This patch adds an optimization in CodeGenPrepare to move an extractelement
right before a store when the target can combine them.
The optimization may promote any scalar operations to vector operations in the
way to make that possible.
** Context **
Some targets use different register files for both vector and scalar operations.
This means that transitioning from one domain to another may incur copy from one
register file to another. These copies are not coalescable and may be expensive.
For example, according to the scheduling model, on cortex-A8 a vector to GPR
move is 20 cycles.
** Motivating Example **
Let us consider an example:
define void @foo(<2 x i32>* %addr1, i32* %dest) {
%in1 = load <2 x i32>* %addr1, align 8
%extract = extractelement <2 x i32> %in1, i32 1
%out = or i32 %extract, 1
store i32 %out, i32* %dest, align 4
ret void
}
As it is, this IR generates the following assembly on armv7:
vldr d16, [r0] @vector load
vmov.32 r0, d16[1] @ cross-register-file copy: 20 cycles
orr r0, r0, #1 @ scalar bitwise or
str r0, [r1] @ scalar store
bx lr
Whereas we could generate much faster code:
vldr d16, [r0] @ vector load
vorr.i32 d16, #0x1 @ vector bitwise or
vst1.32 {d16[1]}, [r1:32] @ vector extract + store
bx lr
Half of the computation made in the vector is useless, but this allows to get
rid of the expensive cross-register-file copy.
** Proposed Solution **
To avoid this cross-register-copy penalty, we promote the scalar operations to
vector operations. The penalty will be removed if we manage to promote the whole
chain of computation in the vector domain.
Currently, we do that only when the chain of computation ends by a store and the
target is able to combine an extract with a store.
Stores are the most likely candidates, because other instructions produce values
that would need to be promoted and so, extracted as some point[1]. Moreover,
this is customary that targets feature stores that perform a vector extract (see
AArch64 and X86 for instance).
The proposed implementation relies on the TargetTransformInfo to decide whether
or not it is beneficial to promote a chain of computation in the vector domain.
Unfortunately, this interface is rather inaccurate for this level of details and
although this optimization may be beneficial for X86 and AArch64, the inaccuracy
will lead to the optimization being too aggressive.
Basically in TargetTransformInfo, everything that is legal has a cost of 1,
whereas, even if a vector type is legal, usually a vector operation is slightly
more expensive than its scalar counterpart. That will lead to too many
promotions that may not be counter balanced by the saving of the
cross-register-file copy. For instance, on AArch64 this penalty is just 4
cycles.
For now, the optimization is just enabled for ARM prior than v8, since those
processors have a larger penalty on cross-register-file copies, and the scope is
limited to basic blocks. Because of these two factors, we limit the effects of
the inaccuracy. Indeed, I did not want to build up a fancy cost model with block
frequency and everything on top of that.
[1] We can imagine targets that can combine an extractelement with other
instructions than just stores. If we want to go into that direction, the current
interfaces must be augmented and, moreover, I think this becomes a global isel
problem.
Differential Revision: http://reviews.llvm.org/D5921
<rdar://problem/14170854>
llvm-svn: 220978
In a case where we have a no {un,}signed wrap flag on the increment, if
RHS - Start is constant then we can avoid inserting a max operation bewteen
the two, since we can statically determine which is greater.
This allows us to unroll loops such as:
void testcase3(int v) {
for (int i=v; i<=v+1; ++i)
f(i);
}
llvm-svn: 220960
Since block address values can be larger than 2GB in 64-bit code, they
cannot be loaded simply using an @l / @ha pair, but instead must be
loaded from the TOC, just like GlobalAddress, ConstantPool, and
JumpTable values are.
The commit also fixes a bug in PPCLinuxAsmPrinter::doFinalization where
temporary labels could not be used as TOC values, since code would
attempt (and fail) to use GetOrCreateSymbol to create a symbol of the
same name as the temporary label.
llvm-svn: 220959
Since JIT->MCJIT migration, most of the ExecutionEngine interface
became deprecated and/or broken. This especially affected the OCaml
bindings, as runFunction is no longer available, and unlike in C,
it is not possible to coerce a pointer to a function and call it
in OCaml.
In practice, LLVM 3.5 shipped completely unusable
Llvm_executionengine.
The GenericValue interface and runFunction were essentially
a poor man's FFI. As such, this interface was removed and instead
a dependency on ctypes >=0.3 added, which handled platform-specific
aspects of accessing data and calling functions.
The new interface does not expose JIT (which is a shim around MCJIT),
as well as the interpreter (which can't handle a lot of valid IR).
Llvm_executionengine.add_global_mapping is currently unusable
due to PR20656.
llvm-svn: 220957
r212242 introduced a legalizer hook, originally to let AArch64 widen
v1i{32,16,8} rather than scalarize, because the legalizer expected, when
scalarizing the result of a conversion operation, to already have
scalarized the operands. On AArch64, v1i64 is legal, so that commit
ensured operations such as v1i32 = trunc v1i64 wouldn't assert.
It did that by choosing to widen v1 types whenever possible. However,
v1i1 types, for which there's no legal widened type, would still trigger
the assert.
This commit fixes that, by only scalarizing a trunc's result when the
operand has already been scalarized, and introducing an extract_elt
otherwise.
This is similar to r205625.
Fixes PR20777.
llvm-svn: 220937
Earlier this summer I fixed an issue where we were incorrectly combining
multiple loads that had different constraints such alignment, invariance,
temporality, etc. Apparently in one case I made copt paste error and swapped
alignment and invariance.
Tests included.
rdar://18816719
llvm-svn: 220933
The langref says:
LLVM explicitly allows declarations of global variables to be marked
constant, even if the final definition of the global is not. This
capability can be used to enable slightly better optimization of the
program, but requires the language definition to guarantee that
optimizations based on the ‘constantness’ are valid for the
translation units that do not include the definition.
Given that definition, when merging two declarations, we have to drop
constantness if of of them is not marked contant, since the Module
without the constant marker might not have the necessary guarantees.
llvm-svn: 220927
If we load from a location with range metadata, we can use information about the ranges of the loaded value for optimization purposes. This helps to remove redundant checks and canonicalize checks for other optimization passes. This particular patch checks whether a value is known to be non-zero from the range metadata.
Currently, these tests are against InstCombine. In theory, all of these should be InstSimplify since we're not inserting any new instructions. Moving the code may follow in a separate change.
Reviewed by: Hal
Differential Revision: http://reviews.llvm.org/D5947
llvm-svn: 220925
Summary:
This patch finishes up support for handling sampling profiles in both
text and binary formats. The new binary format uses uleb128 encoding to
represent numeric values. This makes profiles files about 25% smaller.
The profile writer class can write profiles in the existing text and the
new binary format. In subsequent patches, I will add the capability to
read (and perhaps write) profiles in the gcov format used by GCC.
Additionally, I will be adding support in llvm-profdata to manipulate
sampling profiles.
There was a bit of refactoring needed to separate some code that was in
the reader files, but is actually common to both the reader and writer.
The new test checks that reading the same profile encoded as text or
raw, produces the same results.
Reviewers: bogner, dexonsmith
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D6000
llvm-svn: 220915
Summary:
The previous calling convention prevented custom functions from being able
to access argument labels unless it knew how many variadic arguments there
were, and of which type. This restriction made it impossible to correctly
model functions in the printf family, as it is legal to pass more arguments
than required to those functions. We now pass arguments in the following order:
non-vararg arguments
labels for non-vararg arguments
[if vararg function, pointer to array of labels for vararg arguments]
[if non-void function, pointer to label for return value]
vararg arguments
Differential Revision: http://reviews.llvm.org/D6028
llvm-svn: 220906
Prior to this commit, the Llvm_target tests (ab)used
the Llvm_executionengine as a mechanism to initialize at least some
target. This needlessly restricted tests to builds which can emit
code for their host architecture.
llvm-svn: 220901
This commit updates the OCaml bindings and tests to use ocamlfind.
