Sub-register lane masks are bitmasks that can be used to determine if
two sub-registers of a virtual register will overlap. For example, ARM's
ssub0 and ssub1 sub-register indices don't overlap each other, but both
overlap dsub0 and qsub0.
The lane masks will be accurate on most targets, but on targets that use
sub-register indexes in an irregular way, the masks may conservatively
report that two sub-register indices overlap when the eventually
allocated physregs don't.
Irregular register banks also mean that the bits in a lane mask can't be
mapped onto register units, but the concept is similar.
llvm-svn: 163630
Preserve the Composites map in the CodeGenSubRegIndex class so it can be
used to determine which sub-register indices can actually be composed.
llvm-svn: 163629
Apparently, NumSubRegIndices was completely unused before. Adjust it by
one to include the null subreg index, just like getNumRegs() includes
the null register.
llvm-svn: 163628
- This patch is inspired by the failure of the following code snippet
which is used to convert enumerable values into encoding bits to
improve the readability of td files.
class S<int s> {
bits<2> V = !if(!eq(s, 8), {0, 0},
!if(!eq(s, 16), {0, 1},
!if(!eq(s, 32), {1, 0},
!if(!eq(s, 64), {1, 1}, {?, ?}))));
}
Later, PR8330 is found to report not exactly the same bug relevant
issue to bit/bits values.
- Instead of resolving bit/bits values separately through
resolveBitReference(), this patch adds getBit() for all Inits and
resolves bit value by resolving plus getting the specified bit. This
unifies the resolving of bit with other values and removes redundant
logic for resolving bit only. In addition,
BitsInit::resolveReferences() is optimized to take advantage of this
origanization by resolving VarBitInit's variable reference first and
then getting bits from it.
- The type interference in '!if' operator is revised to support possible
combinations of int and bits/bit in MHS and RHS.
- As there may be illegal assignments from integer value to bit, says
assign 2 to a bit, but we only check this during instantiation in some
cases, e.g.
bit V = !if(!eq(x, 17), 0, 2);
Verbose diagnostic message is generated when invalid value is
resolveed to help locating the error.
- PR8330 is fixed as well.
llvm-svn: 163360
MatchInstructionImpl() function.
These values are used by the ConvertToMCInst() function to index into the
ConversionTable. The values are also needed to call the GetMCInstOperandNum()
function.
llvm-svn: 163101
AsmMatcherEmitter. This function maps inline assembly operands to MCInst
operands.
For example, '__asm mov j, eax' is represented by the follow MCInst:
<MCInst 1460 <MCOperand Reg:0> <MCOperand Imm:1> <MCOperand Reg:0>
<MCOperand Expr:(j)> <MCOperand Reg:0> <MCOperand Reg:43>>
The first 5 MCInst operands are a result of j matching as a memory operand
consisting of a BaseReg (Reg:0), MemScale (Imm:1), MemIndexReg(Reg:0),
Expr (Expr:(j), and a MemSegReg (Reg:0). The 6th MCInst operand represents
the eax register (Reg:43).
This translation is necessary to determine the Input and Output Exprs. If a
single asm operand maps to multiple MCInst operands, the index of the first
MCInst operand is returned. Ideally, it would return the operand we really
care out (i.e., the Expr:(j) in this case), but I haven't found an easy way
of doing this yet.
llvm-svn: 162920
Adding arbitrary records to ARM.td would break
basic-arm-instructions.s because selection of nop vs mov r0,r0 was
ambiguous (this will be tested by a subsequent addition to ARM.td).
An imperfect but sensible fix is to give precedence to match rules
that have more constraints.
llvm-svn: 162824
Previously, instructions without a primary patterns wouldn't get their
properties inferred. Now, we use all single-instruction patterns for
inference, including 'def : Pat<>' instances.
This causes a lot of instruction flags to change.
- Many instructions no longer have the UnmodeledSideEffects flag because
their flags are now inferred from a pattern.
- Instructions with intrinsics will get a mayStore flag if they already
have UnmodeledSideEffects and a mayLoad flag if they already have
mayStore. This is because intrinsics properties are linear.
