Adrian Prantl [Fri, 28 Jul 2017 20:21:02 +0000 (20:21 +0000)]
Remove the obsolete offset parameter from @llvm.dbg.value
There is no situation where this rarely-used argument cannot be
substituted with a DIExpression and removing it allows us to simplify
the DWARF backend. Note that this patch does not yet remove any of
the newly dead code.
[SLP] Allow vectorization of the instruction from the same basic blocks only, NFC.
Summary:
After some changes in SLP vectorizer we missed some additional checks to
limit the instructions for vectorization. We should not perform analysis
of the instructions if the parent of instruction is not the same as the
parent of the first instruction in the tree or it was analyzed already.
Fix conditional tail call branch folding when both edges are the same
The conditional tail call logic did the wrong thing when both
destinations of a conditional branch were the same:
BB#1: derived from LLVM BB %entry
Live Ins: %EFLAGS
Predecessors according to CFG: BB#0
JE_1 <BB#5>, %EFLAGS<imp-use,kill>
JMP_1 <BB#5>
BB#5: derived from LLVM BB %sw.epilog
Predecessors according to CFG: BB#1
TCRETURNdi64 <ga:@mergeable_conditional_tailcall>, 0, ...
We would fold the JE_1 to a TCRETURNdi64cc, and then remove our BB#5
successor. Then BB#5 would be deleted as it had no predecessors, leaving
a dangling "JMP_1 <BB#5>" reference behind to cause assertions later.
This patch checks that both conditional branch destinations are
different before doing the transform. The standard branch folding logic
is able to remove both the JMP_1 and the JE_1, and for my test case we
end up forming a better conditional tail call later.
Matt Arsenault [Fri, 28 Jul 2017 19:06:16 +0000 (19:06 +0000)]
AMDGPU: Look through a bitcast user of an out argument
This allows handling of a lot more of the interesting
cases in Blender. Most of the large functions unlikely
to be inlined have this pattern.
This is a special case for what clang emits for OpenCL 3
element vectors. Annoyingly, these are emitted as
<3 x elt>* pointers, but accessed as <4 x elt>* operations.
This also needs to handle cases where a struct containing
a single vector is used.
Matt Arsenault [Fri, 28 Jul 2017 18:40:05 +0000 (18:40 +0000)]
AMDGPU: Add pass to replace out arguments
It is better to return arguments directly in registers
if we are making a call rather than introducing expensive
stack usage. In one of sample compile from one of
Blender's many kernel variants, this fires on about
~20 different functions. Future improvements may be to
recognize simple cases where the pointer is indexing a small
array. This also fails when the store to the out argument
is in a separate block from the return, which happens in
a few of the Blender functions. This should also probably
be using MemorySSA which might help with that.
I'm not sure this is correct as a FunctionPass, but
MemoryDependenceAnalysis seems to not work with
a ModulePass.
I'm also not sure where it should run.I think it should
run before DeadArgumentElimination, so maybe either
EP_CGSCCOptimizerLate or EP_ScalarOptimizerLate.
[LVI] Constant-propagate a zero extension of the switch condition value through case edges
Summary:
LazyValueInfo currently computes the constant value of the switch condition through case edges, which allows the constant value to be propagated through the case edges.
But we have seen a case where a zero-extended value of the switch condition is used past case edges for which the constant propagation doesn't occur.
This patch adds a small logic to handle such a case in getEdgeValueLocal().
This is motivated by the Python 2.7 eval loop in PyEval_EvalFrameEx() where the lack of the constant propagation causes longer live ranges and more spill code than necessary.
With this patch, we see that the code size of PyEval_EvalFrameEx() decreases by ~5.4% and a performance test improves by ~4.6%.
Matt Arsenault [Fri, 28 Jul 2017 15:52:08 +0000 (15:52 +0000)]
AMDGPU: Annotate implicitarg.ptr usage
We need to pass something to functions for this to work.
It isn't derivable just from the kernarg segment pointer
because the implicit arguments are placed after the
kernel arguments.
Wei Mi [Fri, 28 Jul 2017 15:47:25 +0000 (15:47 +0000)]
[GVN] Recommit the patch "Add phi-translate support in scalarpre"
Recommit after workaround the bug PR31652.
