT entries for weak symbols. NFC
llvm-svn: 267327
Craig Topper [Sun, 24 Apr 2016 04:38:34 +0000 (04:38 +0000)]
[X86] Remove isel patterns for selecting tzcnt/lzcnt from cmove/ne+cttz/ctlz. These are folded by DAG combine now.
llvm-svn: 267326
Craig Topper [Sun, 24 Apr 2016 04:38:32 +0000 (04:38 +0000)]
[CodeGen] Teach DAG combine to fold select_cc seteq X, 0, sizeof(X), ctlz_zero_undef(X) -> ctlz(X). InstCombine already does this for IR and X86 pattern matches this during isel.
A follow up commit will remove the X86 patterns to allow this to be tested.
llvm-svn: 267325
Craig Topper [Sun, 24 Apr 2016 04:38:29 +0000 (04:38 +0000)]
Fix an assertion that can never fire because the condition ANDed with the string is just true or 1.
llvm-svn: 267324
Peter Collingbourne [Sun, 24 Apr 2016 04:29:59 +0000 (04:29 +0000)]
ELF: Simplify preemption logic. Do not include weak undefined symbols in non-DSOs.
Add a test for -Bsymbolic + undefined symbols.
llvm-svn: 267323
Adrian Prantl [Sun, 24 Apr 2016 03:47:37 +0000 (03:47 +0000)]
Revert "Verifier: Verify that each inlinable callsite of a debug-info-bearing function"
This reverts commit r267320 while investigating an OpenMP buildbot failure.
llvm-svn: 267322
Mehdi Amini [Sun, 24 Apr 2016 03:44:55 +0000 (03:44 +0000)]
Make thinlto clang test more robust against LLVM changes.
We should just test the effect of the clang level option here, i.e.
that a summary is correctly emitted with -flto=thin
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 267321
Adrian Prantl [Sun, 24 Apr 2016 03:23:02 +0000 (03:23 +0000)]
Verifier: Verify that each inlinable callsite of a debug-info-bearing function
in a debug-info-bearing function has a debug location attached to it. Failure to
do so causes an "!dbg attachment points at wrong subprogram for function"
assertion failure when the inliner sets up inline scope info.
rdar://problem/
25878916
llvm-svn: 267320
Mehdi Amini [Sun, 24 Apr 2016 03:18:18 +0000 (03:18 +0000)]
Reorganize GlobalValueSummary with a "Flags" bitfield.
Right now it only contains the LinkageType, but will be extended
with "hasSection", "isOptSize", "hasInlineAssembly", etc.
Differential Revision: http://reviews.llvm.org/D19404
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 267319
Mehdi Amini [Sun, 24 Apr 2016 03:18:11 +0000 (03:18 +0000)]
Add a version field in the bitcode for the summary
Differential Revision: http://reviews.llvm.org/D19456
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 267318
Mehdi Amini [Sun, 24 Apr 2016 03:18:01 +0000 (03:18 +0000)]
Add an internalization step to the ThinLTOCodeGenerator
Keeping as much as possible internal/private is
known to help the optimizer. Let's try to benefit from
this in ThinLTO.
Note: this is early work, but is enough to build clang (and
all the LLVM tools). I still need to write some lit-tests...
Differential Revision: http://reviews.llvm.org/D19103
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 267317
Peter Collingbourne [Sun, 24 Apr 2016 02:42:34 +0000 (02:42 +0000)]
Test what I intended to test.
llvm-svn: 267316
Peter Collingbourne [Sun, 24 Apr 2016 02:31:04 +0000 (02:31 +0000)]
ELF: Forbid undefined symbols with non-default visibility in DSOs.
llvm-svn: 267315
Peter Collingbourne [Sun, 24 Apr 2016 02:31:02 +0000 (02:31 +0000)]
ELF: Always include undefined DSO symbols in the symbol table.
Fixes check-llvm when bootstrapping.
