When sinking an instruction it might be moved past the original last use of one
of its operands. This last use has the kill flag set and the verifier will
obviously complain about this.
Before Machine Sinking (AArch64):
%vreg3<def> = ASRVXr %vreg1, %vreg2<kill>
%XZR<def> = SUBSXrs %vreg4, %vreg1<kill>, 160, %NZCV<imp-def>
...
After Machine Sinking:
%XZR<def> = SUBSXrs %vreg4, %vreg1<kill>, 160, %NZCV<imp-def>
...
%vreg3<def> = ASRVXr %vreg1, %vreg2<kill>
This fix clears all the kill flags in all instruction that use the same operands
as the instruction that is being sunk.
This fixes rdar://problem/18180996.
llvm-svn: 216803
This patch checks for DAG patterns that are an add or a sub followed by a
compare on 16 and 8 bit inputs. Since AArch64 does not support those types
natively they are legalized into 32 bit values, which means that mask operations
are inserted into the DAG to emulate overflow behaviour. In many cases those
masks do not change the result of the processing and just introduce a dependent
operation, often in the middle of a hot loop.
This patch detects the relevent DAG patterns and then tests to see if the
transforms are equivalent with and without the mask, removing the mask if
possible. The exact mechanism of this patch was discusses in
http://lists.cs.uiuc.edu/pipermail/llvmdev/2014-July/074444.html
There is a reasonably good chance there are missed oppurtunities due to similiar
(but not identical) DAG patterns that could be funneled into this test, adding
them should be simple if we see test cases.
Tests included.
rdar://13754426
llvm-svn: 216776
When we select a trunc instruction we don't emit any code if the type is already
i32 or smaller. This is because the instruction that uses the truncated value
will deal with it.
This behavior can incorrectly transfer a kill flag, which was meant for the
result of the truncate, onto the source register.
%2 = trunc i32 %1 to i16
... = ... %2 -> ... = ... vreg1 <kill>
... = ... %1 ... = ... vreg1
This commit fixes this by emitting a COPY instruction, so that the result and
source register are distinct virtual registers.
This fixes rdar://problem/18178188.
llvm-svn: 216750
In an llvm-stress generated test, we were trying to create a v0iN type and
asserting when that failed. This case could probably be handled by the
function, but not without added complexity and the situation it arises in is
sufficiently odd that there's probably no benefit anyway.
Should fix PR20775.
llvm-svn: 216725
This fix checks first if the instruction to be folded (e.g. sign-/zero-extend,
or shift) is in the same machine basic block as the instruction we are folding
into.
Not doing so can result in incorrect code, because the value might not be
live-out of the basic block, where the value is defined.
This fixes rdar://problem/18169495.
llvm-svn: 216700
The AArch64 target lowering for [zs]ext of vectors is set up to handle
input simple types and expects the generic SDag path to do something reasonable
with anything that's not a simple type. The code, however, was only
checking that the result type was a simple type and assuming that
implied that the source type would also be a simple type. That's not a
valid assumption, as operations like "zext <1 x i1> %0 to <1 x i32>"
demonstrate. The fix is to simply explicitly validate the source type
as well as the result type.
PR20791
llvm-svn: 216689
The included test case would fail, because the MI PHI node would have two
operands from the same predecessor.
This problem occurs when a switch instruction couldn't be selected. This happens
always, because there is no default switch support for FastISel to begin with.
The problem was that FastISel would first add the operand to the PHI nodes and
then fall-back to SelectionDAG, which would then in turn add the same operands
to the PHI nodes again.
This fix removes these duplicate PHI node operands by reseting the
PHINodesToUpdate to its original state before FastISel tried to select the
instruction.
This fixes <rdar://problem/18155224>.
llvm-svn: 216640
Currently instructions are folded very aggressively for AArch64 into the memory
operation, which can lead to the use of killed operands:
%vreg1<def> = ADDXri %vreg0<kill>, 2
%vreg2<def> = LDRBBui %vreg0, 2
... = ... %vreg1 ...
This usually happens when the result is also used by another non-memory
instruction in the same basic block, or any instruction in another basic block.