Three bugs fixed in previous recommits: The first one is to use CurrentBlock
instead of PREInstr's Parent as param of performScalarPREInsertion because
the Parent of a clone instruction may be uninitialized. The second one is stop
PRE when CurrentBlock to its predecessor is a backedge and an operand of CurInst
is defined inside of CurrentBlock. The same value defined inside of loop in last
iteration can not be regarded as available. The third one is an out-of-bound
array access in a flipped if guard.
Right now scalarpre doesn't have phi-translate support, so it will miss some
simple pre opportunities. Like the following testcase, current scalarpre cannot
recognize the last "a * b" is fully redundent because a and b used by the last
"a * b" expr are both defined by phis.
long a[100], b[100], g1, g2, g3;
__attribute__((pure)) long goo();
void foo(long a, long b, long c, long d) {
g1 = a * b;
if (__builtin_expect(g2 > 3, 0)) {
a = c;
b = d;
g2 = a * b;
}
g3 = a * b; // fully redundant.
}
The patch adds phi-translate support in scalarpre. This is only a temporary
solution before the newpre based on newgvn is available.
Chris Bieneman [Fri, 28 Jul 2017 15:33:35 +0000 (15:33 +0000)]
[CMake] NFC. Add intrinsics_gen target to CMake Exports
By creating a dummy of this target in LLVMConfig.cmake, projects that can build against out-of-tree LLVM can freely depend on the target without needing to have conditionals for if LLVM is in-tree or out-of-tree.
Peter Smith [Fri, 28 Jul 2017 09:21:00 +0000 (09:21 +0000)]
[ARM] Add test to check pcs of ARM ABI runtime floating point helpers
The ARM Runtime ABI document (IHI0043) defines the AEABI floating point
helper functions in section 4.1.2 The floating-point helper functions.
The functions listed in this section must always use the base AAPCS calling
convention.
This test generates calls to all the helper functions that llvm supports
and checks that the base AAPCS calling convention has been used. We test
the equivalent of -mfloat-abi=soft, -mfloat-abi=softfp, -mfloat-abi=hardfp
with an FPU that supports single and double precision, and one that only
supports double precision.
Max Kazantsev [Fri, 28 Jul 2017 06:42:15 +0000 (06:42 +0000)]
[SCEV] Do not visit nodes twice in containsConstantSomewhere
This patch reworks the function that searches constants in Add and Mul SCEV expression
chains so that now it does not visit a node more than once, and also renames this function
for better correspondence between its implementation and semantics.
[MachineOutliner] NFC: Split up getOutliningBenefit
This is some more cleanup in preparation for some actual
functional changes. This splits getOutliningBenefit into
two cost functions: getOutliningCallOverhead and
getOutliningFrameOverhead. These functions return the
number of instructions that would be required to call
a specific function and the number of instructions
that would be required to construct a frame for a
specific funtion. The actual outlining benefit logic
is moved into the outliner, which calls these functions.
The goal of refactoring getOutliningBenefit is to:
- Get us closer to getting rid of the IsTailCall flag
- Further split up "target-specific" things and
"general algorithm" things
JumpThreading claims to preserve LVI, but it doesn't preserve
the analyses which LVI holds a reference to (e.g. the Dominator).
In the current pass manager infrastructure, after JT runs, the
PM frees these analyses (including DominatorTree) but preserves
LVI.
CorrelatedValuePropagation runs immediately after and queries
a corrupted domtree, causing weird miscompiles.
This commit disables the preservation of LVI for the time being.
Eventually, we should either move LVI to a proper dependency
tracking mechanism (i.e. an analyses shouldn't hold references
to other analyses and compute them on demand if needed), or
we should teach all the passes preserving LVI to preserve the
analyses LVI depends on.
The new pass manager has a mechanism to invalidate LVI in case
one of the analyses it depends on becomes invalid, so this problem
shouldn't exist (at least not in this immediate form), but handling
of analyses holding references is still a very delicate subject.
David Blaikie [Fri, 28 Jul 2017 03:06:25 +0000 (03:06 +0000)]
DebugInfo: Consider a CU containing only local imported entities to be 'empty'
This can come up in ThinLTO & wastes space & makes degenerate IR.
