xref: /freebsd/contrib/llvm-project/llvm/lib/Target/AArch64/AArch64ExpandImm.cpp (revision a0ca4af9455b844c5e094fc1b09b1390ffa979fc)
1 //===- AArch64ExpandImm.h - AArch64 Immediate Expansion -------------------===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 //
9 // This file implements the AArch64ExpandImm stuff.
10 //
11 //===----------------------------------------------------------------------===//
12 
13 #include "AArch64.h"
14 #include "AArch64ExpandImm.h"
15 #include "MCTargetDesc/AArch64AddressingModes.h"
16 
17 using namespace llvm;
18 using namespace llvm::AArch64_IMM;
19 
20 /// Helper function which extracts the specified 16-bit chunk from a
21 /// 64-bit value.
22 static uint64_t getChunk(uint64_t Imm, unsigned ChunkIdx) {
23   assert(ChunkIdx < 4 && "Out of range chunk index specified!");
24 
25   return (Imm >> (ChunkIdx * 16)) & 0xFFFF;
26 }
27 
28 /// Check whether the given 16-bit chunk replicated to full 64-bit width
29 /// can be materialized with an ORR instruction.
30 static bool canUseOrr(uint64_t Chunk, uint64_t &Encoding) {
31   Chunk = (Chunk << 48) | (Chunk << 32) | (Chunk << 16) | Chunk;
32 
33   return AArch64_AM::processLogicalImmediate(Chunk, 64, Encoding);
34 }
35 
36 /// Check for identical 16-bit chunks within the constant and if so
37 /// materialize them with a single ORR instruction. The remaining one or two
38 /// 16-bit chunks will be materialized with MOVK instructions.
39 ///
40 /// This allows us to materialize constants like |A|B|A|A| or |A|B|C|A| (order
41 /// of the chunks doesn't matter), assuming |A|A|A|A| can be materialized with
42 /// an ORR instruction.
43 static bool tryToreplicateChunks(uint64_t UImm,
44 				 SmallVectorImpl<ImmInsnModel> &Insn) {
45   using CountMap = DenseMap<uint64_t, unsigned>;
46 
47   CountMap Counts;
48 
49   // Scan the constant and count how often every chunk occurs.
50   for (unsigned Idx = 0; Idx < 4; ++Idx)
51     ++Counts[getChunk(UImm, Idx)];
52 
53   // Traverse the chunks to find one which occurs more than once.
54   for (const auto &Chunk : Counts) {
55     const uint64_t ChunkVal = Chunk.first;
56     const unsigned Count = Chunk.second;
57 
58     uint64_t Encoding = 0;
59 
60     // We are looking for chunks which have two or three instances and can be
61     // materialized with an ORR instruction.
62     if ((Count != 2 && Count != 3) || !canUseOrr(ChunkVal, Encoding))
63       continue;
64 
65     const bool CountThree = Count == 3;
66 
67     Insn.push_back({ AArch64::ORRXri, 0, Encoding });
68 
69     unsigned ShiftAmt = 0;
70     uint64_t Imm16 = 0;
71     // Find the first chunk not materialized with the ORR instruction.
72     for (; ShiftAmt < 64; ShiftAmt += 16) {
73       Imm16 = (UImm >> ShiftAmt) & 0xFFFF;
74 
75       if (Imm16 != ChunkVal)
76         break;
77     }
78 
79     // Create the first MOVK instruction.
80     Insn.push_back({ AArch64::MOVKXi, Imm16,
81 		     AArch64_AM::getShifterImm(AArch64_AM::LSL, ShiftAmt) });
82 
83     // In case we have three instances the whole constant is now materialized
84     // and we can exit.
85     if (CountThree)
86       return true;
87 
88     // Find the remaining chunk which needs to be materialized.