The bindings are migrated in order to use ctypes, which are now
required for MCJIT-backed Llvm_executionengine.
The tests are migrated in order to use OUnit and to verify that
the distributed META.llvm allows to build working executables.
Every OCaml toolchain invocation is now chained through ocamlfind,
which (in theory) allows to cross-compile the OCaml bindings.
The configure script now checks for ctypes (>= 0.2.3) and
OUnit (>= 2). The code depending on these libraries will be added
later. The configure script does not check the package versions
in order to keep changes less invasive.
Additionally, OCaml bindings will now be automatically enabled
if ocamlfind is detected on the system, rather than ocamlc, as it
was before.
llvm-svn: 220899
The also-emit-llvm option only supported getting the IR before optimizations.
This patch replaces it with a more generic save-temps option that saves the IR
both before and after optimizations.
llvm-svn: 220885
Summary: This helps llvm-objdump -r to print out the symbol name along
with the relocation type on x86. Adjust existing tests from checking
for "Unknown" to check for the symbol now.
Test Plan: Adjusted test/Object tests.
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D5987
llvm-svn: 220866
In practice this means:
* Always using -g flag.
* Embedding -cclib -lstdc++ into the corresponding cma/cmxa file.
This also moves -lstdc++ in a single place.
* Using caml_named_value instead of a homegrown mechanism.
llvm-svn: 220843
For example, MS PSDK is not expected to have <cxxabi.h>.
You should introduce the new feature in lit.cfg corresponding to HAVE_CXXABI_H if you would like to test demangler.
llvm-svn: 220840
Remove pointless checks for storage of uninteresting values. Ensure that we
perform basic alias analysis to make the test more correct. Finally, apply a
stylistic change to the test.
llvm-svn: 220839
Previously, tests hardcoded ocamlopt and cmxa, which broke builds on
machines without ocamlopt. Instead, they now fall back to ocamlc.
As a side effect this fixes PR14727, which was caused by a crude hack
that replaced gcc with g++ everywhere in the ocamlopt native compiler
path and passes it back using -cc. Now the tests use the same
technique as META, i.e. -cclib -lstdc++. It might be more fragile
than using g++ explicitly, but it will break when the installed
package will also break, which is good.
llvm-svn: 220828
This restores the commit from SVN r219899 with an additional change to ensure
that the CodeGen is correct for the case that was identified as being incorrect
(originally PR7272).
In the case that during inlining we need to synthesize a value on the stack
(i.e. for passing a value byval), then any function involving that alloca must
be stripped of its tailness as the restriction that it does not access the
parent's stack no longer holds. Unfortunately, a single alloca can cause a
rippling effect through out the inlining as the value may be aliased or may be
mutated through an escaped external call. As such, we simply track if an alloca
has been introduced in the frame during inlining, and strip any tail calls.
llvm-svn: 220811
This transformation worked if selector is produced by SETCC, however SETCC is needed only if we consider to swap operands. So I replaced SETCC check for this case.
Added tests for vselect of <X x i1> values.
llvm-svn: 220777
Ffter commit at rev219046 512-bit broadcasts lowering become non-optimal. Most of tests on broadcasting and embedded broadcasting were changed and they doesn’t produce efficient code.
Example below is from commit changes (it’s the first test from test/CodeGen/X86/avx512-vbroadcast.ll):
define <16 x i32> @_inreg16xi32(i32 %a) {
; CHECK-LABEL: _inreg16xi32:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpbroadcastd %edi, %zmm0
+; CHECK-NEXT: vmovd %edi, %xmm0
+; CHECK-NEXT: vpbroadcastd %xmm0, %ymm0
+; CHECK-NEXT: vinserti64x4 $1, %ymm0, %zmm0, %zmm0
; CHECK-NEXT: retq
%b = insertelement <16 x i32> undef, i32 %a, i32 0
%c = shufflevector <16 x i32> %b, <16 x i32> undef, <16 x i32> zeroinitializer
ret <16 x i32> %c
}
Here, 256-bit broadcast was generated instead of 512-bit one.
In this patch
1) I added vector-shuffle lowering through broadcasts
2) Removed asserts and branches likes because this is incorrect
- assert(Subtarget->hasDQI() && "We can only lower v8i64 with AVX-512-DQI");
3) Fixed lowering tests
llvm-svn: 220774
This is a Microsoft calling convention that supports both x86 and x86_64
subtargets. It passes vector and floating point arguments in XMM0-XMM5,
and passes them indirectly once they are consumed.
Homogenous vector aggregates of up to four elements can be passed in
sequential vector registers, but this part is not implemented in LLVM
and will be handled in Clang.
On 32-bit x86, it is similar to fastcall in that it uses ecx:edx as
integer register parameters and is callee cleanup. On x86_64, it
delegates to the normal win64 calling convention.
Reviewers: majnemer
Differential Revision: http://reviews.llvm.org/D5943
llvm-svn: 220745
Benchmarks have shown that it's harmless to the performance there, and having a
unified set of passes between the two cores where possible helps big.LITTLE
deployment.
Patch by Z. Zheng.
llvm-svn: 220744
For a call to not return in to the stackmap shadow, the shadow must end with the call.
To do this, we must insert any required nops *before* the call, and not after it.
llvm-svn: 220728
This is a minor change to use the immediate version when the operand is a null
value. This should get rid of an unnecessary 'mov' instruction in debug
builds and align the code more with the one generated by SelectionDAG.
This fixes rdar://problem/18785125.
llvm-svn: 220713
To avoid emitting too many nops, a stackmap shadow can include emitted instructions in the shadow, but these must not include branch targets.
A return from a call should count as a branch target as patching over the instructions after the call would lead to incorrect behaviour for threads currently making that call, when they return.
llvm-svn: 220710
The pattern matching for a 'ConstantInt' value was too restrictive. Checking for
a 'Constant' with a bull value is sufficient for using an 'cbz/cbnz' instruction.
This fixes rdar://problem/18784732.
llvm-svn: 220709
This fixes a bug where the input register was not defined for the 'tbz/tbnz'
instruction. This happened, because we folded the 'and' instruction from a
different basic block.
This fixes rdar://problem/18784013.
llvm-svn: 220704
At higher optimization levels the LLVM IR may contain more complex patterns for
loads/stores from/to frame indices. The 'computeAddress' function wasn't able to
handle this and triggered an assertion.
This fix extends the possible addressing modes for frame indices.
This fixes rdar://problem/18783298.
llvm-svn: 220700
Currently, the ARM backend will select the VMAXNM and VMINNM for these C
expressions:
(a < b) ? a : b
(a > b) ? a : b
but not these expressions:
(a > b) ? b : a
(a < b) ? b : a
This patch allows all of these expressions to be matched.
llvm-svn: 220671
First, return true on success, as it is the OCaml convention.
Second, also initialize the native assembly printer, which is,
despite the name, required for MCJIT operation.
Since this function did not initialize the assembly printer earlier
and no function to initialize native assembly printer was available
elsewhere, it is safe to break its interface: it means that it
simply could not be used successfully before.
llvm-svn: 220620
The dividend in "signed % unsigned" is treated as unsigned instead of signed,
causing unexpected behavior such as -64 % (uint64_t)24 == 0.
Added a regression test in split-gep.ll
Patched by Hao Liu.
llvm-svn: 220618
The two operands of the new OR expression should be NextInChain and TheOther
instead of the two original operands.
Added a regression test in split-gep.ll.