This fix teaches hasTrivialKill to not only check the LLVM IR that the value has
a single use, but also to check if the register that represents that value has
already been used. This can happen when the instruction with the use was folded
into another instruction (in this particular case a load instruction).
This fixes rdar://problem/18142857.
llvm-svn: 216634
Currently instructions are folded very aggressively into the memory operation,
which can lead to the use of killed operands:
%vreg1<def> = ADDXri %vreg0<kill>, 2
%vreg2<def> = LDRBBui %vreg0, 2
... = ... %vreg1 ...
This usually happens when the result is also used by another non-memory
instruction in the same basic block, or any instruction in another basic block.
If the computed address is used by only memory operations in the same basic
block, then it is safe to fold them. This is because all memory operations will
fold the address computation and the original computation will never be emitted.
This fixes rdar://problem/18142857.
llvm-svn: 216629
When the address comes directly from a shift instruction then the address
computation cannot be folded into the memory instruction, because the zero
register is not available as a base register. Simplify addess needs to emit the
shift instruction and use the result as base register.
llvm-svn: 216621
Use the zero register directly when possible to avoid an unnecessary register
copy and a wasted register at -O0. This also uses integer stores to store a
positive floating-point zero. This saves us from materializing the positive zero
in a register and then storing it.
llvm-svn: 216617
This teaches the AArch64 backend to deal with the operations required
to deal with the operations on v4f16 and v8f16 which are exposed by
NEON intrinsics, plus the add, sub, mul and div operations.
llvm-svn: 216555
When a shift with extension or an add with shift and extension cannot be folded
into the memory operation, then the address calculation has to be materialized
separately. While doing so the code forgot to consider a possible sign-/zero-
extension. This fix folds now also the sign-/zero-extension into the add or
shift instruction which is used to materialize the address.
This fixes rdar://problem/18141718.
llvm-svn: 216511
This is mostly achieved by providing the correct register class manually,
because getRegClassFor always returns the GPR*AllRegClass for MVT::i32 and
MVT::i64.
Also cleanup the code to use the FastEmitInst_* method whenever possible. This
makes sure that the operands' register class is properly constrained. For all
the remaining cases this adds the missing constrainOperandRegClass calls for
each operand.
llvm-svn: 216225
The AdvSIMD pass may produce copies that are not coalescer-friendly. The
peephole optimizer knows how to fix that as demonstrated in the test case.
<rdar://problem/12702965>
llvm-svn: 216200
This fixes a bug I introduced in a previous commit (r216033). Sign-/Zero-
extension from i1 cannot be folded into the ADDS/SUBS instructions. Instead both
operands have to be sign-/zero-extended with separate instructions.
Related to <rdar://problem/17913111>.
llvm-svn: 216073
legalization stage. With those two optimizations, fewer signed/zero extension
instructions can be inserted, and then we can expose more opportunities to
Machine CSE pass in back-end.
llvm-svn: 216066
LLVM generates illegal `rbit r0, #352` instruction for rbit intrinsic.
According to ARM ARM, rbit only takes register as argument, not immediate.
The correct instruction should be rbit <Rd>, <Rm>.
The bug was originally introduced in r211057.
Differential Revision: http://reviews.llvm.org/D4980
llvm-svn: 216064
Use FMOVWSr/FMOVXDr instead of FMOVSr/FMOVDr, which have the proper register
class to be used with the zero register. This makes the MachineInstruction
verifier happy again.
This is related to <rdar://problem/18027157>.
llvm-svn: 216040
Factor out the ADDS/SUBS instruction emission code into helper functions and
make the helper functions more clever to support most of the different ADDS/SUBS
instructions the architecture support. This includes better immedediate support,
shift folding, and sign-/zero-extend folding.
This fixes <rdar://problem/17913111>.
llvm-svn: 216033
This adds the missing test that I promised for r215753 to test the
materialization of the floating-point value +0.0.
Related to <rdar://problem/18027157>.
llvm-svn: 216019
Note: This was originally reverted to track down a buildbot error. Reapply
without any modifications.
Original commit message:
FastISel didn't take much advantage of the different addressing modes available
to it on AArch64. This commit allows the ComputeAddress method to recognize more
addressing modes that allows shifts and sign-/zero-extensions to be folded into
the memory operation itself.