Also remove mostly dead and most likely incorrect logic regarding preemption
of weak undefined symbols.
llvm-svn: 267314
Craig Topper [Sun, 24 Apr 2016 02:08:22 +0000 (02:08 +0000)]
Fix a couple assertions that can never fire because the condition ANDed with the string is just true or 1.
llvm-svn: 267313
Craig Topper [Sun, 24 Apr 2016 02:01:25 +0000 (02:01 +0000)]
Fix a couple assertions that can never fire because they just contained the text string which always evaluates to true. Add a ! so they'll evaluate to false.
llvm-svn: 267312
Craig Topper [Sun, 24 Apr 2016 02:01:22 +0000 (02:01 +0000)]
[X86] Fix patterns that turn cmove/cmovne+ctlz/cttz into lzcnt/tzcnt instructions. Only one of the conditions should be valid for each pattern, not both. Update tests accordingly.
llvm-svn: 267311
Davide Italiano [Sun, 24 Apr 2016 01:56:02 +0000 (01:56 +0000)]
[ELF] Use llvm-mc rather than checking in a file.
Now that the support for the 'new' relocations is available.
llvm-svn: 267310
Davide Italiano [Sun, 24 Apr 2016 01:36:37 +0000 (01:36 +0000)]
[RuntimeDyldELF] Handle GOTPCRELX/REX_GOTPCRELX.
llvm-svn: 267309
Davide Italiano [Sun, 24 Apr 2016 01:08:35 +0000 (01:08 +0000)]
[MC/ELF] Make the relaxation test more interesting.
Add a case where we can't relax.
llvm-svn: 267308
Davide Italiano [Sun, 24 Apr 2016 01:03:57 +0000 (01:03 +0000)]
[MC/ELF] Implement support for GOTPCRELX/REX_GOTPCRELX.
The option to control the emission of the new relocations
is -relax-relocations (blatantly copied from GNU as).
It can't be enabled by default because it breaks relatively
recent versions of ld.bfd/ld.gold (late 2015).
llvm-svn: 267307
Mehdi Amini [Sun, 24 Apr 2016 00:25:15 +0000 (00:25 +0000)]
Relax test using CHECK-DAG instead of CHECK-NEXT
It seems we still have some ordering issue in the combined index
emission, but I can't figure out why right now.
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 267306
Mehdi Amini [Sun, 24 Apr 2016 00:03:57 +0000 (00:03 +0000)]
Fix test stability (was sensitive to the path)
This is a fixup for r267304.
The test was sensitive to the path in a subtle way:
the index in memory is sorted by GUID, which are hashes
that include the source filename for local globals.
Teresa recently added a directive at the IR level, so
we can specify it here to make the test independent of
the path.
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 267305
Mehdi Amini [Sat, 23 Apr 2016 23:38:17 +0000 (23:38 +0000)]
Store and emit original name in combined index
Summary:
As discussed in D18298, some local globals can't
be renamed/promoted (because they have a section, or because
they are referenced from inline assembly).
To be able to detect naming collision, we need to keep around
the "GUID" using their original name without taking the linkage
into account.
Reviewers: tejohnson
Subscribers: joker.eph, llvm-commits
Differential Revision: http://reviews.llvm.org/D19454
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 267304
Mehdi Amini [Sat, 23 Apr 2016 23:29:24 +0000 (23:29 +0000)]
Always traverse GlobalVariable initializer when computing the export list
Summary:
We are always importing the initializer for a GlobalVariable.
So if a GlobalVariable is in the export-list, we pull in any
refs as well.
Reviewers: tejohnson
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D19102
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 267303
Duncan P. N. Exon Smith [Sat, 23 Apr 2016 22:29:26 +0000 (22:29 +0000)]
DebugInfo: DIGlobalVariables became 'distinct' in LLVM r267301
Update testcases due to DIBuilder change.
llvm-svn: 267302
Duncan P. N. Exon Smith [Sat, 23 Apr 2016 22:29:09 +0000 (22:29 +0000)]
DebugInfo: Change DIBuilder to make distinct DIGlobalVariables
A long overdue change to make DIGlobalVariable distinct. Much like
DISubprogram definitions (changed in r246098), it isn't logical to
unique DIGlobalVariable definitions from two different compile units.