89     for (ShiftAmt += 16; ShiftAmt < 64; ShiftAmt += 16) {
90       Imm16 = (UImm >> ShiftAmt) & 0xFFFF;
91 
92       if (Imm16 != ChunkVal)
93         break;
94     }
95     Insn.push_back({ AArch64::MOVKXi, Imm16,
96                      AArch64_AM::getShifterImm(AArch64_AM::LSL, ShiftAmt) });
97     return true;
98   }
99 
100   return false;
101 }
102 
103 /// Check whether this chunk matches the pattern '1...0...'. This pattern
104 /// starts a contiguous sequence of ones if we look at the bits from the LSB
105 /// towards the MSB.
106 static bool isStartChunk(uint64_t Chunk) {
107   if (Chunk == 0 || Chunk == std::numeric_limits<uint64_t>::max())
108     return false;
109 
110   return isMask_64(~Chunk);
111 }
112 
113 /// Check whether this chunk matches the pattern '0...1...' This pattern
114 /// ends a contiguous sequence of ones if we look at the bits from the LSB
115 /// towards the MSB.
116 static bool isEndChunk(uint64_t Chunk) {
117   if (Chunk == 0 || Chunk == std::numeric_limits<uint64_t>::max())
118     return false;
119 
120   return isMask_64(Chunk);
121 }
122 
123 /// Clear or set all bits in the chunk at the given index.
124 static uint64_t updateImm(uint64_t Imm, unsigned Idx, bool Clear) {
125   const uint64_t Mask = 0xFFFF;
126 
127   if (Clear)
128     // Clear chunk in the immediate.
129     Imm &= ~(Mask << (Idx * 16));
130   else
131     // Set all bits in the immediate for the particular chunk.
132     Imm |= Mask << (Idx * 16);
133 
134   return Imm;
135 }
136 
137 /// Check whether the constant contains a sequence of contiguous ones,
138 /// which might be interrupted by one or two chunks. If so, materialize the
139 /// sequence of contiguous ones with an ORR instruction.
140 /// Materialize the chunks which are either interrupting the sequence or outside
141 /// of the sequence with a MOVK instruction.
142 ///
143 /// Assuming S is a chunk which starts the sequence (1...0...), E is a chunk
144 /// which ends the sequence (0...1...). Then we are looking for constants which
145 /// contain at least one S and E chunk.
146 /// E.g. |E|A|B|S|, |A|E|B|S| or |A|B|E|S|.
147 ///
148 /// We are also looking for constants like |S|A|B|E| where the contiguous
149 /// sequence of ones wraps around the MSB into the LSB.
150 static bool trySequenceOfOnes(uint64_t UImm,
151                               SmallVectorImpl<ImmInsnModel> &Insn) {
152   const int NotSet = -1;
153   const uint64_t Mask = 0xFFFF;
154 
155   int StartIdx = NotSet;
156   int EndIdx = NotSet;
157   // Try to find the chunks which start/end a contiguous sequence of ones.
158   for (int Idx = 0; Idx < 4; ++Idx) {
159     int64_t Chunk = getChunk(UImm, Idx);
160     // Sign extend the 16-bit chunk to 64-bit.
161     Chunk = (Chunk << 48) >> 48;
162 
163     if (isStartChunk(Chunk))
164       StartIdx = Idx;
165     else if (isEndChunk(Chunk))
166       EndIdx = Idx;
167   }
168 
169   // Early exit in case we can't find a start/end chunk.
170   if (StartIdx == NotSet || EndIdx == NotSet)
171     return false;
172 
173   // Outside of the contiguous sequence of ones everything needs to be zero.
174   uint64_t Outside = 0;
175   // Chunks between the start and end chunk need to have all their bits set.
176   uint64_t Inside = Mask;
177 
178   // If our contiguous sequence of ones wraps around from the MSB into the LSB,
179   // just swap indices and pretend we are materializing a contiguous sequence
180   // of zeros surrounded by a contiguous sequence of ones.
181   if (StartIdx > EndIdx) {
182     std::swap(StartIdx, EndIdx);
183     std::swap(Outside, Inside);
184   }
185 
186   uint64_t OrrImm = UImm;
187   int FirstMovkIdx = NotSet;
188   int SecondMovkIdx = NotSet;
189 
190   // Find out which chunks we need to patch up to obtain a contiguous sequence
191   // of ones.