- Instructions with atomic_load patterns get a mayStore flag because
atomic loads can't be reordered. The correct workaround is to create
pseudo-instructions instead of using normal loads. PR13693.
llvm-svn: 162614
Instructions are now only marked as variadic if they use variable_ops in
their ins list.
A variadic SDNode is typically used for call nodes that have the call
arguments as operands.
A variadic MachineInstr can actually encode a variable number of
operands, for example ARM's stm/ldm instructions. A call instruction
does not have to be variadic. The call argument registers are added as
implicit operands.
This change remove the MCID::Variadic flags from most call and return
instructions, allowing us to better verify their operands.
llvm-svn: 162599
It is now allowed to explicitly set hasSideEffects, mayStore, and
mayLoad on instructions with patterns.
Verify that the patterns are consistent with the explicit flags.
llvm-svn: 162569
Emit TableGen errors if guessInstructionProperties is 0 and
instruction properties can't be inferred from patterns.
Allow explicit instruction properties even when they can be inferred.
This patch doesn't change the TableGen output. Redundant properties
are not yet verified because the tree has errors.
llvm-svn: 162516
Currently, TableGen just guesses instruction properties when it can't
infer them form patterns.
This adds a guessInstructionProperties flag to the instruction set
definition that will be used to disable guessing. The flag is intended
as a migration aid. It will be removed again when no more targets need
their properties guessed.
llvm-svn: 162460
When reporting an error for a defm, we would previously only report the
location of the outer defm, which is not always where the error is.
Now we also print the location of the expanded multiclass defs:
lib/Target/X86/X86InstrSSE.td:2902:12: error: foo
defm ADD : basic_sse12_fp_binop_s<0x58, "add", fadd, SSE_ALU_ITINS_S>,
^
lib/Target/X86/X86InstrSSE.td:2801:11: note: instantiated from multiclass
defm PD : sse12_fp_packed<opc, !strconcat(OpcodeStr, "pd"), OpNode, VR128,
^
lib/Target/X86/X86InstrSSE.td:194:5: note: instantiated from multiclass
def rm : PI<opc, MRMSrcMem, (outs RC:$dst), (ins RC:$src1, x86memop:$src2),
^
llvm-svn: 162409
No change in interface or functionality. Purely under-the-hood
details of the generated function that change.
The X86 assembly parser is reduced in size by over 15% and ARM by
over 10%.
No performance change by my measurements.
llvm-svn: 162337
Select instructions pick one of two virtual registers based on a
condition, like x86 cmov. On targets like ARM that support predication,
selects can sometimes be eliminated by predicating the instruction
defining one of the operands.
Teach PeepholeOptimizer to recognize select instructions, and ask the
target to optimize them.
llvm-svn: 162059
TableGen sometimes synthesizes missing sub-register indexes. Emit these
indexes as enumerators in the target namespace along with the
user-defined ones.
Also take this opportunity to stop creating new Record objects for
synthetic indexes.
llvm-svn: 161964
Refactor the TableGen'erated fixed length disassemblmer to use a
table-driven state machine rather than a massive set of nested
switch() statements.
As a result, the ARM Disassembler (ARMDisassembler.cpp) builds much more
quickly and generates a smaller end result. For a Release+Asserts build on
a 16GB 3.4GHz i7 iMac w/ SSD:
Time to compile at -O2 (averaged w/ hot caches):
Previous: 35.5s
New: 8.9s
TEXT size:
Previous: 447,251
New: 297,661
Builds in 25% of the time previously required and generates code 66% of
the size.
Execution time of the disassembler is only slightly slower (7% disassembling
10 million ARM instructions, 19.6s vs 21.0s). The new implementation has
not yet been tuned, however, so the performance should almost certainly
be recoverable should it become a concern.
llvm-svn: 161888
These tables were indexed by [register][subreg index] which made them,
very large and sparse.