(Longer-term, we should also find a way to reverse the link between
GlobalVariable and DIGlobalVariable, and between DIGlobalVariable and
DICompileUnit, so that debug info to do with optimized-out globals
disappears. Admittedly it's harder than with Function/DISubprogram,
since global variables may be constant-folded and the debug info should
still describe that somehow.)
llvm-svn: 267301
Davide Italiano [Sat, 23 Apr 2016 22:26:31 +0000 (22:26 +0000)]
[MC/ELF] Pass Fixup to getRelocType64.
In preparation for other changes.
llvm-svn: 267300
Duncan P. N. Exon Smith [Sat, 23 Apr 2016 21:36:59 +0000 (21:36 +0000)]
BitcodeReader: Avoid std::vector with non-movable types from r267296
r267298 didn't quite fix the build errors. Use SmallVector instead of
std::vector, the latter of which I think is trying to maintain a strong
exception safety guarantee.
http://lab.llvm.org:8011/builders/lldb-amd64-ninja-freebsd11/builds/6228
llvm-svn: 267299
Duncan P. N. Exon Smith [Sat, 23 Apr 2016 21:23:41 +0000 (21:23 +0000)]
BitcodeReader: Avoid non-moving std::piecewise_construct from r267296
Not exactly sure why the host tries to use a copy constructor here, but
it's easy enough to work around it.
http://lab.llvm.org:8011/builders/lldb-amd64-ninja-freebsd11/builds/6227
llvm-svn: 267298
Duncan P. N. Exon Smith [Sat, 23 Apr 2016 21:08:27 +0000 (21:08 +0000)]
DebugInfo: Adapt to loss of DITypeRef in LLVM r267296
LLVM stopped using MDString-based type references, and DIBuilder no
longer fills 'retainedTypes:' with every DICompositeType that has an
'identifier:' field. There are just minor changes to keep the same
behaviour in CFE.
Leaving 'retainedTypes:' unfilled has a dramatic impact on the output
order of the IR though. There are a huge number of testcase changes,
which were unfortunately not really scriptable.
llvm-svn: 267297
Duncan P. N. Exon Smith [Sat, 23 Apr 2016 21:08:00 +0000 (21:08 +0000)]
DebugInfo: Remove MDString-based type references
Eliminate DITypeIdentifierMap and make DITypeRef a thin wrapper around
DIType*. It is no longer legal to refer to a DICompositeType by its
'identifier:', and DIBuilder no longer retains all types with an
'identifier:' automatically.
Aside from the bitcode upgrade, this is mainly removing logic to resolve
an MDString-based reference to an actualy DIType. The commits leading
up to this have made the implicit type map in DICompileUnit's
'retainedTypes:' field superfluous.
This does not remove DITypeRef, DIScopeRef, DINodeRef, and
DITypeRefArray, or stop using them in DI-related metadata. Although as
of this commit they aren't serving a useful purpose, there are patchces
under review to reuse them for CodeView support.
The tests in LLVM were updated with deref-typerefs.sh, which is attached
to the thread "[RFC] Lazy-loading of debug info metadata":
http://lists.llvm.org/pipermail/llvm-dev/2016-April/098318.html
llvm-svn: 267296
Sanjay Patel [Sat, 23 Apr 2016 20:01:22 +0000 (20:01 +0000)]
replace duplicated static functions for profile metadata access with BranchInst member function; NFCI
llvm-svn: 267295
Renato Golin [Sat, 23 Apr 2016 19:30:52 +0000 (19:30 +0000)]
Revert "[AArch64] Fix optimizeCondBranch logic."
This reverts commit r267206, as it broke self-hosting on AArch64.
llvm-svn: 267294
Derek Bruening [Sat, 23 Apr 2016 16:41:24 +0000 (16:41 +0000)]
[esan] EfficiencySanitizer libc interceptors
Summary:
Adds libc interceptors to the runtime library for the new
EfficiencySanitizer ("esan") family of tools. The interceptors cover
the memory operations in most common library calls and will be shared
among all esan tools.