As per the added FIXME, ultimately, local imported entities should hang
off the function and that way the imported entity list on the CU can be
tested for emptiness like all the other CU lists.
(function-attached local imported entities are probably also the best
path forward for fixing how imported entities are handled both in
cross-module use (currently, while ThinLTO preserves the imported
entities, they would not get used at the imported inlined location -
only in the abstract origin that appears in the partial CU created by
the import (which isn't emitted under Fission due to cross-CU
limitations there)) and to reduce the number of points where imported
entities are emitted (they're currently emitted into every inlined
instance, concrete instance, and abstract origin - they should only go
in teh abstract origin if there is one, otherwise in the concrete
instance - but this requires lots of delayed handling and wiring up,
same as abstract variables & subprograms))
ARMFrameLowering: Only set ExtraCSSpill for actually unused registers.
The code assumed that unclobbered/unspilled callee saved registers are
unused in the function. This is not true for callee saved registers that are
also used to pass parameters such as swiftself.
Summary:
The technique of directly calling subprocess.Popen on a python script
doesn't work on Windows. The executable path of the command must refer
to a valid win32 executable.
Instead, rename all the python scripts masquerading as gtest executables
to have .py extensions, so we can easily detect then and call the python
executable for them. Do this on Linux as well as Windows for
consistency.
The test suite directory names also come out in lower-case on Windows.
We can consider removing that in a later patch. This change just updates
the FileCheck lines to match on Windows.
Changing the default MaxNumPromotions from 2 to 3.
Summary: In performance tuning, we see performance benefits when enlarge the maximum num promotion targets to 3. This is safe as soon as we have total percentage threshold properly setup (https://reviews.llvm.org/D35962)
Separate the ICP total threshold and remaining threshold.
Summary: In the current implementation, isPromotionProfitable only checks if the call count to a direct target is no less than a certain percentage threshold of the remaining call counts that have not been promoted. This causes code size problems when the target count is small but greater than a large portion of remaining counts. E.g. target1 takes 99.9%, while target2 takes 0.1%. Both targets will be promoted and inlined, makes the function size too large, which potentially prevents it from further inlining into its callers. This patch adds another percentage threshold against the total indirect call count. If the target count needs to be no less than both thresholds in order to be promoted speculatively.
Summary: The original 3.0 hot mupltiplier is too small, and would prevent hot callsites from being inline. This patch increases the hot multilier to 10.0
The X86 tail call eligibility logic was correct when it was written, but
the addition of inalloca and argument copy elision broke its
assumptions. It was assuming that fixed stack objects were immutable.
Currently, we aim to emit a tail call if no arguments have to be
re-arranged in memory. This code would trace the outgoing argument
values back to check if they are loads from an incoming stack object.
If the stack argument is immutable, then we won't need to store it back
to the stack when we tail call.
Fortunately, stack objects track their mutability, so we can just make
the obvious check to fix the bug.
[sanitizer-coverage] add a feature sanitizer-coverage-create-pc-table=1 (works with trace-pc-guard and inline-8bit-counters) that adds a static table of instrumented PCs to be used at run-time
[MachineOutliner] Cleanup: move findCandidates out of suffix tree
Doing some cleanup in preparation for some functional changes.
This commit moves findCandidates out of the suffix tree and into the
MachineOutliner class. This is much easier to follow, and removes
the burden of candidate choice from the suffix tree.
It also adds a couple FIXMEs and simplifies building outlined function
names.
[PDB] Initialize the std::array<ulittle32_t> used for the gsi bitmap
With ASan, we would write about 512 bytes of malloc fill value to the
PDB, with some random bits ORed in here and there. Dumping the PDB would
always fail reliably.
Davide Italiano [Thu, 27 Jul 2017 22:20:44 +0000 (22:20 +0000)]
[ConstantFolder] Don't try to fold gep when the idx is a vector.
The code in ConstantFoldGetElementPtr() assumes integers, and
therefore it crashes trying to get the integer bidwith of a vector
type (in this case <4 x i32>. I just changed the code to prevent
the folding in case of vectors and I didn't bother to generalize
as this doesn't seem to me something that really happens in
practice, but I'm willing to change the patch if you think
it's worth it.