192   for (int Idx = 0; Idx < 4; ++Idx) {
193     const uint64_t Chunk = getChunk(UImm, Idx);
194 
195     // Check whether we are looking at a chunk which is not part of the
196     // contiguous sequence of ones.
197     if ((Idx < StartIdx || EndIdx < Idx) && Chunk != Outside) {
198       OrrImm = updateImm(OrrImm, Idx, Outside == 0);
199 
200       // Remember the index we need to patch.
201       if (FirstMovkIdx == NotSet)
202         FirstMovkIdx = Idx;
203       else
204         SecondMovkIdx = Idx;
205 
206       // Check whether we are looking a chunk which is part of the contiguous
207       // sequence of ones.
208     } else if (Idx > StartIdx && Idx < EndIdx && Chunk != Inside) {
209       OrrImm = updateImm(OrrImm, Idx, Inside != Mask);
210 
211       // Remember the index we need to patch.
212       if (FirstMovkIdx == NotSet)
213         FirstMovkIdx = Idx;
214       else
215         SecondMovkIdx = Idx;
216     }
217   }
218   assert(FirstMovkIdx != NotSet && "Constant materializable with single ORR!");
219 
220   // Create the ORR-immediate instruction.
221   uint64_t Encoding = 0;
222   AArch64_AM::processLogicalImmediate(OrrImm, 64, Encoding);
223   Insn.push_back({ AArch64::ORRXri, 0, Encoding });
224 
225   const bool SingleMovk = SecondMovkIdx == NotSet;
226   Insn.push_back({ AArch64::MOVKXi, getChunk(UImm, FirstMovkIdx),
227                    AArch64_AM::getShifterImm(AArch64_AM::LSL,
228                                              FirstMovkIdx * 16) });
229 
230   // Early exit in case we only need to emit a single MOVK instruction.
231   if (SingleMovk)
232     return true;
233 
234   // Create the second MOVK instruction.
235   Insn.push_back({ AArch64::MOVKXi, getChunk(UImm, SecondMovkIdx),
236 	           AArch64_AM::getShifterImm(AArch64_AM::LSL,
237                                              SecondMovkIdx * 16) });
238 
239   return true;
240 }
241 
242 static uint64_t GetRunOfOnesStartingAt(uint64_t V, uint64_t StartPosition) {
243   uint64_t NumOnes = llvm::countr_one(V >> StartPosition);
244 
245   uint64_t UnshiftedOnes;
246   if (NumOnes == 64) {
247     UnshiftedOnes = ~0ULL;
248   } else {
249     UnshiftedOnes = (1ULL << NumOnes) - 1;
250   }
251   return UnshiftedOnes << StartPosition;
252 }
253 
254 static uint64_t MaximallyReplicateSubImmediate(uint64_t V, uint64_t Subset) {
255   uint64_t Result = Subset;
256 
257   // 64, 32, 16, 8, 4, 2
258   for (uint64_t i = 0; i < 6; ++i) {
259     uint64_t Rotation = 1ULL << (6 - i);
260     uint64_t Closure = Result | llvm::rotl<uint64_t>(Result, Rotation);
261     if (Closure != (Closure & V)) {
262       break;
263     }
264     Result = Closure;
265   }
266 
267   return Result;
268 }
269 
270 // Find the logical immediate that covers the most bits in RemainingBits,
271 // allowing for additional bits to be set that were set in OriginalBits.
272 static uint64_t maximalLogicalImmWithin(uint64_t RemainingBits,
273                                         uint64_t OriginalBits) {
274   // Find the first set bit.
275   uint32_t Position = llvm::countr_zero(RemainingBits);
276 
277   // Get the first run of set bits.
278   uint64_t FirstRun = GetRunOfOnesStartingAt(OriginalBits, Position);
279 
280   // Replicate the run as many times as possible, as long as the bits are set in
281   // RemainingBits.