Reviewers: aizatsky
Subscribers: zhaoqin, tberghammer, danalbert, srhines, llvm-commits, vitalybuka, eugenis, kcc
Differential Revision: http://reviews.llvm.org/D19411
llvm-svn: 267293
Sanjay Patel [Sat, 23 Apr 2016 16:31:48 +0000 (16:31 +0000)]
improve documentation comments; NFC
llvm-svn: 267292
Saleem Abdulrasool [Sat, 23 Apr 2016 16:00:15 +0000 (16:00 +0000)]
ObjectFile: parse EABI Attributes
This adds basic parsing of the EABI attributes section. This section contains
additional information about the target for which the file was built. Attempt
to infer additional architecture information from that section.
llvm-svn: 267291
NAKAMURA Takumi [Sat, 23 Apr 2016 14:54:28 +0000 (14:54 +0000)]
clang-tidy/cppcoreguidelines/ProTypeMemberInitCheck.cpp: Use raw_string_ostream::str() to flush the buffer explicitly.
llvm-svn: 267290
Johannes Doerfert [Sat, 23 Apr 2016 14:32:34 +0000 (14:32 +0000)]
Add an invalid domain to memory accesses
Memory accesses can have non-precisely modeled access functions that
would cause us to build incorrect execution context for hoisted loads.
This is the same issue that occurred during the domain construction for
statements and it is dealt with the same way.
llvm-svn: 267289
Johannes Doerfert [Sat, 23 Apr 2016 14:31:17 +0000 (14:31 +0000)]
Translate SCEVs to isl_pw_aff and their invalid domain
The SCEVAffinator will now produce not only the isl representaiton of
a SCEV but also the domain under which it is invalid. This is used to
record possible overflows that can happen in the statement domains in
the statements invalid domain. The result is that invalid loads have
an accurate execution contexts with regards to the validity of their
statements domain. While the SCEVAffinator currently is only taking
"no-wrapping" assumptions, we can add more withouth worrying about the
execution context of loads that are optimistically hoisted.
llvm-svn: 267288
Marshall Clow [Sat, 23 Apr 2016 14:31:00 +0000 (14:31 +0000)]
Rename a few tests that had typos in their names. No functional change. Thanks to STL for the catch
llvm-svn: 267287
Johannes Doerfert [Sat, 23 Apr 2016 13:02:23 +0000 (13:02 +0000)]
Track invalid domains not invalid contexts for statements
The invalid context is not enough to describe the parameter constraints under
which a statement is not modeled precisely. The reason is that during the
domain construction the bounds on the induction variables are not known but
needed to check if e.g., an overflow can actually happen. To this end we
replace the invalid context of a statement with an invalid domain. It is
initialized during domain construction and intersected with the domain once
it was completely build. Later this invalid domain allows to eliminate
falsely assumed wrapping cases and other falsely assumed mismatches in the
modeling.
llvm-svn: 267286
Johannes Doerfert [Sat, 23 Apr 2016 13:00:27 +0000 (13:00 +0000)]
Improve accuracy of Scop::hasFeasibleRuntimeContext
If the AssumptionContext is a subset of the InvalidContext the runtime
context is not feasible.
llvm-svn: 267285
Johannes Doerfert [Sat, 23 Apr 2016 12:59:18 +0000 (12:59 +0000)]
Simplify the execution context for dereferencable loads
If we know it is safe to execute a load we do not need an execution
context, however only if we are sure it was modeled correctly.
llvm-svn: 267284
Johannes Doerfert [Sat, 23 Apr 2016 12:56:58 +0000 (12:56 +0000)]
Remove simplification calls for the execution domain [NFC]
These calls were sometimes costly and do not show any improvements on our
small test cases.
llvm-svn: 267283
Simon Pilgrim [Sat, 23 Apr 2016 11:14:18 +0000 (11:14 +0000)]
[X86][XOP] Added VPPERM -> BLEND-WITH-ZERO Test
Currently failing due to poor blend matching, found whilst investigating PR27472
llvm-svn: 267282
Benjamin Kramer [Sat, 23 Apr 2016 11:01:36 +0000 (11:01 +0000)]
Use %T instead of cd'ing to Output directly.