This is hard to trigger from -instsimplify or -instcombine
only as the second instruction is dead, so the test uses loop-unroll.
Ahmed Bougacha [Thu, 27 Jul 2017 21:28:59 +0000 (21:28 +0000)]
[X86] Don't lie about legality to TLI's demanded bits.
Like r309323, X86 had a typo where it passed the wrong flags to TLO.
Found by inspection; I haven't been able to tickle this into having
observable behavior. I don't think it does, given that X86 doesn't have
custom demanded bits logic, and the generic logic doesn't have a lot of
exposure to illegal constructs.
Ahmed Bougacha [Thu, 27 Jul 2017 21:27:25 +0000 (21:27 +0000)]
[AArch64] Fix legality info passed to demanded bits for TBI opt.
The (seldom-used) TBI-aware optimization had a typo lying dormant since
it was first introduced, in r252573: when asking for demanded bits, it
told TLI that it was running after legalize, where the opposite was
true.
This is an important piece of information, that the demanded bits
analysis uses to make assumptions about the node. r301019 added such an
assumption, which was broken by the TBI combine.
Michal Gorny [Thu, 27 Jul 2017 21:13:25 +0000 (21:13 +0000)]
[OCaml] Fix undefined reference to LLVMDumpType() with NDEBUG
Account for the possibility of LLVMDumpType() not being available with
NDEBUG in the OCaml bindings. If it is not built into LLVM, make
the dump function raise an exception.
Since rL293359, the dump functions are built only if either NDEBUG is
not defined, or LLVM_ENABLE_DUMP is defined. As a result, if the dump
functions are not built in LLVM, the dynamic OCaml libraries fail to
load due to undefined LLVMDumpType symbol.
Michal Gorny [Thu, 27 Jul 2017 21:13:19 +0000 (21:13 +0000)]
[OCaml] Respect CMAKE_C_FLAGS for OCaml C files
Pass the values of CMAKE_C_FLAGS and CMAKE_C_FLAGS_${CMAKE_BUILD_TYPE}
as -ccopt to ocamlc. This enforces the specific flags used for the LLVM
build to be used for OCaml bindings as well, notably -O and -march
flags.
This also solves the issue of the user being unable to force specific
flags for OCaml bindings builds. Gentoo needs this to enforce -DNDEBUG
consistently between the LLVM build and the split OCaml bindings build.
[ARM] Add use-misched feature, to enable the MachineScheduler.
Summary:
This change makes it easier to experiment with the MachineScheduler in
the ARM backend and also makes it very explicit which CPUs use the
MachineScheduler (currently only swift and cyclone).
The alias support was dead code since 2011. It was last touched
in r124182, where it was reintroduced after being removed
in r110434, and since then it was gated behind a HasGlobalAliases
flag that was permanently stuck as `false`.
It is also broken. I'm not sure if it bitrotted or was just broken
in the first place because it appears to have never been tested,
but the following IR results in a crash:
Brian Gesiak [Thu, 27 Jul 2017 19:27:10 +0000 (19:27 +0000)]
[lit] Fix TestRunner unit test on Windows
Summary:
Normally Python converts all newline characters, Windows or Unix,
to Unix newlines when opening a file. However, lit opens files in
binary mode, which does not perform this conversion. As a result,
trailing Windows newlines are not stripped from test input, which
caused a failure in the TestRunner unit test:
```
FAIL: test_custom (__main__.TestIntegratedTestKeywordParser)
----------------------------------------------------------------------
Traceback (most recent call last):
File "C:\Users\bgesiak\src\llvm\llvm\utils\lit\tests\unit\TestRunner.py", line 109, in test_custom
self.assertItemsEqual(value, ['a', 'b', 'c'])
AssertionError: Element counts were not equal:
First has 1, Second has 0: 'c\r'
First has 0, Second has 1: 'c'
```
Fix the discrepancy in behavior across the two platforms by
manually stripping Windows newlines before yielding each line in
the test file.
Brian Gesiak [Thu, 27 Jul 2017 19:18:35 +0000 (19:18 +0000)]
Un-revert "Teach the CMake build system to run lit's test suite. These can be run"
Summary:
Depends on https://reviews.llvm.org/D35879.