282   uint64_t MaximalImm = MaximallyReplicateSubImmediate(OriginalBits, FirstRun);
283 
284   return MaximalImm;
285 }
286 
287 static std::optional<std::pair<uint64_t, uint64_t>>
288 decomposeIntoOrrOfLogicalImmediates(uint64_t UImm) {
289   if (UImm == 0 || ~UImm == 0)
290     return std::nullopt;
291 
292   // Make sure we don't have a run of ones split around the rotation boundary.
293   uint32_t InitialTrailingOnes = llvm::countr_one(UImm);
294   uint64_t RotatedBits = llvm::rotr<uint64_t>(UImm, InitialTrailingOnes);
295 
296   // Find the largest logical immediate that fits within the full immediate.
297   uint64_t MaximalImm1 = maximalLogicalImmWithin(RotatedBits, RotatedBits);
298 
299   // Remove all bits that are set by this mask.
300   uint64_t RemainingBits = RotatedBits & ~MaximalImm1;
301 
302   // Find the largest logical immediate covering the remaining bits, allowing
303   // for additional bits to be set that were also set in the original immediate.
304   uint64_t MaximalImm2 = maximalLogicalImmWithin(RemainingBits, RotatedBits);
305 
306   // If any bits still haven't been covered, then give up.
307   if (RemainingBits & ~MaximalImm2)
308     return std::nullopt;
309 
310   // Make sure to un-rotate the immediates.
311   return std::make_pair(rotl(MaximalImm1, InitialTrailingOnes),
312                         rotl(MaximalImm2, InitialTrailingOnes));
313 }
314 
315 // Attempt to expand an immediate as the ORR of a pair of logical immediates.
316 static bool tryOrrOfLogicalImmediates(uint64_t UImm,
317                                       SmallVectorImpl<ImmInsnModel> &Insn) {
318   auto MaybeDecomposition = decomposeIntoOrrOfLogicalImmediates(UImm);
319   if (MaybeDecomposition == std::nullopt)
320     return false;
321   uint64_t Imm1 = MaybeDecomposition->first;
322   uint64_t Imm2 = MaybeDecomposition->second;
323 
324   uint64_t Encoding1, Encoding2;
325   bool Imm1Success = AArch64_AM::processLogicalImmediate(Imm1, 64, Encoding1);
326   bool Imm2Success = AArch64_AM::processLogicalImmediate(Imm2, 64, Encoding2);
327 
328   if (Imm1Success && Imm2Success) {
329     // Create the ORR-immediate instructions.
330     Insn.push_back({AArch64::ORRXri, 0, Encoding1});
331     Insn.push_back({AArch64::ORRXri, 1, Encoding2});
332     return true;
333   }
334 
335   return false;
336 }
337 
338 // Attempt to expand an immediate as the AND of a pair of logical immediates.
339 // This is done by applying DeMorgan's law, under which logical immediates
340 // are closed.
341 static bool tryAndOfLogicalImmediates(uint64_t UImm,
342                                       SmallVectorImpl<ImmInsnModel> &Insn) {
343   // Apply DeMorgan's law to turn this into an ORR problem.
344   auto MaybeDecomposition = decomposeIntoOrrOfLogicalImmediates(~UImm);
345   if (MaybeDecomposition == std::nullopt)
346     return false;
347   uint64_t Imm1 = MaybeDecomposition->first;
348   uint64_t Imm2 = MaybeDecomposition->second;
349 
350   uint64_t Encoding1, Encoding2;
351   bool Imm1Success = AArch64_AM::processLogicalImmediate(~Imm1, 64, Encoding1);
352   bool Imm2Success = AArch64_AM::processLogicalImmediate(~Imm2, 64, Encoding2);
353 
354   if (Imm1Success && Imm2Success) {
355     // Materialize Imm1, the LHS of the AND
356     Insn.push_back({AArch64::ORRXri, 0, Encoding1});
357     // AND Imm1 with Imm2
358     Insn.push_back({AArch64::ANDXri, 1, Encoding2});
359     return true;
360   }
361 
362   return false;
363 }
364 
365 // Check whether the constant can be represented by exclusive-or of two 64-bit
366 // logical immediates. If so, materialize it with an ORR instruction followed
367 // by an EOR instruction.