%T expands to Output if not configured differently.
llvm-svn: 267281
Craig Topper [Sat, 23 Apr 2016 05:20:47 +0000 (05:20 +0000)]
[CodeGen] When promoting CTTZ operations to larger type, don't insert a select to detect if the input is zero to return the original size instead of the extended size. Instead just set the first bit in the zero extended part.
llvm-svn: 267280
Teresa Johnson [Sat, 23 Apr 2016 05:15:59 +0000 (05:15 +0000)]
[gold] Gate value name discarding under save-temps
Summary:
This removes a couple of flags added to control this behavior, and
simply keeps all value names when save-temps is specified.
Reviewers: rafael
Subscribers: llvm-commits, pcc, davide
Differential Revision: http://reviews.llvm.org/D19384
llvm-svn: 267279
Duncan P. N. Exon Smith [Sat, 23 Apr 2016 04:59:22 +0000 (04:59 +0000)]
BitcodeWriter: Emit uniqued subgraphs after all distinct nodes
Since forward references for uniqued node operands are expensive (and
those for distinct node operands are cheap due to
DistinctMDOperandPlaceholder), minimize forward references in uniqued
node operands.
Moreover, guarantee that when a cycle is broken by a distinct node, none
of the uniqued nodes have any forward references. In
ValueEnumerator::EnumerateMetadata, enumerate uniqued node subgraphs
first, delaying distinct nodes until all uniqued nodes have been
handled. This guarantees that uniqued nodes only have forward
references when there is a uniquing cycle (since r267276 changed
ValueEnumerator::organizeMetadata to partition distinct nodes in front
of uniqued nodes as a post-pass).
Note that a single uniqued subgraph can hit multiple distinct nodes at
its leaves. Ideally these would themselves be emitted in post-order,
but this commit doesn't attempt that; I think it requires an extra pass
through the edges, which I'm not convinced is worth it (since
DistinctMDOperandPlaceholder makes forward references quite cheap
between distinct nodes).
I've added two testcases:
- test/Bitcode/mdnodes-distinct-in-post-order.ll is just like
test/Bitcode/mdnodes-in-post-order.ll, except with distinct nodes
instead of uniqued ones. This confirms that, in the absence of
uniqued nodes, distinct nodes are still emitted in post-order.
- test/Bitcode/mdnodes-distinct-nodes-break-cycles.ll is the minimal
example where a naive post-order traversal would cause one uniqued
node to forward-reference another. IOW, it's the motivating test.
llvm-svn: 267278
Duncan P. N. Exon Smith [Sat, 23 Apr 2016 04:52:47 +0000 (04:52 +0000)]
Avoid MSVC failure with default arguments in lambdas from r267270
http://lab.llvm.org:8011/builders/clang-x64-ninja-win7/builds/11700
llvm-svn: 267277
Duncan P. N. Exon Smith [Sat, 23 Apr 2016 04:42:39 +0000 (04:42 +0000)]
BitcodeWriter: Emit distinct nodes before uniqued nodes
When an operand of a distinct node hasn't been read yet, the reader can
use a DistinctMDOperandPlaceholder. This is much cheaper than forward
referencing from a uniqued node. Change
ValueEnumerator::organizeMetadata to partition distinct nodes and
uniqued nodes to reduce the overhead of cycles broken by distinct nodes.
Mehdi measured this for me; this removes most of the RAUW from the
importing step of -flto=thin, even after a WIP patch that removes
string-based DITypeRefs (introducing many more cycles to the metadata
graph).
llvm-svn: 267276
Duncan P. N. Exon Smith [Sat, 23 Apr 2016 04:34:11 +0000 (04:34 +0000)]
Add #ifndef NDEBUG markers around EXPECT_DEATH after r267270
http://bb.pgr.jp/builders/cmake-llvm-x86_64-linux/builds/36076
llvm-svn: 267275
Teresa Johnson [Sat, 23 Apr 2016 04:31:20 +0000 (04:31 +0000)]
Address comments.
llvm-svn: 267274
Teresa Johnson [Sat, 23 Apr 2016 04:30:47 +0000 (04:30 +0000)]
Refactor bitcode writer into classes (NFC)
Summary:
As discussed in on the mailing list yesterday, I have refactored
BitcodeWriter.cpp to use classes to manage the bitcode writing process,
instead of passing around long lists of parameters between static
functions. See:
http://lists.llvm.org/pipermail/llvm-dev/2016-April/098610.html
I created a parent BitcodeWriter class to own the BitstreamWriter,
write the header, and contain the main entry point into the writing
process. There are two derived classes, one for writing a module and one
for writing a combined index file (for ThinLTO), which manage the
writing process specific to those bitcode file types.