Hao Liu reported this bug, and provded the test case and an initial patch.
Thanks!
llvm-svn: 220615
Summary:
Fixes PR21100 which is caused by inconsistency between the declared return type
and the expected return type at the call site. The new behavior is consistent
with nvcc and the NVPTXTargetLowering::getPrototype function.
Test Plan: test/Codegen/NVPTX/vector-return.ll
Reviewers: jholewinski
Reviewed By: jholewinski
Subscribers: llvm-commits, meheff, eliben, jholewinski
Differential Revision: http://reviews.llvm.org/D5612
llvm-svn: 220607
In a Mach-O object file a relocatable expression of the form
SymbolA - SymbolB + constant is allowed when both symbols are
defined in a section. But when either symbol is undefined it
is an error.
The code was crashing when it had an undefined symbol in this case.
And should have printed a error message using the location information
in the relocation entry.
rdar://18678402
llvm-svn: 220599
Minor patch to fix an issue in XFormVExtractWithShuffleIntoLoad where a load is unary shuffled, then bitcast (to a type with the same number of elements) before extracting an element.
An undef was created for the second shuffle operand using the original (post-bitcasted) vector type instead of the pre-bitcasted type like the rest of the shuffle node - this was then causing an assertion on the different types later on inside SelectionDAG::getVectorShuffle.
Differential Revision: http://reviews.llvm.org/D5917
llvm-svn: 220592
Modified library structure to deal with circular dependency between HexagonInstPrinter and HexagonMCInst.
Adding encoding bits for add opcode.
Adding llvm-mc tests.
Removing unit tests.
http://reviews.llvm.org/D5624
llvm-svn: 220584
This is a first step for generating SSE rsqrt instructions for
reciprocal square root calcs when fast-math is allowed.
For now, be conservative and only enable this for AMD btver2
where performance improves significantly - for example, 29%
on llvm/projects/test-suite/SingleSource/Benchmarks/BenchmarkGame/n-body.c
(if we convert the data type to single-precision float).
This patch adds a two constant version of the Newton-Raphson
refinement algorithm to DAGCombiner that can be selected by any target
via a parameter returned by getRsqrtEstimate()..
See PR20900 for more details:
http://llvm.org/bugs/show_bug.cgi?id=20900
Differential Revision: http://reviews.llvm.org/D5658
llvm-svn: 220570
Summary:
Most structs were fixed by r218451 but those of between >32-bits and
<64-bits remained broken since they were not marked with [ASZ]ExtUpper.
This patch fixes the remaining cases by using
CCPromoteToUpperBitsInType<i64> on i64's in addition to i32 and smaller.
Reviewers: vmedic
Reviewed By: vmedic
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D5963
llvm-svn: 220556
This fixes a miscompilation in the AArch64 fast-isel which was
triggered when a branch is based on an icmp with condition eq or ne,
and type i1, i8 or i16. The cbz instruction compares the whole 32-bit
register, so values with the bottom 1, 8 or 16 bits clear would cause
the wrong branch to be taken.
llvm-svn: 220553
This is asm/diasm-only support, similar to AVX.
For ISeling the register variant, they are no different from 213 other than
whether the multiplication or the addition operand is destructed.
For ISeling the memory variant, i.e. to fold a load, they are no different
than the 132 variant. The addition operand (op3) in both cases can come from
memory. Again the ony difference is which operand is destructed.
There could be a post-RA pass that would convert a 213 or 132 into a 231.
Part of <rdar://problem/17082571>
llvm-svn: 220540
This adds support for legalization of instructions of the form:
[fp_conv] <1 x i1> %op to <1 x double>
where fp_conv is one of fpto[us]i, [us]itofp. This used to assert
because they were simply missing from the vector operand scalarizer.
A similar problem arose in r190830, with trunc instead.
Fixes PR20778.
Differential Revision: http://reviews.llvm.org/D5810
llvm-svn: 220533
x86's CMPXCHG -> EFLAGS consumer wasn't being recorded as a real EFLAGS
dependency because it was represented by a pair of CopyFromReg(EFLAGS) ->
CopyToReg(EFLAGS) nodes. ScheduleDAG was expecting the source to be an
implicit-def on the instruction, where the result numbers in the DAG and the
Uses list in TableGen matched up precisely.
The Copy notation seems much more robust, so this patch extends ScheduleDAG
rather than refactoring x86.
Should fix PR20376.
llvm-svn: 220529
While refactoring this code I was confused by both the name I had
introduced (addNonArgumentVariable... but it has all this logic to
handle argument numbering and keep things in order?) and by the
redundancy. Seems when I fixed the misordered inlined argument handling,
I didn't realize it was mostly redundant with the argument ordering code
(which I may've also written, I'm not sure). So let's just rely on the
more general case.
The only oddity in output this produces is that it means when we emit
all the variables for the current function, we don't track when we've
finished the argument variables and are about to start the local
variables and insert DW_AT_unspecified_parameters (for varargs
functions) there. Instead it ends up after the local variables, scopes,
etc. But this isn't invalid and doesn't cause DWARF consumers problems
that I know of... so we'll just go with that because it makes the code
nice & simple.
(though, let's see what the buildbots have to say about this - *crosses
fingers*)
There will be some cleanup commits to follow to remove the now trivial
wrappers, etc.
llvm-svn: 220527
Currently, @llvm.smul.with.overflow.i8 expands to 9 instructions, where
3 are really needed.
This adds X86ISD::UMUL8/SMUL8 SD nodes, and custom lowers them to
MUL8/IMUL8 + SETO.
i8 is a special case because there is no two/three operand variants of
(I)MUL8, so the first operand and return value need to go in AL/AX.
Also, we can't write patterns for these instructions: TableGen refuses
patterns where output operands don't match SDNode results. In this case,
instructions where the output operand is an implicitly defined register.
A related special case (and FIXME) exists for MUL8 (X86InstrArith.td):
// FIXME: Used for 8-bit mul, ignore result upper 8 bits.
// This probably ought to be moved to a def : Pat<> if the
// syntax can be accepted.
[(set AL, (mul AL, GR8:$src)), (implicit EFLAGS)]
Ideally, these go away with UMUL8, but we still need to improve TableGen
support of implicit operands in patterns.
Before this change:
movsbl %sil, %eax
movsbl %dil, %ecx
imull %eax, %ecx
movb %cl, %al
sarb $7, %al
movzbl %al, %eax
movzbl %ch, %esi
cmpl %eax, %esi
setne %al
After:
movb %dil, %al
imulb %sil
seto %al
Also, remove a made-redundant testcase for PR19858, and enable more FastISel
ALU-overflow tests for SelectionDAG too.
Differential Revision: http://reviews.llvm.org/D5809
llvm-svn: 220516
This patch removes a chunk of special case logic for folding
(float)sqrt((double)x) -> sqrtf(x)
in InstCombineCasts and handles it in the mainstream path of SimplifyLibCalls.
No functional change intended, but I loosened the restriction on the existing
sqrt testcases to allow for this optimization even without unsafe-fp-math because
that's the existing behavior.
I also added a missing test case for not shrinking the llvm.sqrt.f64 intrinsic
in case the result is used as a double.
Differential Revision: http://reviews.llvm.org/D5919
llvm-svn: 220514
This fixes a bug (introduced by fixing the IR emitted from Clang where
the definition of a static member would be scoped within the class,
rather than within its lexical decl context) where the definition of a
static variable would be placed inside a class.
It also improves source fidelity by scoping static class member
definitions inside the lexical decl context in which tehy are written
(eg: namespace n { class foo { static int i; } int foo::i; } - the
definition of 'i' will be within the namespace 'n' in the DWARF output
now).