This reverts rL257268, which in turn was a revert of rL257221.
https://reviews.llvm.org/D35879 marks the tests in the lit test suite
that fail on Windows as XFAIL, which should allow these tests to pass
on Windows-based buildbots.
[InstCombine] Simplify pointer difference subtractions (GEP-GEP) where GEPs have other uses and one non-constant index
Summary:
Pointer difference simplifications currently happen only if input GEPs don't have other uses or their indexes are all constants, to avoid duplicating indexing arithmetic.
This patch enables cases with exactly one non-constant index among input GEPs to happen where there is no duplicated arithmetic or code size increase even if input GEPs have other uses.
For example, this patch allows "(&A[42][i]-&A[42][0])" --> "i", which didn't happen previously, if the input GEP(s) have other uses.
Improve DAGTypeLegalizer::convertMask's isSETCCorConvertedSETCC assertion to properly check for any mixture of SETCC or BUILD_VECTOR of constants, or a logical mask op of them.
Adam Nemet [Thu, 27 Jul 2017 16:54:15 +0000 (16:54 +0000)]
[ICP] Migrate to OptimizationRemarkEmitter
This is a module pass so for the old PM, we can't use ORE, the function
analysis pass. Instead ORE is created on the fly.
A few notes:
- isPromotionLegal is folded in the caller since we want to emit the Function
in the remark but we can only do that if the symbol table look-up succeeded.
- There was good test coverage for remarks in this pass.
- promoteIndirectCall uses ORE conditionally since it's also used from
SampleProfile which does not use ORE yet.
Brian Gesiak [Thu, 27 Jul 2017 16:50:40 +0000 (16:50 +0000)]
[lit] Fix order of checks in shtest-shell.py test
Summary:
An expectation in `utils/lit/tests/Inputs/shtest-shell/redirects.txt`
expects that first a string printed to stdout is seen, and then a
string printed to stderr. Add `flush()` calls to ensure that stdout is
printed before stderr, as expected.
Daniel Neilson [Thu, 27 Jul 2017 16:49:39 +0000 (16:49 +0000)]
All libcalls should be considered to be GC-leaf functions.
Summary:
It is possible for some passes to materialize a call to a libcall (ex: ldexp, exp2, etc),
but these passes will not mark the call as a gc-leaf-function. All libcalls are
actually gc-leaf-functions, so we change llvm::callsGCLeafFunction() to tell us that
available libcalls are equivalent to gc-leaf-function calls.
[TargetParser] Use enum classes for various ARM kind enums.
Summary:
Using c++11 enum classes ensures that only valid enum values are used
for ArchKind, ProfileKind, VersionKind and ISAKind. This removes the
need for checks that the provided values map to a proper enum value,
allows us to get rid of AK_LAST and prevents comparing values from
different enums. It also removes a bunch of static_cast
from unsigned to enum values and vice versa, at the cost of introducing
static casts to access AArch64ARCHNames and ARMARCHNames by ArchKind.
FPUKind and ArchExtKind are the only remaining old-style enum in
TargetParser.h. I think it's beneficial to keep ArchExtKind as old-style
enum, but FPUKind can be converted too, but this patch is quite big, so
could do this in a follow-up patch. I could also split this patch up a
bit, if people would prefer that.
David Blaikie [Thu, 27 Jul 2017 15:09:06 +0000 (15:09 +0000)]
ThinLTO: Don't import aliases of any kind (even linkonce_odr)
Summary:
Until a more advanced version of importing can be implemented for
aliases (one that imports an alias as an available_externally definition
of the aliasee), skip the narrow subset of cases that was possible but
came at a cost: aliases of linkonce_odr functions could be imported
because the linkonce_odr function could be safely duplicated from the
source module. This came/comes at the cost of not being able to 'home'
imported linkonce functions (they had to be emitted linkonce_odr in all
the destination modules (even if they weren't used by an alias) rather
than as available_externally - causing extra object size).
Tangentially, this also was the only reason ThinLTO would emit multiple
CUs in to the resulting DWARF - which happens to be a problem for
Fission (there's a fix for this in GDB but not released yet, etc).