For Example:
lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3]
ldr x0, [x0, x1]
sxtw x1, w1
lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3]
ldr x0, [x0, x1]
llvm-svn: 216013
Note: This was originally reverted to track down a buildbot error. Reapply
without any modifications.
Original commit message:
This change materializes now the value "0" from the zero register.
The zero register can be folded by several instruction, so no
materialization is need at all.
Fixes <rdar://problem/17924413>.
llvm-svn: 216009
This fixes a few BuildMI callsites where the result register was added by
using addReg, which is per default a use and therefore an operand register.
Also use the zero register as result register when emitting a compare
instruction (SUBS with unused result register).
llvm-svn: 215997
Externally-defined functions with weak linkage should not be
tail-called on ARM or AArch64, as the AAELF spec requires normal calls
to undefined weak functions to be replaced with a NOP or jump to the
next instruction. The behaviour of branch instructions in this
situation (as used for tail calls) is implementation-defined, so we
cannot rely on the linker replacing the tail call with a return.
llvm-svn: 215890
This reverts:
r215595 "[FastISel][X86] Add large code model support for materializing floating-point constants."
r215594 "[FastISel][X86] Use XOR to materialize the "0" value."
r215593 "[FastISel][X86] Emit more efficient instructions for integer constant materialization."
r215591 "[FastISel][AArch64] Make use of the zero register when possible."
r215588 "[FastISel] Let the target decide first if it wants to materialize a constant."
r215582 "[FastISel][AArch64] Cleanup constant materialization code. NFCI."
llvm-svn: 215673
Certain functions such as objc_autoreleaseReturnValue have to be called as
tail-calls even at -O0. Since normal fast-isel doesn't emit calls as tail calls,
we have to fall back to SelectionDAG to select calls that are marked as tail.
<rdar://problem/17991614>
llvm-svn: 215600
FastISel didn't take much advantage of the different addressing modes available
to it on AArch64. This commit allows the ComputeAddress method to recognize more
addressing modes that allows shifts and sign-/zero-extensions to be folded into
the memory operation itself.
For Example:
lsl x1, x1, #3 --> ldr x0, [x0, x1, lsl #3]
ldr x0, [x0, x1]
sxtw x1, w1
lsl x1, x1, #3 --> ldr x0, [x0, x1, sxtw #3]
ldr x0, [x0, x1]
llvm-svn: 215597
This change materializes now the value "0" from the zero register.
The zero register can be folded by several instruction, so no
materialization is need at all.
Fixes <rdar://problem/17924413>.
llvm-svn: 215591
The combiner ignored DBG nodes when checking
the uses of a virtual register.
It combined a sequence like
%vreg1 = madd %vreg2, %vreg3,...
DBG_VALUE (%vreg1 ...)
%vreg4 = add %vreg1,...
to
%vreg4 = madd %vreg2, %vreg3
leaving behind a dangling DBG_VALUE with
a definition. This triggered an assertion
in the MachineTraceMetrics.cpp module.
llvm-svn: 215431
be propagated to all its users, and this propagation could increase the
probability of finding common subexpressions. If the COPY has only one user,
the COPY itself can be removed.
llvm-svn: 215344
For best-case performance on Cortex-A57, we should try to use a balanced mix of odd and even D-registers when performing a critical sequence of independent, non-quadword FP/ASIMD floating-point multiply or multiply-accumulate operations.
This pass attempts to detect situations where the register allocation may adversely affect this load balancing and to change the registers used so as to better utilize the CPU.
Ideally we'd just take each multiply or multiply-accumulate in turn and allocate it alternating even or odd registers. However, multiply-accumulates are most efficiently performed in the same functional unit as their accumulation operand. Therefore this pass tries to find maximal sequences ("Chains") of multiply-accumulates linked via their accumulation operand, and assign them all the same "color" (oddness/evenness).
This optimization affects S-register and D-register floating point multiplies and FMADD/FMAs, as well as vector (floating point only) muls and FMADD/FMA. Q register instructions (and 128-bit vector instructions) are not affected.
llvm-svn: 215199
This short-circuited our error reporting for incorrectly specified
target triples (you'd get AArch64 code instead).