Lastly, and the original goal, this reduces debug info size slightly
(and makes debug info easier to read, etc) by placing the definitions of
non-member global variables within their namespace, rather than using a
separate namespace-scoped declaration along with a definition at global
scope.
Based on patches and discussion with Frédéric.
llvm-svn: 220497
This reverts commit r214981.
I'm not sure what I was thinking when I wrote this. Testing with MSVC
shows that this function is mangled to '@f@8':
int __fastcall f(int a, int b);
llvm-svn: 220492
This updates check for double precision zero floating point constant to allow
use of instruction with immediate value rather than temporary register.
Currently "a == 0.0", where "a" is of "double" type generates:
vmov.i32 d16, #0x0
vcmpe.f64 d0, d16
With this change it becomes:
vcmpe.f64 d0, #0
Patch by Sergey Dmitrouk.
llvm-svn: 220486
Currently, the ARM disassembler will disassemble the Thumb2 memory hint
instructions (PLD, PLDW and PLI), even for targets which do not have
these instructions. This patch adds the required checks to the
disassmebler.
llvm-svn: 220472
This tool lets us build LLVM components within the tree by setting up a
$GOPATH that resembles a tree fetched in the normal way with "go get".
It is intended that components such as the Go frontend will be built in-tree
using this tool.
Differential Revision: http://reviews.llvm.org/D5902
llvm-svn: 220462
Summary:
Currently when emitting a label, a new data fragment is created for it if the
current fragment isn't a data fragment.
This change instead enqueues the label and attaches it to the next fragment
(e.g. created for the next instruction) if possible.
When bundle alignment is not enabled, this has no functionality change (it
just results in fewer extra fragments being created). For bundle alignment,
previously labels would point to the beginning of the bundle padding instead
of the beginning of the emitted instruction. This was not only less efficient
(e.g. jumping to the nops instead of past them) but also led to miscalculation
of the address of the GOT (since MC uses a label difference rather than
emitting a "." symbol).
Fixes https://code.google.com/p/nativeclient/issues/detail?id=3982
Test Plan: regression test attached
Reviewers: jvoung, eliben
Subscribers: jfb, llvm-commits
Differential Revision: http://reviews.llvm.org/D5915
llvm-svn: 220439
This has been implement using the MCTargetStreamer interface as is done in the
ARM, Mips and PPC backends.
Phabricator: http://reviews.llvm.org/D5891
PR20964
llvm-svn: 220422
Jenkins likes to use directories with names involving the '@'
character, which breaks the sed expression in this test. Switch to use
'|' on the assumption that it's less likely to show up in a path.
llvm-svn: 220401
A previous patch enabled SELECT_VSRC and SELECT_CC_VSRC for VSX to
handle <2 x double> cases. This patch adds SELECT_VSFRC and
SELECT_CC_VSFRC to allow use of all 64 vector-scalar registers for the
f64 type when VSX is enabled. The changes are analogous to those in
the previous patch. I've added a new variant to vsx.ll to test the
code generation.
(I also cleaned up a little formatting in PPCInstrVSX.td from the
previous patch.)
llvm-svn: 220395
When a call to a double-precision libm function has fast-math semantics
(via function attribute for now because there is no IR-level FMF on calls),
we can avoid fpext/fptrunc operations and use the float version of the call
if the input and output are both float.
We already do this optimization using a command-line option; this patch just
adds the ability for fast-math to use the existing functionality.
I moved the cl::opt from InstructionCombining into SimplifyLibCalls because
it's only ever used internally to that class.
Modified the existing test cases to use the unsafe-fp-math attribute rather
than repeating all tests.
This patch should solve: http://llvm.org/bugs/show_bug.cgi?id=17850
Differential Revision: http://reviews.llvm.org/D5893
llvm-svn: 220390
When the profile for a function cannot be applied, we use to emit an
error. This seems extreme. The compiler can continue, it's just that the
optimization opportunities won't include profile information.
llvm-svn: 220386
Summary:
When using a profile, we used to require the use -gmlt so that we could
get access to the line locations. This is used to match line numbers in
the input profile to the line numbers in the function's IR.
But this is actually not necessary. The driver can provide source
location tracking without the emission of debug information. In these
cases, the annotation 'llvm.dbg.cu' is missing from the IR, but the
actual line location annotations are still present.
This patch adds a new way of looking for the start of the current
function. Instead of looking through the compile units in llvm.dbg.cu,
we can walk up the scope for the first instruction in the function with
a debug loc. If that describes the function, we use it. Otherwise, we
keep looking until we find one.
If no such instruction is found, we then give up and produce an error.
Reviewers: echristo, dblaikie
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D5887
llvm-svn: 220382
ConstantFolding crashes when trying to InstSimplify the following load:
@a = private unnamed_addr constant %mst {
i8* inttoptr (i64 -1 to i8*),
i8* inttoptr (i64 -1 to i8*)
}, align 8
%x = load <2 x i8*>* bitcast (%mst* @a to <2 x i8*>*), align 8
This patch fix this by adding support to this type of folding:
%x = load <2 x i8*>* bitcast (%mst* @a to <2 x i8*>*), align 8
==> gets folded to:
%x = <2 x i8*> <i8* inttoptr (i64 -1 to i8*), i8* inttoptr (i64 -1 to i8*)>
llvm-svn: 220380
On AArch64, GOT references are page relative (ADRP + LDR), so they can't be
applied until we know exactly where, within a page, the GOT entry will be in
the target address space.
Fixes <rdar://problem/18693976>.
llvm-svn: 220347
These are named following the IEEE-754 names for these
functions, rather than the libm fmin / fmax to avoid
possible ambiguities. Some languages may implement something
resembling fmin / fmax which return NaN if either operand is
to propagate errors. These implement the IEEE-754 semantics
of returning the other operand if either is a NaN representing
missing data.
llvm-svn: 220341
This enables targets to adapt their pass pipeline to the register
allocator in use. For example, with the AArch64 backend, using PBQP
with the cortex-a57, the FPLoadBalancing pass is no longer necessary.
llvm-svn: 220321
This function was complicated by the fact that it tried to perform
canonicalizations that were already preformed by InstSimplify. Remove
this extra code and move the tests over to InstSimplify. Add asserts to
make sure our preconditions hold before we make any assumptions.
llvm-svn: 220314
With VSX enabled, test/CodeGen/PowerPC/recipest.ll exposes a bug in
the FMA mutation pass. If we have a situation where a killed product
register is the same register as the FMA target, such as:
%vreg5<def,tied1> = XSNMSUBADP %vreg5<tied0>, %vreg11, %vreg5,
%RM<imp-use>; VSFRC:%vreg5 F8RC:%vreg11
then the substitution makes no sense. We end up getting a crash when
we try to extend the interval associated with the killed product
register, as there is already a live range for %vreg5 there. This
patch just disables the mutation under those circumstances.
Since recipest.ll generates different code with VMX enabled, I've
modified that test to use -mattr=-vsx. I've borrowed the code from
that test that exposed the bug and placed it in fma-mutate.ll, where
it tests several mutation opportunities including the "bad" one.
llvm-svn: 220290
The 32-bit variants of the NEON scalar<->GPR move instructions are
also available in VFPv2. The 8- and 16-bit variants do require NEON.
Note that the checks in the test file are all -DAG because they are
checking a mixture of stdout and stderr, and the ordering is not
guaranteed.
llvm-svn: 220288
Summary: Fixed memory accesses with rbp as a base or an index register.
Reviewers: eugenis
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D5819
llvm-svn: 220283
inttoptr or ptrtoint cast provided there is datalayout available.
Eventually, the datalayout can just be required but in practice it will
always be there today.
To go with the ability to expose available values requiring a ptrtoint
or inttoptr cast, helpers are added to perform one of these three casts.