368 //
369 // This encoding allows all remaining repeated byte patterns, and many repeated
370 // 16-bit values, to be encoded without needing four instructions. It can also
371 // represent some irregular bitmasks (although those would mostly only need
372 // three instructions otherwise).
373 static bool tryEorOfLogicalImmediates(uint64_t Imm,
374                                       SmallVectorImpl<ImmInsnModel> &Insn) {
375   // Determine the larger repetition size of the two possible logical
376   // immediates, by finding the repetition size of Imm.
377   unsigned BigSize = 64;
378 
379   do {
380     BigSize /= 2;
381     uint64_t Mask = (1ULL << BigSize) - 1;
382 
383     if ((Imm & Mask) != ((Imm >> BigSize) & Mask)) {
384       BigSize *= 2;
385       break;
386     }
387   } while (BigSize > 2);
388 
389   uint64_t BigMask = ((uint64_t)-1LL) >> (64 - BigSize);
390 
391   // Find the last bit of each run of ones, circularly. For runs which wrap
392   // around from bit 0 to bit 63, this is the bit before the most-significant
393   // zero, otherwise it is the least-significant bit in the run of ones.
394   uint64_t RunStarts = Imm & ~rotl<uint64_t>(Imm, 1);
395 
396   // Find the smaller repetition size of the two possible logical immediates by
397   // counting the number of runs of one-bits within the BigSize-bit value. Both
398   // sizes may be the same. The EOR may add one or subtract one from the
399   // power-of-two count that can be represented by a logical immediate, or it
400   // may be left unchanged.
401   int RunsPerBigChunk = popcount(RunStarts & BigMask);
402 
403   static const int8_t BigToSmallSizeTable[32] = {
404       -1, -1, 0,  1,  2,  2,  -1, 3,  3,  3,  -1, -1, -1, -1, -1, 4,
405       4,  4,  -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 5,
406   };
407 
408   int BigToSmallShift = BigToSmallSizeTable[RunsPerBigChunk];
409 
410   // Early-exit if the big chunk couldn't be a power-of-two number of runs
411   // EORed with another single run.
412   if (BigToSmallShift == -1)
413     return false;
414 
415   unsigned SmallSize = BigSize >> BigToSmallShift;
416 
417   // 64-bit values with a bit set every (1 << index) bits.
418   static const uint64_t RepeatedOnesTable[] = {
419       0xffffffffffffffff, 0x5555555555555555, 0x1111111111111111,
420       0x0101010101010101, 0x0001000100010001, 0x0000000100000001,
421       0x0000000000000001,
422   };
423 
424   // This RepeatedOnesTable lookup is a faster implementation of the division
425   // 0xffffffffffffffff / ((1 << SmallSize) - 1), and can be thought of as
426   // dividing the 64-bit value into fields of width SmallSize, and placing a
427   // one in the least significant bit of each field.
428   uint64_t SmallOnes = RepeatedOnesTable[countr_zero(SmallSize)];
429 
430   // Now we try to find the number of ones in each of the smaller repetitions,
431   // by looking at runs of ones in Imm. This can take three attempts, as the
432   // EOR may have changed the length of the first two runs we find.
433 
434   // Rotate a run of ones so we can count the number of trailing set bits.