Should fix PR20567.
llvm-svn: 215191
__stack_chk_guard.
Handle the case where the pointer operand of the load instruction that loads the
stack guard is not a global variable but instead a bitcast.
%StackGuard = load i8** bitcast (i64** @__stack_chk_guard to i8**)
call void @llvm.stackprotector(i8* %StackGuard, i8** %StackGuardSlot)
Original test case provided by Ana Pazos.
This fixes PR20558.
llvm-svn: 215167
Re-commit of r214832,r21469 with a work-around that
avoids the previous problem with gcc build compilers
The work-around is to use SmallVector instead of ArrayRef
of basic blocks in preservesResourceLen()/MachineCombiner.cpp
llvm-svn: 215151
Instruction prefetch is not implemented for AArch64, it is incorrectly
translated into data prefetch instruction.
Differential Revision: http://reviews.llvm.org/D4777
llvm-svn: 214860
This implements basic argument lowering for AArch64 in FastISel. It only
handles a small subset of the C calling convention. It supports simple
arguments that can be passed in GPR and FPR registers.
This should cover most of the trivial cases without falling back to
SelectionDAG.
This fixes <rdar://problem/17890986>.
llvm-svn: 214846
sequence on AArch64
Re-commit of r214669 without changes to test cases
LLVM::CodeGen/AArch64/arm64-neon-mul-div.ll and
LLVM:: CodeGen/AArch64/dp-3source.ll
This resolves the reported compfails of the original commit.
llvm-svn: 214832
This fix changes the parameters #r and #s that are passed to the UBFM/SBFM
instruction to get the zero/sign-extension for free.
The original problem was that the shift left would use the 32-bit shift even for
i8/i16 value types, which could leave the upper bits set with "garbage" values.
The arithmetic shift right on the other side would use the wrong MSB as sign-bit
to determine what bits to shift into the value.
This fixes <rdar://problem/17907720>.
llvm-svn: 214788
scalar integer instruction pass.
This is a patch I had lying around from a few months ago. The pass is
currently disabled by default, so nothing to interesting.
llvm-svn: 214779
sequence - AArch64 target support
This patch turns off madd/msub generation in the DAGCombiner and generates
them in the MachineCombiner instead. It replaces the original code sequence
with the combined sequence when it is beneficial to do so.
When there is no machine model support it always generates the madd/msub
instruction. This is true also when the objective is to optimize for code
size: when the combined sequence is shorter is always chosen and does not
get evaluated.
When there is a machine model the combined instruction sequence
is evaluated for critical path and resource length using machine
trace metrics and the original code sequence is replaced when it is
determined to be faster.
rdar://16319955
llvm-svn: 214669
Add branch weights to branch instructions, so that the following passes can
optimize based on it (i.e. basic block ordering).
Fixes <rdar://problem/17887137>.
llvm-svn: 214537
The tbz/tbnz checks the sign bit to convert
op w1, w1, w10
cmp w1, #0
b.lt .LBB0_0
to
op w1, w1, w10
tbnz w1, #31, .LBB0_0
Differential Revision: http://reviews.llvm.org/D4440
llvm-svn: 214518
ADDS and SUBS cannot encode negative immediates or immediates larger than 12bit.
This fix checks if the immediate version can be used under this constraints and
if we can convert ADDS to SUBS or vice versa to support negative immediates.
Also update the test cases to test the immediate versions.
llvm-svn: 214470
This commit updates the existing SelectionDAG tests for the stackmap and patchpoint
intrinsics and enables FastISel testing. It also splits up the tests into separate
files, due to different codegen between SelectionDAG and FastISel.
llvm-svn: 214382
Currently the large code model for MachO uses the GOT to make function calls.
Emit the required adrp and ldr instructions to load the address from the GOT.
Related to <rdar://problem/17733076>.
llvm-svn: 214381
This improves the code generation for the XALU intrinsics when the
condition is feeding a select instruction.
This also updates and enables the XALU unit tests for FastISel.