These smarts are necessary to finish canonicalizing loads and stores to
the operational type requirements without regressing fundamental
combines.
I've added some test cases. These should actually improve as the load
combining and store combining improves, but they may fundamentally be
highlighting some missing combines for select in addition to exercising
the specific added logic to load analysis.
llvm-svn: 220277
Every target we support has support for assembly that looks like
a = b - c
.long a
What is special about MachO is that the above combination suppresses the
production of a relocation.
With this change we avoid producing the intermediary labels when they don't
add any value.
llvm-svn: 220256
When functions are inlined, instructions without debug information are
attributed to the call site's DebugLoc. After inlining, inlined static
allocas are moved to the caller's entry block, adjacent to the caller's
original static alloca instructions. By retaining the call site's
DebugLoc, these instructions could cause instructions that were
subsequently inserted at the entry block to pick up the same DebugLoc.
Patch by Wolfgang Pieb!
llvm-svn: 220255
the CGO build environment. This lets things like -rpath propagate down
to the C++ code that is built along side the Go bindings when testing
them.
Patch by Peter Collingbourne, and verified that it works by me.
llvm-svn: 220252
Range metadata applies to loads, call, and invokes. We were validating that metadata applied to loads was correct according to the LangRef, but we were not validating metadata applied to calls or invokes. This change extracts the checking functionality to a common location, reuses it for all valid locations, and adds a simple test to ensure a misused range on a call gets reported.
llvm-svn: 220246
X86 code to lower VSELECT messed a bit with the bits set in the mask of VSELECT
when it knows it can be lowered into BLEND. Indeed, only the high bits need to be
set for those and it optimizes those accordingly.
However, when the mask is a compile time constant, the lowering will be handled
by the generic optimizer and those modifications will generate bad code in the
generic optimizer.
This patch fixes that by preventing the optimization if the VSELECT will be
handled by the generic optimizer.
<rdar://problem/18675020>
llvm-svn: 220242
The newly introduced 'nonnull' metadata is analogous to existing 'nonnull' attributes, but applies to load instructions rather than call arguments or returns. Long term, it would be nice to combine these into a single construct. The value of the load is allowed to vary between successive loads, but null is not a valid value to be loaded by any load marked nonnull.
Reviewed by: Hal Finkel
Differential Revision: http://reviews.llvm.org/D5220
llvm-svn: 220240
This patch improves support for commutative instructions in the x86 memory folding implementation by attempting to fold a commuted version of the instruction if the original folding fails - if that folding fails as well the instruction is 're-commuted' back to its original order before returning.
Updated version of r219584 (reverted in r219595) - the commutation attempt now explicitly ensures that neither of the commuted source operands are tied to the destination operand / register, which was the source of all the regressions that occurred with the original patch attempt.
Added additional regression test case provided by Joerg Sonnenberger.
Differential Revision: http://reviews.llvm.org/D5818
llvm-svn: 220239
The previous code had a few problems, motivating the choices here.
1. It could create instructions clobbering CPSR, but the incoming MachineInstr
didn't reflect this. A potential source of corruption. This is why the patch
has a new PseudoInst for before lowering.
2. Similarly, there was some code to handle the incoming instruction not being
ARMCC::AL, but this would have caused massive problems if it was actually
invoked when a complex offset needing more than one instruction was requested.
3. It wasn't designed to handle unaligned pointers (or offsets). These should
probably be minimised anyway, but the code needs to deal with them properly
regardless.
4. It had some rather dubious ad-hoc code to avoid calling
emitThumbRegPlusImmediate, a function which should be designed to do precisely
this job.
We seem to cover the common cases correctly now, and hopefully can enhance
emitThumbRegPlusImmediate to handle any extra optimisations we need to add in
future.
llvm-svn: 220236
The current instruction selection patterns for SMULW[BT] and SMLAW[BT]
are incorrect. These instructions multiply a 32-bit and a 16-bit value
(both signed) and return the top 32 bits of the 48-bit result. This
preserves the 16 bits of overflow, whereas the patterns they currently
match truncate the result to 16 bits then sign extend.
To select these instructions, we would need to match an ISD::SMUL_LOHI,
a sign extend, two shifts and an or. There is no way to match SMUL_LOHI
in an instruction pattern as it defines multiple values, so this would
have to be done in C++. I have raised
http://llvm.org/bugs/show_bug.cgi?id=21297 to cover allowing correct
selection of these instructions.
This fixes http://llvm.org/bugs/show_bug.cgi?id=19396
llvm-svn: 220196
This function can, for some offsets from the SP, split one instruction
into two. Since it re-uses the original instruction as the first
instruction of the result, we need ensure its result register is not
marked as dead before we use it in the second instruction.
llvm-svn: 220194
The original code had an implicit assumption that if the test for
allocas or globals was reached, the two pointers were not equal. With my
changes to make the pointer analysis more powerful here, I also had to
guard against circumstances where the results weren't useful. That in
turn violated the assumption and gave rise to a circumstance in which we
could have a store with both the queried pointer and stored pointer
rooted at *the same* alloca. Clearly, we cannot ignore such a store.
There are other things we might do in this code to better handle the
case of both pointers ending up at the same alloca or global, but it
seems best to at least make the test explicit in what it intends to
check.
I've added tests for both the alloca and global case here.
llvm-svn: 220190
r220178. First, the creation routine doesn't insert prior to the
terminator of the basic block provided, but really at the end of the
basic block. Instead, get the terminator and insert before that. The
next issue was that we need to ensure multiple PHI node entries for
a single predecessor re-use the same cast instruction rather than
creating new ones.
All of the logic here was without tests previously. I've reduced and
added a test case from the test suite that crashed without both of these
fixes.
llvm-svn: 220186
logic to look through pointer casts, making them trivially stronger in
the face of loads and stores with intervening pointer casts.
I've included a few test cases that demonstrate the kind of folding
instcombine can do without pointer casts and then variations which
obfuscate the logic through bitcasts. Without this patch, the variations
all fail to optimize fully.
This is more important now than it has been in the past as I've started
moving the load canonicialization to more closely follow the value type
requirements rather than the pointer type requirements and thus this
needs to be prepared for more pointer casts. When I made the same change
to stores several test cases regressed without logic along these lines
so I wanted to systematically improve matters first.
llvm-svn: 220178
of InstCombine rather than just the bits enabled when datalayout is
optional.
The primary fixes here are because now things are little endian.
In good news, silliness like this seems like it will be going away as
we've got pretty stong consensus on dropping optional datalayout
entirely.
llvm-svn: 220176
We recently discovered an issue that reinforces what a good idea it is
to always specify -mcpu in our code generation tests, particularly for
-mattr=+vsx. This patch ensures that all tests that specify
-mattr=+vsx also specify -mcpu=pwr7 or -mcpu=pwr8, as appropriate.
Some of the uses of -mattr=+vsx added recently don't make much sense
(when specified for -mtriple=powerpc-apple-darwin8 or -march=ppc32,
for example). For cases like this I've just removed the extra VSX
test commands; there's enough coverage without them.
llvm-svn: 220173
Patch by Bill Seurer; some comment formatting changes by me.
There are a few PowerPC test cases for FastISel support that currently
fail with VSX support enabled. The temporary workaround under
discussion in http://reviews.llvm.org/D5362 helps, but the tests still
fail because they specify -fast-isel-abort, and the VSX workaround
punts back to SelectionDAG. We have plans to fix FastISel permanently
for VSX, but until that's in place these tests are preventing us from
enabling VSX by default. Therefore we are adding -mattr=-vsx to these
tests until the full support is ready.
llvm-svn: 220172
The VSX testing variant in test/CodeGen/PowerPC/fma.ll had to be
disabled because of unexpected behavior on many of the builders. I
tracked this down to a situation that occurs when the VSX attribute is
enabled for a target that disables the MI early scheduling pass. This
patch adds -mcpu=pwr7 to make this predictable. The other issue will
be addressed separately.
llvm-svn: 220171
loads.