Replace them with lists of sub-register indexes that match the existing
lists of sub-registers. MCRI::getSubReg() becomes a very short linear
search, like getSubRegIndex() already was.
llvm-svn: 160843
Now that the weird X86 sub_ss and sub_sd sub-register indexes are gone,
there is no longer a need for the CompositeIndices construct in .td
files. Sub-register index composition can be specified on the
SubRegIndex itself using the ComposedOf field.
Also enforce unique names for sub-registers in TableGen. The same
sub-register cannot be available with multiple sub-register indexes.
llvm-svn: 160842
A standalone pattern defined in a multiclass expansion should handle
null_frag references just like patterns on instructions. Follow-up to
r160333.
llvm-svn: 160384
Define a 'null_frag' SDPatternOperator node, which if referenced in an
instruction Pattern, results in the pattern being collapsed to be as-if
'[]' had been specified instead. This allows supporting a multiclass
definition where some instaniations have ISel patterns associated and
others do not.
For example,
multiclass myMulti<RegisterClass rc, SDPatternOperator OpNode = null_frag> {
def _x : myI<(outs rc:), (ins rc:), []>;
def _r : myI<(outs rc:), (ins rc:), [(set rc:, (OpNode rc:))]>;
}
defm foo : myMulti<GRa, not>;
defm bar : myMulti<GRb>;
llvm-svn: 160333
Make sure the tblgen'erated asm matcher correctly returns numoperands+1
as the ErrorInfo when the problem was that there weren't enough operands
specified.
rdar://9142751
llvm-svn: 160144
subtarget CPU descriptions and support new features of
MachineScheduler.
MachineModel has three categories of data:
1) Basic properties for coarse grained instruction cost model.
2) Scheduler Read/Write resources for simple per-opcode and operand cost model (TBD).
3) Instruction itineraties for detailed per-cycle reservation tables.
These will all live side-by-side. Any subtarget can use any
combination of them. Instruction itineraries will not change in the
near term. In the long run, I expect them to only be relevant for
in-order VLIW machines that have complex contraints and require a
precise scheduling/bundling model. Once itineraries are only actively
used by VLIW-ish targets, they could be replaced by something more
appropriate for those targets.
This tablegen backend rewrite sets things up for introducing
MachineModel type #2: per opcode/operand cost model.
llvm-svn: 159891
The TargetInstrInfo::getNumMicroOps API does not change, but soon it
will be used by MachineScheduler. Now each subtarget can specify the
number of micro-ops per itinerary class. For ARM, this is currently
always dynamic (-1), because it is used for load/store multiple which
depends on the number of register operands.
Zero is now a valid number of micro-ops. This can be used for
nop pseudo-instructions or instructions that the hardware can squash
during dispatch.
llvm-svn: 159406
When generating selection tables for Pat instances, TableGen relied on
an output Instruction's Pattern field being set to infer whether a
chain should be added.
This patch adds additional logic to check various flag fields so that
correct code can be generated even if Pattern is unset.
llvm-svn: 159217
"Invalid operand" may be a completely correct diagnostic, but it's often
insufficiently specific to really help identify and fix the problem in
assembly source. Allow a target to specify a more-specific diagnostic kind
for each AsmOperandClass derived definition and use that to provide
more detailed diagnostics when an operant of that class resulted in a
match failure.
rdar://8987109
llvm-svn: 159050
Original commit message:
Allow up to 64 functional units per processor itinerary.
This patch changes the type used to hold the FU bitset from unsigned to uint64_t.
This will be needed for some upcoming PowerPC itineraries.
llvm-svn: 159027
This makes it explicit when ScoreboardHazardRecognizer will be used.
"GenericItineraries" would only make sense if it contained real
itinerary values and still required ScoreboardHazardRecognizer.
llvm-svn: 158963
This patch changes the type used to hold the FU bitset from unsigned to uint64_t.
This will be needed for some upcoming PowerPC itineraries.
llvm-svn: 158679
When returning a 'cannot match due to missing CPU features' error code,
if there are multiple potential matches with different feature sets,
return the smallest set of missing features from the alternatives as
that's most likely to be the one that's desired.
llvm-svn: 158673
LLVM is now -Wunused-private-field clean except for
- lib/MC/MCDisassembler/Disassembler.h. Not sure why it keeps all those unaccessible fields.
