stack slots and giving them different PseudoSourceValue's did not fix the
problem of post-alloc scheduling miscompiling llvm itself.
- Apply Dan's conservative workaround by assuming any non fixed stack slots can
alias other memory locations. This means a load from spill slot #1 cannot
move above a store of spill slot #2.
- Enable post-alloc scheduling for x86 at optimization leverl Default and above.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84424 91177308-0d34-0410-b5e6-96231b3b80d8
so get rid of eh.selector.i64 and rename eh.selector.i32 to eh.selector.
Likewise for eh.typeid.for. This aligns us with gcc, which always uses a
32 bit value for the selector on all platforms. My understanding is that
the register allocator used to assert if the selector intrinsic size didn't
match the pointer size, and this was the reason for introducing the two
variants. However my testing shows that this is no longer the case (I
fixed some bugs in selector lowering yesterday, and some more today in the
fastisel path; these might have caused the original problems).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84106 91177308-0d34-0410-b5e6-96231b3b80d8
truncating an SDValue (depending on whether the target
type is bigger or smaller than the value's type); or zero
extending or truncating it. Use it in a few places (this
seems to be a popular operation, but I only modified cases
of it in SelectionDAGBuild). In particular, the eh_selector
lowering was doing this wrong due to a repeated rather than
inverted test, fixed with this change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84027 91177308-0d34-0410-b5e6-96231b3b80d8
into MachineInstrs. This is mostly just moving the code from
ScheduleDAGSDNodesEmit.cpp into a new class. This decouples MachineInstr
emitting from scheduling.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83699 91177308-0d34-0410-b5e6-96231b3b80d8
While recording beginning of a function, use scope info from the first location entry instead of just relying on first location entry itself.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83684 91177308-0d34-0410-b5e6-96231b3b80d8
basic blocks that are so long that their size overflows a short.
Also assert that overflow does not happen in the future, as requested by Evan.
This fixes PR4401.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83159 91177308-0d34-0410-b5e6-96231b3b80d8
- Allocate MachineMemOperands and MachineMemOperand lists in MachineFunctions.
This eliminates MachineInstr's std::list member and allows the data to be
created by isel and live for the remainder of codegen, avoiding a lot of
copying and unnecessary translation. This also shrinks MemSDNode.
- Delete MemOperandSDNode. Introduce MachineSDNode which has dedicated
fields for MachineMemOperands.
- Change MemSDNode to have a MachineMemOperand member instead of its own
fields with the same information. This introduces some redundancy, but
it's more consistent with what MachineInstr will eventually want.
- Ignore alignment when searching for redundant loads for CSE, but remember
the greatest alignment.
Target-specific code which previously used MemOperandSDNodes with generic
SDNodes now use MemIntrinsicSDNodes, with opcodes in a designated range
so that the SelectionDAG framework knows that MachineMemOperand information
is available.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82794 91177308-0d34-0410-b5e6-96231b3b80d8
naming scheme used in SelectionDAG, where there are multiple kinds
of "target" nodes, but "machine" nodes are nodes which represent
a MachineInstr.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82790 91177308-0d34-0410-b5e6-96231b3b80d8
before producing FSIN, FCOS, FSQRT. If they aren't
so marked we have to assume they might set errno.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82781 91177308-0d34-0410-b5e6-96231b3b80d8
allows appropriate backends to generate a sqrt instruction.
On x86, this isn't done at -O0 because we go through
FastISel instead. This is a behavior change from before
this series of sqrt patches started. I think this is OK
considering that compile speed is most important at -O0, but
could be convinced otherwise.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82778 91177308-0d34-0410-b5e6-96231b3b80d8
two different places for printing MachineMemOperands.
Drop the virtual from Value::dump and instead give Value a
protected virtual hook that can be overridden by subclasses
to implement custom printing. This lets printing be more
consistent, and simplifies printing of PseudoSourceValue
values.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82599 91177308-0d34-0410-b5e6-96231b3b80d8
the base pointer, without the offset. This matches MemSDNode's
new alignment behavior, and holds more interesting information.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82473 91177308-0d34-0410-b5e6-96231b3b80d8
This can be seen on CodeGen/Generic/2006-09-06-SwitchLowering.ll. But it's not known to cause any real regression (but I have added an assertion for it now).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82214 91177308-0d34-0410-b5e6-96231b3b80d8
1. Switch from an std::set to a SmallPtrSet for visited chain nodes.
2. Do not force the recursive flattening of token factor nodes, regardless of
use count.