This handles many more cases than just the AA metadata, some of them
suggested by Hal in his review of the AA metadata handling patch. I've
tried to test this behavior where tractable to do so.
I'll point out that I have specifically *not* included a test for
debuginfo because it was going to require 2 or 3 times as much work to
craft some input which would survive the "helpful" stripping of debug
info metadata that doesn't match the desired schema. This is another
good example of why the current state of write-ability for our debug
info metadata is unacceptable. I spent over 30 minutes trying to conjure
some test case that would survive, even copying from other debug info
tests, but it always failed to survive with no explanation of why or how
I might fix it. =[
llvm-svn: 220165
up to where it actually works as intended. The problem is that
a GlobalAlias isa GlobalValue and so the prior block handled all of the
cases.
This allows us to constant fold based on the actual constant expression
in the global alias. As an example, see the last function in the newly
added test case which explicitly aligns an unaligned pointer using
constant expression math. Without this change, we fail to see that and
fold an alignment test to zero.
llvm-svn: 220164
The following implements the transformation:
(sub (or A B) (xor A B)) --> (and A B).
Patch by Ankur Garg!
Differential Revision: http://reviews.llvm.org/D5719
llvm-svn: 220163
The following implements the optimization for sequences of the form:
icmp eq/ne (shl Const2, A), Const1
Such sequences can be transformed to:
icmp eq/ne A, (TrailingZeros(Const1) - TrailingZeros(Const2))
This handles only the equality operators for now. Other operators need
to be handled.
Patch by Ankur Garg!
llvm-svn: 220162
by my refactoring of this code.
The method isSafeToLoadUnconditionally assumes that the load will
proceed with the preferred type alignment. Given that, it has to ensure
that the alloca or global is at least that aligned. It has always done
this historically when a datalayout is present, but has never checked it
when the datalayout is absent. When I refactored the code in r220156,
I exposed this path when datalayout was present and that turned the
latent bug into a patent bug.
This fixes the issue by just removing the special case which allows
folding things without datalayout. This isn't worth the complexity of
trying to tease apart when it is or isn't safe without actually knowing
the preferred alignment.
llvm-svn: 220161
...)) and (load (cast ...)): canonicalize toward the former.
Historically, we've tried to load using the type of the *pointer*, and
tried to match that type as closely as possible removing as many pointer
casts as we could and trading them for bitcasts of the loaded value.
This is deeply and fundamentally wrong.
Repeat after me: memory does not have a type! This was a hard lesson for
me to learn working on SROA.
There is only one thing that should actually drive the type used for
a pointer, and that is the type which we need to use to load from that
pointer. Matching up pointer types to the loaded value types is very
useful because it minimizes the physical size of the IR required for
no-op casts. Similarly, the only thing that should drive the type used
for a loaded value is *how that value is used*! Again, this minimizes
casts. And in fact, the *only* thing motivating types in any part of
LLVM's IR are the types used by the operations in the IR. We should
match them as closely as possible.
I've ended up removing some tests here as they were testing bugs or
behavior that is no longer present. Mostly though, this is just cleanup
to let the tests continue to function as intended.
The only fallout I've found so far from this change was SROA and I have
fixed it to not be impeded by the different type of load. If you find
more places where this change causes optimizations not to fire, those
too are likely bugs where we are assuming that the type of pointers is
"significant" for optimization purposes.
llvm-svn: 220138
This test is pretty awesome. It is claiming to test devirtualization.
However, the code in question is not in fact devirtualized by LLVM. If
you take the original C++ test case and run it through Clang at -O3 we
fail to devirtualize it completely. It also isn't a sufficiently focused
test case.
The *reason* we fail to devirtualize it isn't because of any missing
instcombine though. Instead, it is because we fail to emit an available
externally vtable and thus the vtable is just an external and completely
opaque. If I cause the vtable to be emitted, we successfully
devirtualize things.
Anyways, I'm just removing it because it is providing negative value at
this point: it isn't representative of the output of Clang really, LLVM
isn't doing the transform it claims to be testing, LLVM's failure to do
the transform isn't actually an LLVM bug at all and we shouldn't be
testing for it here, and finally the test is written in such a way that
it will trivially pass even when the point of the test is failing.
llvm-svn: 220137
cases where the alloca type, the load types, and the store types used
all disagree.
Previously, the only way that vector-based promotion occured was if the
alloca type was a vector type. This was one of the *very* few remaining
uses of the alloca's type to guide SROA/mem2reg left in LLVM. It turns
out it was a bad idea.
The alloca type can change very easily based on the mixture of types
loaded and stored to that alloca. We shouldn't be relying on it as
a signal for very much. Instead, the source of truth should be loads and
stores. We should canonicalize the loads and stores as much as possible
and then rely on them exclusively in SROA.
When looking and loads and stores, we may find many different candidate
vector types. This change will let SROA try all of them to find a vector
type which is a viable way to promote the entire alloca to a vector
register.
With this change, it becomes possible to do better canonicalization and
optimization of loads and stores without breaking SROA in random ways,
and that should allow fixing a core source of performance loss in hot
numerical loops such as those in Eigen.
llvm-svn: 220116
The previous tests claimed to test constant offsets in the function name,
but the tests weren't actually testing them.
Clone the tests, and do testing of all combinations of the following:
1) with/without constant pointer offset
2) 32/64-bit addressing modes
3) Usage and non-usage of the return value from the atomicrmw
Reviewed-by: Matt Arsenault <matthew.arsenault@amd.com>
llvm-svn: 220103
The function name now matches what it's actually testing.
Signed-off-by: Aaron Watry <awatry@gmail.com>
Reviewed-by: Matt Arsenault <matthew.arsenault@amd.com>
llvm-svn: 220102
TL;DR: Indexing maps with [] creates missing entries.
The long version:
When selecting lifetime intrinsics, we index the *static* alloca map with the AllocaInst we find for that lifetime. Trouble is, we don't first check to see if this is a dynamic alloca.
On the attached example, this causes a dynamic alloca to create an entry in the static map, and returns 0 (the default) as the frame index for that lifetime. 0 was used for the frame index of the stack protector, which given that it now has a lifetime, is coloured, and merged with other stack slots.
PEI would later trigger an assert because it expects the stack protector to not be dead.
This fix ensures that we only get frame indices for static allocas, ie, those in the map. Dynamic ones are effectively dropped, which is suboptimal, but at least isn't completely broken.
rdar://problem/18672951
llvm-svn: 220099
This reverts commit r219899.
This also updates byval-tail-call.ll to make it clear what was breaking.
Adding r219899 again will cause the load/store to disappear.
llvm-svn: 220093
With VSX enabled, LLVM crashes when compiling
test/CodeGen/PowerPC/fma.ll. I traced this to the liveness test
that's revised in this patch. The interval test is designed to only
work for virtual registers, but in this case the AddendSrcReg is
physical. Since there is already a walk of the MIs between the
AddendMI and the FMA, I added a check for def/kill of the AddendSrcReg
in that loop. At Hal Finkel's request, I converted the liveness test
to an assert restricted to virtual registers.
I've changed the fma.ll test to have VSX and non-VSX variants so we
can test both kinds of multiply-adds.
llvm-svn: 220090
When the input to a store instruction was a zero vector, the backend
always selected a normal vector store regardless of the non-temporal
hint. This is fixed by this patch.