- gtest.
llvm-svn: 158096
There are some that I didn't remove this round because they looked like
obvious stubs. There are dead variables in gtest too, they should be
fixed upstream.
llvm-svn: 158090
This allows a subtarget to explicitly specify the issue width and
other properties without providing pipeline stage details for every
instruction.
llvm-svn: 157979
Each register unit has one or two root registers. The full set of
registers containing a given register unit can be computed as the union
of the root registers and their super-registers.
Provide an MCRegUnitRootIterator class to enumerate the roots.
llvm-svn: 157753
Register units are already used internally in TableGen to compute
register pressure sets and overlapping registers. This patch makes them
available to the code generators.
The register unit lists are differentially encoded so they can be reused
for many related registers. This keeps the total size of the lists below
200 bytes for most targets. ARM has the largest table at 560 bytes.
Add an MCRegUnitIterator for traversing the register unit lists. It
provides an abstract interface so the representation can be changed in
the future without changing all clients.
llvm-svn: 157650
This required light surgery on the assembler and disassembler
because the instructions use an uncommon encoding. They are
the only two instructions in x86 that use register operands
and two immediates.
llvm-svn: 157634
making it stronger and more sane.
Delete the code from tblgen that produced the old code.
Besides being a path forward in intrinsic sanity, this also eliminates a bunch of
machine generated code that was compiled into Function.o
llvm-svn: 157545
separate side table, using the handy SequenceToOffsetTable class. This encodes all
these weird things into another 256 bytes, allowing all intrinsics to be encoded this way.
llvm-svn: 156995
TableGen already computes register units as the basic unit of
interference. We can use that to compute the set of overlapping
registers.
This means that we can easily compute overlap sets for one register at a
time. There is no benefit to computing all registers at once.
llvm-svn: 156960
generated code (for Intrinsic::getType) into a table. This handles common cases right now,
but I plan to extend it to handle all cases and merge in type verification logic as well
in follow-on patches.
llvm-svn: 156905
Many targets always use the same bitwise encoding value for physical
registers in all (or most) instructions. Add this mapping to the
.td files and TableGen'erate the information and expose an accessor
in MCRegisterInfo.
patch by Tom Stellard.
llvm-svn: 156829
Besides the weight, we also want to store up to two root registers per
unit. Most units will have a single root, the leaf register they
represent. Units created for ad hoc aliasing get two roots: The two
aliasing registers.
The root registers can be used to compute the set of overlapping
registers.
llvm-svn: 156792
Register units can be used to compute if two registers overlap:
A overlaps B iff units(A) intersects units(B).
With this change, the above holds true even on targets that use ad hoc
aliasing (currently only ARM). This means that register units can be
used to implement regsOverlap() more efficiently, and the register
allocator can use the concept to model interference.
When there is no ad hoc aliasing, the register units correspond to the
maximal cliques in the register overlap graph. This is optimal, no other
register unit assignment can have fewer units.
With ad hoc aliasing, weird things are possible, and we don't try too
hard to compute the maximal cliques. The current approach is always
correct, and it works very well (probably optimally) as long as the ad
hoc aliasing doesn't have cliques larger than pairs. It seems unlikely
that any target would need more.
llvm-svn: 156763
The ad hoc aliasing specified in the 'Aliases' list in .td files is
currently only used by computeOverlaps(). It will soon be needed to
build accurate register units as well, so build the undirected graph in
CodeGenRegister::buildObjectGraph() instead.
Aliasing is a symmetric relationship with only one direction specified
in the .td files. Make sure both directions are represented in
getExplicitAliases().
llvm-svn: 156762
TableGen creates new register classes and sub-register indices based on
the sub-register structure present in the register bank. So far, it has
been doing that on a per-register basis, but that is not very efficient.
This patch teaches TableGen to compute topological signatures for
registers, and use that to reduce the amount of redundant computation.
Registers get the same TopoSig if they have identical sub-register
structure.