This fixes <rdar://problem/17831117>.
llvm-svn: 214350
Currently the shift-immediate versions are not supported by tblgen and
hopefully this can be later removed, once the required support has been
added to tblgen.
llvm-svn: 214345
'J' represents a negative number suitable for an add/sub alias
instruction, but while preparing it to become an int64_t we were
mangling the sign extension. So "i32 -1" became 0xffffffffLL, for
example.
Should fix one half of PR20456.
llvm-svn: 214052
address of the stack guard was being spilled to the stack.
Previously the address of the stack guard would get spilled to the stack if it
was impossible to keep it in a register. This patch introduces a new target
independent node and pseudo instruction which gets expanded post-RA to a
sequence of instructions that load the stack guard value. Register allocator
can now just remat the value when it can't keep it in a register.
<rdar://problem/12475629>
llvm-svn: 213967
This commit implements the frameaddress intrinsic for the AArch64 architecture
in FastISel.
There were two test cases that pretty much tested the same, so I combined them
to a single test case.
Fixes <rdar://problem/17811834>
llvm-svn: 213959
which have successfully round-tripped through the combine phase, and use
this to ensure all operands to DAG nodes are visited by the combiner,
even if they are only added during the combine phase.
This is critical to have the combiner reach nodes that are *introduced*
during combining. Previously these would sometimes be visited and
sometimes not be visited based on whether they happened to end up on the
worklist or not. Now we always run them through the combiner.
This fixes quite a few bad codegen test cases lurking in the suite while
also being more principled. Among these, the TLS codegeneration is
particularly exciting for programs that have this in the critical path
like TSan-instrumented binaries (although I think they engineer to use
a different TLS that is faster anyways).
I've tried to check for compile-time regressions here by running llc
over a merged (but not LTO-ed) clang bitcode file and observed at most
a 3% slowdown in llc. Given that this is essentially a worst case (none
of opt or clang are running at this phase) I think this is tolerable.
The actual LTO case should be even less costly, and the cost in normal
compilation should be negligible.
With this combining logic, it is possible to re-legalize as we combine
which is necessary to implement PSHUFB formation on x86 as
a post-legalize DAG combine (my ultimate goal).
Differential Revision: http://reviews.llvm.org/D4638
llvm-svn: 213898
This bug is introduced by r211144. The element of operand may be
smaller than the element of result, but previous commit can
only handle the contrary condition. This commit is to handle this
scenario and generate optimized codes like ZIP1.
llvm-svn: 213830
The transform to constant fold unary operations with an AND across a
vector comparison applies when the constant is not a splat of a scalar
as well.
llvm-svn: 213800
The folding of unary operations through a vector compare and mask operation
is only safe if the unary operation result is of the same size as its input.
For example, it's not safe for [su]itofp from v4i32 to v4f64.
llvm-svn: 213799
I used the wrong method to obtain the return type inside FinishCall. This fix
simply uses the return type from FastLowerCall, which we already determined to
be a valid type.
Reduced test case from Chad. Thanks.
llvm-svn: 213788
The target-independent DAGcombiner will generate:
asr w1, X, #31 w1 = splat sign bit.
add X, X, w1, lsr #28 X = X + 0 or pow2-1
asr w0, X, asr #4 w0 = X/pow2
However, the add + shifts is expensive, so generate:
add w0, X, 15 w0 = X + pow2-1
cmp X, wzr X - 0
csel X, w0, X, lt X = (X < 0) ? X + pow2-1 : X;
asr w0, X, asr 4 w0 = X/pow2
llvm-svn: 213758
There really is no arm64_be: it was a useful fiction to test big-endian support
while both backends existed in parallel, but now the only platform that uses
the name (iOS) doesn't have a big-endian variant, let alone one called
"arm64_be".
llvm-svn: 213748
This commit modifies the existing call lowering functions to be used as the
FastLowerCall and FastLowerIntrinsicCall target-hooks instead.
This enables patchpoint intrinsic lowering for AArch64.
This fixes <rdar://problem/17733076>
llvm-svn: 213704
This makes the first stage DAG for @llvm.convert.to.fp16 an fptrunc,
and correspondingly @llvm.convert.from.fp16 an fpext. The legalisation
path is now uniform, regardless of the input IR:
fptrunc -> FP_TO_FP16 (if f16 illegal) -> libcall
fpext -> FP16_TO_FP (if f16 illegal) -> libcall
Each target should be able to select the version that best matches its
operations and not be required to duplicate patterns for both fptrunc
and FP_TO_FP16 (for example).