This fixes PR19370.
llvm-svn: 220054
We should be talking about the number of source elements, not the number of destination elements, given we know at this point that the source and dest element numbers are not the same.
While we're at it, avoid writing to std::vector::end()...
Bug found with random testing and a lot of coffee.
llvm-svn: 220051
Currently the VSX support enables use of lxvd2x and stxvd2x for 2x64
types, but does not yet use lxvw4x and stxvw4x for 4x32 types. This
patch adds that support.
As with lxvd2x/stxvd2x, this involves straightforward overriding of
the patterns normally recognized for lvx/stvx, with preference given
to the VSX patterns when VSX is enabled.
In addition, the logic for permitting misaligned memory accesses is
modified so that v4r32 and v4i32 are treated the same as v2f64 and
v2i64 when VSX is enabled. Finally, the DAG generation for unaligned
loads is changed to just use a normal LOAD (which will become lxvw4x)
on P8 and later hardware, where unaligned loads are preferred over
lvsl/lvx/lvx/vperm.
A number of tests now generate the VSX loads/stores instead of
lvx/stvx, so this patch adds VSX variants to those tests. I've also
added <4 x float> tests to the vsx.ll test case, and created a
vsx-p8.ll test case to be used for testing code generation for the
P8Vector feature. For now, that simply tests the unaligned load/store
behavior.
This has been tested along with a temporary patch to enable the VSX
and P8Vector features, with no new regressions encountered with or
without the temporary patch applied.
llvm-svn: 220047
v2: use dyn_cast
fixup comments
v3: use cast
Reviewed-by: Matt Arsenault <arsenm2@gmail.com>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 220044
DSE's overlap checking contained special logic, used only when no DataLayout
was available, which inferred a complete overwrite when the pointee types were
equal. This logic seems fine for regular loads/stores, but does not work for
memcpy and friends. Instead of fixing this, I'm just removing it.
Philosophically, transformations should not contain enhanced behavior used only
when data layout is lacking (data layout should be strictly additive), and
maintaining these rarely-tested code paths seems not worthwhile at this stage.
Credit to Aliaksei Zasenka for the bug report and the diagnosis. The test case
(slightly reduced from that provided by Aliaksei) replaces the original
contents of test/Transforms/DeadStoreElimination/no-targetdata.ll -- a few
other tests have been updated to have a data layout.
llvm-svn: 220035
The only difference from r219829 is using
getOrCreateSectionSymbol(*ELFSec)
instead of
GetOrCreateSymbol(ELFSec->getSectionName())
in ELFObjectWriter which causes us to use the correct section symbol even if
we have multiple sections with the same name.
Original messages:
r219829:
Correctly handle references to section symbols.
When processing assembly like
.long .text
we were creating a new undefined symbol .text. GAS on the other hand would
handle that as a reference to the .text section.
This patch implements that by creating the section symbols earlier so that
they are visible during asm parsing.
The patch also updates llvm-readobj to print the symbol number in the relocation
dump so that the test can differentiate between two sections with the same name.
r219835:
Allow forward references to section symbols.
llvm-svn: 220021
Patch by Bill Seurer; committed on his behalf.
These test cases generate slightly different code sequences when VSX
is activated and thus fail. The update turns off VSX explicitly for
the existing checks and then adds a second set of checks for most of
them that test the VSX instruction output.
llvm-svn: 220019
The bug is in ARMConstantIslands::createNewWater where the upper bound of the
new water split point is computed:
// This could point off the end of the block if we've already got constant
// pool entries following this block; only the last one is in the water list.
// Back past any possible branches (allow for a conditional and a maximally
// long unconditional).
if (BaseInsertOffset + 8 >= UserBBI.postOffset()) {
BaseInsertOffset = UserBBI.postOffset() - UPad - 8;
DEBUG(dbgs() << format("Move inside block: %#x\n", BaseInsertOffset));
}
The split point is supposed to be somewhere between the machine instruction that
loads from the constant pool entry and the end of the basic block, before branch
instructions. The code above is fine if the basic block is large enough and
there are a sufficient number of instructions following the machine instruction.
However, if the machine instruction is near the end of the basic block,
BaseInsertOffset can point to the machine instruction or another instruction
that precedes it, and this can lead to convergence failure.
This commit fixes this bug by ensuring BaseInsertOffset is larger than the
offset of the instruction following the constant-loading instruction.
rdar://problem/18581150
llvm-svn: 220015
Revert "Correctly handle references to section symbols."
Revert "Allow forward references to section symbols."
Rui found a regression I am debugging.
llvm-svn: 220010
llvm-symbolizer will consult one of the .dSYM paths passed via -dsym-hint
if it fails to find the .dSYM bundle at the default location.
llvm-svn: 220004
This code is based on the existing LLVM Go bindings project hosted at:
https://github.com/go-llvm/llvm
Note that all contributors to the gollvm project have agreed to relicense
their changes under the LLVM license and submit them to the LLVM project.
Differential Revision: http://reviews.llvm.org/D5684
llvm-svn: 219976
Summary:
Backends can use setInsertFencesForAtomic to signal to the middle-end that
montonic is the only memory ordering they can accept for
stores/loads/rmws/cmpxchg. The code lowering those accesses with a stronger
ordering to fences + monotonic accesses is currently living in
SelectionDAGBuilder.cpp. In this patch I propose moving this logic out of it
for several reasons:
- There is lots of redundancy to avoid: extremely similar logic already
exists in AtomicExpand.
- The current code in SelectionDAGBuilder does not use any target-hooks, it
does the same transformation for every backend that requires it
- As a result it is plain *unsound*, as it was apparently designed for ARM.
It happens to mostly work for the other targets because they are extremely
conservative, but Power for example had to switch to AtomicExpand to be
able to use lwsync safely (see r218331).
- Because it produces IR-level fences, it cannot be made sound ! This is noted
in the C++11 standard (section 29.3, page 1140):
```
Fences cannot, in general, be used to restore sequential consistency for atomic
operations with weaker ordering semantics.
```
It can also be seen by the following example (called IRIW in the litterature):
```
atomic<int> x = y = 0;
int r1, r2, r3, r4;
Thread 0:
x.store(1);
Thread 1:
y.store(1);
Thread 2:
r1 = x.load();
r2 = y.load();
Thread 3:
r3 = y.load();
r4 = x.load();
```
r1 = r3 = 1 and r2 = r4 = 0 is impossible as long as the accesses are all seq_cst.
But if they are lowered to monotonic accesses, no amount of fences can prevent it..
This patch does three things (I could cut it into parts, but then some of them
would not be tested/testable, please tell me if you would prefer that):
- it provides a default implementation for emitLeadingFence/emitTrailingFence in
terms of IR-level fences, that mimic the original logic of SelectionDAGBuilder.
As we saw above, this is unsound, but the best that can be done without knowing
the targets well (and there is a comment warning about this risk).
- it then switches Mips/Sparc/XCore to use AtomicExpand, relying on this default
implementation (that exactly replicates the logic of SelectionDAGBuilder, so no
functional change)
- it finally erase this logic from SelectionDAGBuilder as it is dead-code.
Ideally, each target would define its own override for emitLeading/TrailingFence
using target-specific fences, but I do not know the Sparc/Mips/XCore memory model
well enough to do this, and they appear to be dealing fine with the ARM-inspired
default expansion for now (probably because they are overly conservative, as
Power was). If anyone wants to compile fences more agressively on these
platforms, the long comment should make it clear why he should first override
emitLeading/TrailingFence.
Test Plan: make check-all, no functional change
Reviewers: jfb, t.p.northover
Subscribers: aemerson, llvm-commits
Differential Revision: http://reviews.llvm.org/D5474
llvm-svn: 219957
Summary:
Currently, call slot optimization requires that if the destination is an
argument, the argument has the sret attribute. This is to ensure that
the memory access won't trap. In addition to sret, we can also allow the
optimization to happen for arguments that have the new dereferenceable
attribute, which gives the same guarantee.