TopoSigs are not currently exposed outside TableGen.
llvm-svn: 156761
Don't compute the SuperRegs list until the sub-register graph is
completely finished. This guarantees that the list of super-registers is
properly topologically ordered, and has no duplicates.
llvm-svn: 156629
The sub-registers explicitly listed in SubRegs in the .td files form a
tree. In a complicated register bank, it is possible to have
sub-register relationships across sub-trees. For example, the ARM NEON
double vector Q0_Q1 is a tree:
Q0_Q1 = [Q0, Q1], Q0 = [D0, D1], Q1 = [D2, D3]
But we also define the DPair register D1_D2 = [D1, D2] which is fully
contained in Q0_Q1.
This patch teaches TableGen to find such sub-register relationships, and
assign sub-register indices to them. In the example, TableGen will
create a dsub_1_dsub_2 sub-register index, and add D1_D2 as a
sub-register of Q0_Q1.
This will eventually enable the coalescer to handle copies of skewed
sub-registers.
llvm-svn: 156587
The .td files specify a tree of sub-registers. Store that tree as
ExplicitSubRegs lists in CodeGenRegister instead of extracting it from
the Record when needed.
llvm-svn: 156555
This mapping is for internal use by TableGen. It will not be exposed in
the generated files.
Unfortunately, the mapping is not completely well-defined. The X86 xmm
registers appear with multiple sub-register indices in the ymm
registers. This is because of the odd idempotent sub_sd and sub_ss
sub-register indices. I hope to be able to eliminate them entirely, so
we can require the sub-registers to form a tree.
For now, just place the canonical sub_xmm index in the mapping, and
ignore the idempotents.
llvm-svn: 156519
Previously, if an instruction definition was missing the mnemonic,
the next line would just assert(). Issue a real diagnostic instead.
llvm-svn: 156263
This is still a topological ordering such that every register class gets
a smaller enum value than its sub-classes.
Placing the smaller spill sizes first makes a difference for the
super-register class bit masks. When looking for a super-register class,
we usually want the smallest possible kind of super-register. That is
now available as the first bit set in the bit mask.
llvm-svn: 156222
This manually enumerated list of super-register classes has been
superceeded by the automatically computed super-register class masks
available through SuperRegClassIterator.
llvm-svn: 156151
This is a pointer into one of the tables used by
getMatchingSuperRegClass(). It makes it possible to use a shared
implementation of that function.
llvm-svn: 156121
The RC->getSubClassMask() pointer now points to a sequence of register
class bit masks. The first bit mask is the normal sub-class mask. The
following masks are super-reg class masks used by
getMatchingSuperRegClass().
llvm-svn: 156120
Many register classes only have a few super-registers, so it is not
necessary to keep individual bit masks for all possible sub-register
indices.
llvm-svn: 156083
Some targets have no sub-registers at all. Use the TargetRegisterInfo
versions of composeSubRegIndices(), getSubClassWithSubReg(), and
getMatchingSuperRegClass() for those targets.
llvm-svn: 156075
When an instruction match is found, but the subtarget features it
requires are not available (missing floating point unit, or thumb vs arm
mode, for example), issue a diagnostic that identifies what the feature
mismatch is.
rdar://11257547
llvm-svn: 155499
Assembly matchers for instructions with a two-operand form. ARM is full
of these, for example:
add {Rd}, Rn, Rm // Rd is optional and is the same as Rn if omitted.
The property TwoOperandAliasConstraint on the instruction definition controls
when, and if, an alias will be formed. No explicit InstAlias definitions
are required.
rdar://11255754
llvm-svn: 155172
This is a new algorithm that finds sets of register units that can be
used to model registers pressure. This handles arbitrary, overlapping
register classes. Each register class is associated with a (small)
list of pressure sets. These are the dimensions of pressure affected
by the register class's liveness.
llvm-svn: 154374
This is a new algorithm that associates registers with weighted
register units to accuretely model their effect on register
pressure. This handles registers with multiple overlapping
subregisters. It is possible, but almost inconceivable that the
algorithm fails to find an exact solution for a target description. If
an exact solution cannot be found, an inexact, but reasonable solution
will be chosen.
llvm-svn: 154373