I also changed the functions to conform to LLVM coding standards
(lowercase function name first letter). The only two routines that still
start with an uppercase letter are the two external interfaces, which
can be fixed as a follow-on (I wanted to keep this round just within
BitcodeWriter.cpp).
Reviewers: dexonsmith, joker.eph
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D19447
llvm-svn: 267273
Duncan P. N. Exon Smith [Sat, 23 Apr 2016 04:23:57 +0000 (04:23 +0000)]
Avoid ternery statement to please g++ after r267270, NFC
http://bb.pgr.jp/builders/cmake-llvm-x86_64-linux/builds/36074
llvm-svn: 267272
Duncan P. N. Exon Smith [Sat, 23 Apr 2016 04:22:38 +0000 (04:22 +0000)]
ValueEnumerator: Use std::find_if, NFC
Mehdi's pattern recognition pulled this one out. This is cleaner with
std::find_if than with the strange helper function that took an iterator
by reference and updated it.
llvm-svn: 267271
Duncan P. N. Exon Smith [Sat, 23 Apr 2016 04:15:56 +0000 (04:15 +0000)]
BitcodeReader: Avoid referencing unresolved nodes from distinct ones
Each reference to an unresolved MDNode is expensive, since the RAUW
support in MDNode uses a separate allocation and side map. Since
a distinct MDNode doesn't require its operands on creation (unlike
uniuqed nodes, there's no need to check for structural equivalence),
use nullptr for any of its unresolved operands. Besides reducing the
burden on MDNode maps, this can avoid allocating temporary MDNodes in
the first place.
We need some way to track operands. Invent DistinctMDOperandPlaceholder
for this purpose, which is a Metadata subclass that holds an ID and
points at its single user. DistinctMDOperandPlaceholder::replaceUseWith
is just like RAUW, but its name highlights that there is only ever
exactly one use.
There is no support for moving (or, obviously, copying) these. Move
support would be possible but expensive; leaving it unimplemented
prevents user error. In the BitcodeReader I originally considered
allocating on a BumpPtrAllocator and keeping a vector of pointers to
them, and then I realized that std::deque implements exactly this.
A couple of obvious follow-ups:
- Change ValueEnumerator to emit distinct nodes first to take more
advantage of this optimization. (How convenient... I think I might
have a couple of patches for this.)
- Change DIBuilder and its consumers (like CGDebugInfo in clang) to
use something like this when constructing debug info in the first
place.
llvm-svn: 267270
Duncan P. N. Exon Smith [Sat, 23 Apr 2016 04:02:39 +0000 (04:02 +0000)]
Tag the end of an anonymous namespace, NFC
Prevent clang-format from moving the closing branch.
llvm-svn: 267269
Duncan P. N. Exon Smith [Sat, 23 Apr 2016 04:01:57 +0000 (04:01 +0000)]
BitcodeReader: Consistently use IsDistinct, NFC
Consistently use the IsDistinct variable and start relying on it in
GET_OR_DISTINCT. This change has NFC, but prepares for using IsDistinct
to optimize the behaviour of the getMD() and getMDOrNull() helpers.
llvm-svn: 267268
Duncan P. N. Exon Smith [Sat, 23 Apr 2016 03:55:14 +0000 (03:55 +0000)]
BitcodeReader: Use getMD/getMDOrNull helpers consistently, almost NFC
The only functionality change was removing an error check from the
BitcodeReader (and an assertion from DILocation::getImpl) that is
already caught by Verifier::visitDILocation. The Verifier is a better
place for this anyway, and being inconsistent with other subclasses of
MDNode isn't serving anyone.
llvm-svn: 267267