(actually it's not the only reason - but I'm sending a patch to fix the
other reason shortly)
There's no reason to believe this particularly narrow alias importing
was especially/meaningfully important, only that it was /possible/ to
implement in this way. When a more general solution is done, it should
still satisfy the DWARF concerns above, since the import will still be
available_externally, and thus not create extra CUs.
Since now all aliases are treated the same, I removed/simplified some
test cases since they were testing corner cases where there are no
longer any corners.
[ARM] Mark labels in skipAlignedDPRCS2Spills as fallthrough (NFC).
The comment at the top of the switch statement indicates that the
fall-through behavior is intentional. By using LLVM_FALLTHROUGH,
-Wimplicit-fallthrough are silenced, which is enabled by default in GCC
7.
[InlineCost, NFC] Change CallAnalyzer::isGEPFree to use TTI::getUserCost instead of TTI::getGEPCost
Currently CallAnalyzer::isGEPFree uses TTI::getGEPCost to check if GEP is free.
TTI::getGEPCost cannot handle cases when GEPs participate in Def-Use dependencies
(see https://reviews.llvm.org/D31186 for example).
There is TTI::getUserCost which can calculate the cost more accurately by
taking dependencies into account.
Daniel Sanders [Thu, 27 Jul 2017 11:03:45 +0000 (11:03 +0000)]
Re-commit: r309094 [globalisel][tablegen] Fuse the generated tables together.
Summary:
Now that we have control flow in place, fuse the per-rule tables into a
single table. This is a compile-time saving at this point. However, this will
also enable the optimization of a table so that similar instructions can be
tested together, reducing the time spent on the matching the code.
This is NFC in terms of externally visible behaviour but some internals have
changed slightly. State.MIs is no longer reset between each rule that is
attempted because it's not necessary to do so. As a consequence of this the
restriction on the order that instructions are added to State.MIs has been
relaxed to only affect recorded instructions that require new elements to be
added to the vector. GIM_RecordInsn can now write to any element from 1 to
State.MIs.size() instead of just State.MIs.size().
The compile-time regressions from the last commit were caused by the ARM target
including a non-const variable (zero_reg) in the table and therefore generating
an initializer for it. That variable is now const.
[TTI] fixing a bug in the isLegalMaskedScatter API
isLegalMaskedScatter called the Gather version which is a bug.
use test case is provided within the patch of AVX2 gathers at: https://reviews.llvm.org/D35772
[PowerPC] enable optimizeCompareInstr for branch with static branch hint
In optimizeCompareInstr, a compare instruction is eliminated by using a record form instruction if possible.
If the branch instruction that uses the result of the compare has a static branch hint, the optimization does not happen.
This patch makes this optimization happen regardless of the branch hint by splitting branch hint and branch condition before checking the predicate to identify the possible optimizations.
Petr Hosek [Thu, 27 Jul 2017 04:35:30 +0000 (04:35 +0000)]
Reland "[LLVM][llvm-objcopy] Added basic plumbing to get things started"
As discussed on llvm-dev I've implemented the first basic steps towards
llvm-objcopy/llvm-objtool (name pending).
This change adds the ability to copy (without modification) 64-bit
little endian ELF executables that have SHT_PROGBITS, SHT_NOBITS,
SHT_NULL and SHT_STRTAB sections.
David Blaikie [Thu, 27 Jul 2017 00:06:53 +0000 (00:06 +0000)]
DebugInfo: Ensure imported entities at the top level of an inlined function don't cause degenerate concrete definitions
Local imported entities at the top level of a subprogram were being
handled differently from those in nested scopes - that different
handling would cause pseudo concrete out-of-line definitions to be
created (but without any of their attributes, nor an abstract_origin) in
the case where there was no real concrete definition.
These local imported entities also only appeared in the concrete
definition where those imported entities in nested scopes appear in all
cases (abstract, concrete, and inlined). This change at least makes top
level case handle the same as the others - though there's a FIXME to
improve this to /only/ emit them into the abstract origin (though this
requires more plumbing - like the abstract subprogram and variable
handling that must defer population until the end of the unit to
discover if there is an abstract origin, or only a standalone concrete
definition).
projects / llvm / log