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D5832
llvm-svn: 219950
If a square root call has an FP multiplication argument that can be reassociated,
then we can hoist a repeated factor out of the square root call and into a fabs().
In the simplest case, this:
y = sqrt(x * x);
becomes this:
y = fabs(x);
This patch relies on an earlier optimization in instcombine or reassociate to put the
multiplication tree into a canonical form, so we don't have to search over
every permutation of the multiplication tree.
Because there are no IR-level FastMathFlags for intrinsics (PR21290), we have to
use function-level attributes to do this optimization. This needs to be fixed
for both the intrinsics and in the backend.
Differential Revision: http://reviews.llvm.org/D5787
llvm-svn: 219944
When the constant divisor was larger than 32bits, then the optimized code
generated for the AArch64 backend would emit the wrong code, because the shift
was defined as a shift of a 32bit constant '(1<<Lg2(divisor))' and we would
loose the upper 32bits.
This fixes rdar://problem/18678801.
llvm-svn: 219934
Summary:
In order to support big endian targets for the BuildPairF64 nodes we
just need to swap the low/high pair registers. Additionally, for the
ExtractElementF64 nodes we have to calculate the correct stack offset
with respect to the node's register/operand that we want to extract.
Reviewers: dsanders
Reviewed By: dsanders
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D5753
llvm-svn: 219931
Make tail recursion elimination a bit more aggressive. This allows us to get
tail recursion on functions that are just branches to a different function. The
fact that the function takes a byval argument does not restrict it from being
optimised into just a tail call.
llvm-svn: 219899
For pointer-typed function arguments, enhanced alignment can be asserted using
the 'align' attribute. When inlining, if this enhanced alignment information is
not otherwise available, preserve it using @llvm.assume-based alignment
assumptions.
llvm-svn: 219876
In AVX512f we support 64x2 and 32x8 inserts via matching them to 32x4 and 64x4
respectively. These are matched by "Alt" Pat<>'s (Alt stands for alternative
VTs).
Since DQ has native support for these intructions, I peeled off the non-"Alt"
part of the baseclass into vinsert_for_size_no_alt. The DQ instructions are
derived from this multiclass. The "Alt" Pat<>'s are disabled with DQ.
Fixes <rdar://problem/18426089>
llvm-svn: 219874
The SelectDS1Addr1Offset complex pattern always tries to store constant
lds pointers in the offset operand and store a zero value in the addr operand.
Since the addr operand does not accept immediates, the zero value
needs to first be copied to a register.
This newly created zero value will not go through normal instruction
selection, so we need to manually insert a V_MOV_B32_e32 in the complex
pattern.
This bug was hidden by the fact that if there was another zero value
in the DAG that had not been selected yet, then the CSE done by the DAG
would use the unselected node for the addr operand rather than the one
that was just created. This would lead to the zero value being selected
and the DAG automatically inserting a V_MOV_B32_e32 instruction.
llvm-svn: 219848
If x is known to have the range [a, b) in a loop predicated by (icmp
ne x, a), its range can be sharpened to [a + 1, b). Get
ScalarEvolution and hence IndVars to exploit this fact.
This change triggers an optimization to widen-loop-comp.ll, so it had
to be edited to get it to pass.
phabricator: http://reviews.llvm.org/D5639
llvm-svn: 219834
Truncate the operands of a switch instruction to a narrower type if the upper
bits are known to be all ones or zeros.
rdar://problem/17720004
llvm-svn: 219832
This is mostly a copy of the existing FastISel GEP code, but we have to
duplicate it for AArch64, because otherwise we would bail out even for simple
cases. This is because the standard fastEmit functions don't cover MUL at all
and ADD is lowered very inefficientily.
The original commit had a bug in the add emit logic, which has been fixed.
llvm-svn: 219831
When processing assembly like
.long .text
we were creating a new undefined symbol .text. GAS on the other hand would
handle that as a reference to the .text section.
This patch implements that by creating the section symbols earlier so that
they are visible during asm parsing.
The patch also updates llvm-readobj to print the symbol number in the relocation
dump so that the test can differentiate between two sections with the same name.
llvm-svn: 219829
The SLP vectorizer should not vectorize ephemeral values. These are used to
express information to the optimizer, and vectorizing them does not lead to
faster code (because the ephemeral values are dropped prior to code generation,
vectorized or not), and obscures the information the instructions are
attempting to communicate (the logic that interprets the arguments to
@llvm.assume generically does not understand vectorized conditions).
Also, uses by ephemeral values are free (because they, and the necessary
extractelement instructions, will be dropped prior to code generation).
llvm-svn: 219816
Summary:
Currently an error is thrown if bundle alignment mode is set more than once
per module (either via the API or the .bundle_align_mode directive). This
change allows setting it multiple times as long as the alignment doesn't
change.
Also nested bundle_lock groups are currently not allowed. This change allows
them, with the effect that the group stays open until all nests are exited,
and if any of the bundle_lock directives has the align_to_end flag, the
group becomes align_to_end.
These changes make the bundle aligment simpler to use in the compiler, and
also better match the corresponding support in GNU as.
Reviewers: jvoung, eliben
Differential Revision: http://reviews.llvm.org/D5801
llvm-svn: 219811
Summary:
Fixes a FIXME in MachineSinking. Instead of using the simple heuristics in
isPostDominatedBy, use the real MachinePostDominatorTree and MachineLoopInfo.
The old heuristics caused instructions to sink unnecessarily, and might create
register pressure.
This is the second try of the fix. The first one (D4814) caused a performance
regression due to failing to sink instructions out of loops (PR21115). This
patch fixes PR21115 by sinking an instruction from a deeper loop to a shallower
one regardless of whether the target block post-dominates the source.
Thanks Alexey Volkov for reporting PR21115!
Test Plan:
Added a NVPTX codegen test to verify that our change prevents the backend from
over-sinking. It also shows the unnecessary register pressure caused by
over-sinking.
Added an X86 test to verify we can sink instructions out of loops regardless of
the dominance relationship. This test is reduced from Alexey's test in PR21115.
Updated an affected test in X86.
Also ran SPEC CINT2006 and llvm-test-suite for compilation time and runtime
performance. Results are attached separately in the review thread.
Reviewers: Jiangning, resistor, hfinkel
Reviewed By: hfinkel
Subscribers: hfinkel, bruno, volkalexey, llvm-commits, meheff, eliben, jholewinski
Differential Revision: http://reviews.llvm.org/D5633
llvm-svn: 219773
Peephole optimization that generates a single conditional branch
for csinc-branch sequences like in the examples below. This is
possible when the csinc sets or clears a register based on a condition
code and the branch checks that register. Also the condition
code may not be modified between the csinc and the original branch.
Examples:
1. Convert csinc w9, wzr, wzr, <CC>;tbnz w9, #0, 0x44
to b.<invCC>
2. Convert csinc w9, wzr, wzr, <CC>; tbz w9, #0, 0x44
to b.<CC>
rdar://problem/18506500
llvm-svn: 219742
A few minor changes to prevent @llvm.assume from interfering with loop
vectorization. First, treat @llvm.assume like the lifetime intrinsics, which
are scalarized (but don't otherwise interfere with the legality checking).
Second, ignore the cost of ephemeral instructions in the loop (these will go
away anyway during CodeGen).
Alignment assumptions and other uses of @llvm.assume can often end up inside of
loops that should be vectorized (this is not uncommon for assumptions generated
by __attribute__((align_value(n))), for example).
llvm-svn: 219741
Quentin Colombet