As a result we can remove some redundant AArch64 patterns.
llvm-svn: 213507
Because i16 is illegal, there's no native DAG method to
represent a bitcast to or from an f16 type. This meant LLVM was
inserting a stack store/load pair which is really not ideal.
llvm-svn: 213378
Actual support for softening f16 operations is still limited, and can be added
when it's needed. But Soften is much closer to being a useful thing to try
than keeping it Legal when no registers can actually hold such values.
Longer term, we probably want something between Soften and Promote semantics
for most targets, it'll be more efficient to promote the 4 basic operations to
f32 than libcall them.
llvm-svn: 213372
Since the result of a SETCC for AArch64 is 0 or -1 in each lane, we can
move unary operations, in this case [su]int_to_fp through the mask
operation and constant fold the operation away. Generally speaking:
UNARYOP(AND(VECTOR_CMP(x,y), constant))
--> AND(VECTOR_CMP(x,y), constant2)
where constant2 is UNARYOP(constant).
This implements the transform where UNARYOP is [su]int_to_fp.
For example, consider the simple function:
define <4 x float> @foo(<4 x float> %val, <4 x float> %test) nounwind {
%cmp = fcmp oeq <4 x float> %val, %test
%ext = zext <4 x i1> %cmp to <4 x i32>
%result = sitofp <4 x i32> %ext to <4 x float>
ret <4 x float> %result
}
Before this change, the code is generated as:
fcmeq.4s v0, v0, v1
movi.4s v1, #0x1 // Integer splat value.
and.16b v0, v0, v1 // Mask lanes based on the comparison.
scvtf.4s v0, v0 // Convert each lane to f32.
ret
After, the code is improved to:
fcmeq.4s v0, v0, v1
fmov.4s v1, #1.00000000 // f32 splat value.
and.16b v0, v0, v1 // Mask lanes based on the comparison.
ret
The svvtf.4s has been constant folded away and the floating point 1.0f
vector lanes are materialized directly via fmov.4s.
Rather than do the folding manually in the target code, teach getNode()
in the generic SelectionDAG to handle folding constant operands of
vector [su]int_to_fp nodes. It is reasonable (as noted in a FIXME) to do
additional constant folding there as well, but I don't have test cases
for those operations, so leaving them for another time when it becomes
appropriate.
rdar://17693791
llvm-svn: 213341
This makes the two intrinsics @llvm.convert.from.f16 and
@llvm.convert.to.f16 accept types other than simple "float". This is
only strictly needed for the truncate operation, since otherwise
double rounding occurs and there's no way to represent the strict IEEE
conversion. However, for symmetry we allow larger types in the extend
too.
During legalization, we can expand an "fp16_to_double" operation into
two extends for convenience, but abort when the truncate isn't legal. A new
libcall is probably needed here.
Even after this commit, various target tweaks are needed to actually use the
extended intrinsics. I've put these into separate commits for clarity, so there
are no actual tests of f64 conversion here.
llvm-svn: 213248
Memory barrier __builtin_arm_[dmb, dsb, isb] intrinsics are required to
implement their corresponding ACLE and MSVC intrinsics.
This patch ports ARM dmb, dsb, isb intrinsic to AArch64.
Differential Revision: http://reviews.llvm.org/D4520
llvm-svn: 213247
This adds a llvm.aarch64.hint intrinsic to mirror the llvm.arm.hint in order to
support the various hint intrinsic functions in the ACLE.
Add an optional pattern field that permits the subclass to specify the pattern
that matches the selection. The intrinsic pattern is set as mayLoad, mayStore,
so overload the value for the definition of the hint instruction.
llvm-svn: 212883
Storing will generally be immediately preceded by rounding from an f32
or f64, so make sure to match those patterns directly to convert into the
FPR16 register class directly rather than going through the integer GPRs.
This also eliminates an extra step in the convert-from-f64 path
which was first converting to f32 and then to f16 from there.
rdar://17594379
llvm-svn: 212638
Juergen Ributzka