3. Immediately process newly created TokenFactor nodes.
Also, improve combiner-aa by teaching it that loads to non-overlapping offsets
of relatively aligned objects cannot alias.
These changes result in a >5x speedup for combiner-aa on most testcases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@81816 91177308-0d34-0410-b5e6-96231b3b80d8
DAG Combiner to disambiguate chains for loads and stores of types which are
broken up by the Legalizer into smaller pieces.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@81813 91177308-0d34-0410-b5e6-96231b3b80d8
to instructions instead of zero extended ones. This makes the asmprinter
print signed values more consistently. This apparently only really affects
the X86 backend.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@81265 91177308-0d34-0410-b5e6-96231b3b80d8
from floating-point to integer first, and bitcast the result
back to floating-point. Previously, this test was passing by
falling back to SelectionDAG lowering. The resulting code isn't
as nice, but it's correct and CodeGen now stays on the fast path.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@81171 91177308-0d34-0410-b5e6-96231b3b80d8
Do not use DenseMap operator[] because it inserts new entry if lookup fails. Use find() to check an entry in a DenseMap first.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@81058 91177308-0d34-0410-b5e6-96231b3b80d8
This change speeds up llvm-gcc by more then 6% at "-O0 -g" (measured by compiling InstructionCombining.cpp!)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@79977 91177308-0d34-0410-b5e6-96231b3b80d8
be of (dynamically) constant values, so races on it are immaterial. We just need
to ensure that at least one write has completed before return the pointer into it.
With this change, parllc exhibits essentially no overhead on 403.gcc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@79708 91177308-0d34-0410-b5e6-96231b3b80d8
If two uses of a CopyFromReg want different regclasses, first try a common
sub-class, then fall back on the copy emitted in AddRegisterOperand. There is
no need for an assert here. The cross-class joiner usually cleans up nicely.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@79193 91177308-0d34-0410-b5e6-96231b3b80d8
libcall. Take advantage of this in the ARM backend to rectify broken
choice of CC when hard float is in effect. PIC16 may want to see if
it could be of use in MakePIC16Libcall, which works unchanged.
Patch by Sandeep!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@79033 91177308-0d34-0410-b5e6-96231b3b80d8
and short. Well, it's kinda short. Definitely nasty and brutish.
The front-end generates the register/unregister calls into the SjLj runtime,
call-site indices and landing pad dispatch. The back end fills in the LSDA
with the call-site information provided by the front end. Catch blocks are
not yet implemented.
Built on Darwin and verified no llvm-core "make check" regressions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78625 91177308-0d34-0410-b5e6-96231b3b80d8
MERGE_VALUES nodes. Replacing the result values with the
operands in one MERGE_VALUES node may cause another
MERGE_VALUES node be CSE'd with the first one, and bring
its uses along, so that the first one isn't dead, as this
code expects. Fix this by iterating until the node is
really dead. This fixes PR4699.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78619 91177308-0d34-0410-b5e6-96231b3b80d8
Blackfin supports and/or/xor on i32 but not on i16. Teach
DAGCombiner::SimplifyBinOpWithSameOpcodeHands to not produce illegal nodes
after legalize ops.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78497 91177308-0d34-0410-b5e6-96231b3b80d8
and high-bits values in ways that weren't correct for integer
types wider than 64 bits. This fixes a miscompile in
PPMacroExpansion.cpp in clang on x86-64.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78295 91177308-0d34-0410-b5e6-96231b3b80d8
Instead of awkwardly encoding calling-convention information with ISD::CALL,
ISD::FORMAL_ARGUMENTS, ISD::RET, and ISD::ARG_FLAGS nodes, TargetLowering
provides three virtual functions for targets to override:
LowerFormalArguments, LowerCall, and LowerRet, which replace the custom
lowering done on the special nodes. They provide the same information, but
in a more immediately usable format.
This also reworks much of the target-independent tail call logic. The
decision of whether or not to perform a tail call is now cleanly split
between target-independent portions, and the target dependent portion
in IsEligibleForTailCallOptimization.
This also synchronizes all in-tree targets, to help enable future
refactoring and feature work.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78142 91177308-0d34-0410-b5e6-96231b3b80d8
TLI.computeMaskedBitsForTargetNode from ComputeMaskedBits, since
the former may call back into the latter. This fixes a major
compile time problem on a testcase that happnened to hit this
in a particularly bad way, PR4643.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@78023 91177308-0d34-0410-b5e6-96231b3b80d8
support. This isn't immediately interesting, because Legalize
ends up lowering SELECT_CC if the target doesn't support it,
but this simplifies the process.
Also, if the SELECT_CC would be expanded in Legalize, it
can potentially end up with two copies of the condition
expression. By leaving it as SELECT+SETCC, the SELECT can be
expanded into two SELECTs that use a single SETCC.
The two comparisons are usually CSE'd, but depending on
when various expressions get legalized, the comparison
expression could involve calls to library functions, such
that the comparison expression may not be able to be CSE'd.
This will be needed by a future patch.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@77896 91177308-0d34-0410-b5e6-96231b3b80d8
in SelectionDAGLowering::visitTargetIntrinsic.
This removes a bit of special-case code for vector types. After staring
at it for a while, I managed to convince myself that it is not necessary.
The only case where TLI.getValueType() differs from MVT::getMVT is for iPTR,
so this code could potentially make a difference for a vector of pointers.
But, it looks like that is not supported. Calling TLI.getValueType() on
a vector of pointers leads to the following sequence of calls:
TargetLowering::getValueType
MVT::getMVT
MVT::getVectorVT(iPTR, num elements)
MVT::getExtendedVectorVT
MVT::getTypeForMVT for iPTR
assertion fails "Type is not extended!"
So, unless I'm really missing something, this bit of code is irrelevant to
the current version of LLVM, which is consistent with the fact that I don't
see this code in other similar places.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@77747 91177308-0d34-0410-b5e6-96231b3b80d8
shouldn't do AU.setPreservesCFG(), because even though CodeGen passes
don't modify the LLVM IR CFG, they may modify the MachineFunction CFG,
and passes like MachineLoop are registered with isCFGOnly set to true.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@77691 91177308-0d34-0410-b5e6-96231b3b80d8
failures when building assorted projects with clang.
--- Reverse-merging r77654 into '.':
U include/llvm/CodeGen/Passes.h
U include/llvm/CodeGen/MachineFunctionPass.h
U include/llvm/CodeGen/MachineFunction.h
U include/llvm/CodeGen/LazyLiveness.h
U include/llvm/CodeGen/SelectionDAGISel.h
D include/llvm/CodeGen/MachineFunctionAnalysis.h
U include/llvm/Function.h
U lib/Target/CellSPU/SPUISelDAGToDAG.cpp
U lib/Target/PowerPC/PPCISelDAGToDAG.cpp
U lib/CodeGen/LLVMTargetMachine.cpp
U lib/CodeGen/MachineVerifier.cpp
U lib/CodeGen/MachineFunction.cpp
U lib/CodeGen/PrologEpilogInserter.cpp
U lib/CodeGen/MachineLoopInfo.cpp
U lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp
D lib/CodeGen/MachineFunctionAnalysis.cpp
D lib/CodeGen/MachineFunctionPass.cpp
U lib/CodeGen/LiveVariables.cpp
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@77661 91177308-0d34-0410-b5e6-96231b3b80d8
When the return value is not used (i.e. only care about the value in the memory), x86 does not have to use add to implement these. Instead, it can use add, sub, inc, dec instructions with the "lock" prefix.
This is currently implemented using a bit of instruction selection trick. The issue is the target independent pattern produces one output and a chain and we want to map it into one that just output a chain. The current trick is to select it into a merge_values with the first definition being an implicit_def. The proper solution is to add new ISD opcodes for the no-output variant. DAG combiner can then transform the node before it gets to target node selection.
Problem #2 is we are adding a whole bunch of x86 atomic instructions when in fact these instructions are identical to the non-lock versions. We need a way to add target specific information to target nodes and have this information carried over to machine instructions. Asm printer (or JIT) can use this information to add the "lock" prefix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@77582 91177308-0d34-0410-b5e6-96231b3b80d8
it is highly specific to the object file that will be generated in the end,
this introduces a new TargetLoweringObjectFile interface that is implemented
for each of ELF/MachO/COFF/Alpha/PIC16 and XCore.
Though still is still a brutal and ugly refactoring, this is a major step
towards goodness.
This patch also:
1. fixes a bunch of dangling pointer problems in the PIC16 backend.
2. disables the TargetLowering copy ctor which PIC16 was accidentally using.
3. gets us closer to xcore having its own crazy target section flags and
pic16 not having to shadow sections with its own objects.
4. fixes wierdness where ELF targets would set CStringSection but not
CStringSection_. Factor the code better.
5. fixes some bugs in string lowering on ELF targets.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@77294 91177308-0d34-0410-b5e6-96231b3b80d8