435   int Rotation = countr_zero(RunStarts);
436   uint64_t RotatedImm = rotr<uint64_t>(Imm, Rotation);
437   for (int Attempt = 0; Attempt < 3; ++Attempt) {
438     unsigned RunLength = countr_one(RotatedImm);
439 
440     // Construct candidate values BigImm and SmallImm, such that if these two
441     // values are encodable, we have a solution. (SmallImm is constructed to be
442     // encodable, but this isn't guaranteed when RunLength >= SmallSize)
443     uint64_t SmallImm =
444         rotl<uint64_t>((SmallOnes << RunLength) - SmallOnes, Rotation);
445     uint64_t BigImm = Imm ^ SmallImm;
446 
447     uint64_t BigEncoding = 0;
448     uint64_t SmallEncoding = 0;
449     if (AArch64_AM::processLogicalImmediate(BigImm, 64, BigEncoding) &&
450         AArch64_AM::processLogicalImmediate(SmallImm, 64, SmallEncoding)) {
451       Insn.push_back({AArch64::ORRXri, 0, SmallEncoding});
452       Insn.push_back({AArch64::EORXri, 1, BigEncoding});
453       return true;
454     }
455 
456     // Rotate to the next run of ones
457     Rotation += countr_zero(rotr<uint64_t>(RunStarts, Rotation) & ~1);
458     RotatedImm = rotr<uint64_t>(Imm, Rotation);
459   }
460 
461   return false;
462 }
463 
464 /// \brief Expand a MOVi32imm or MOVi64imm pseudo instruction to a
465 /// MOVZ or MOVN of width BitSize followed by up to 3 MOVK instructions.
466 static inline void expandMOVImmSimple(uint64_t Imm, unsigned BitSize,
467 				      unsigned OneChunks, unsigned ZeroChunks,
468 				      SmallVectorImpl<ImmInsnModel> &Insn) {
469   const unsigned Mask = 0xFFFF;
470 
471   // Use a MOVZ or MOVN instruction to set the high bits, followed by one or
472   // more MOVK instructions to insert additional 16-bit portions into the
473   // lower bits.
474   bool isNeg = false;
475 
476   // Use MOVN to materialize the high bits if we have more all one chunks
477   // than all zero chunks.
478   if (OneChunks > ZeroChunks) {
479     isNeg = true;
480     Imm = ~Imm;
481   }
482 
483   unsigned FirstOpc;
484   if (BitSize == 32) {
485     Imm &= (1LL << 32) - 1;
486     FirstOpc = (isNeg ? AArch64::MOVNWi : AArch64::MOVZWi);
487   } else {
488     FirstOpc = (isNeg ? AArch64::MOVNXi : AArch64::MOVZXi);
489   }
490   unsigned Shift = 0;     // LSL amount for high bits with MOVZ/MOVN
491   unsigned LastShift = 0; // LSL amount for last MOVK
492   if (Imm != 0) {
493     unsigned LZ = llvm::countl_zero(Imm);
494     unsigned TZ = llvm::countr_zero(Imm);
495     Shift = (TZ / 16) * 16;
496     LastShift = ((63 - LZ) / 16) * 16;
497   }
498   unsigned Imm16 = (Imm >> Shift) & Mask;
499 
500   Insn.push_back({ FirstOpc, Imm16,
501                    AArch64_AM::getShifterImm(AArch64_AM::LSL, Shift) });
502 
503   if (Shift == LastShift)
504     return;
505 
506   // If a MOVN was used for the high bits of a negative value, flip the rest
507   // of the bits back for use with MOVK.
508   if (isNeg)
509     Imm = ~Imm;
510 
511   unsigned Opc = (BitSize == 32 ? AArch64::MOVKWi : AArch64::MOVKXi);
512   while (Shift < LastShift) {
513     Shift += 16;
514     Imm16 = (Imm >> Shift) & Mask;
515     if (Imm16 == (isNeg ? Mask : 0))
516       continue; // This 16-bit portion is already set correctly.
517 
518     Insn.push_back({ Opc, Imm16,
519                      AArch64_AM::getShifterImm(AArch64_AM::LSL, Shift) });
520   }
521 }
522 
523 /// Expand a MOVi32imm or MOVi64imm pseudo instruction to one or more
524 /// real move-immediate instructions to synthesize the immediate.
525 void AArch64_IMM::expandMOVImm(uint64_t Imm, unsigned BitSize,
526                                SmallVectorImpl<ImmInsnModel> &Insn) {
527   const unsigned Mask = 0xFFFF;
528 
529   // Scan the immediate and count the number of 16-bit chunks which are either
530   // all ones or all zeros.
531   unsigned OneChunks = 0;
532   unsigned ZeroChunks = 0;
533   for (unsigned Shift = 0; Shift < BitSize; Shift += 16) {
534     const unsigned Chunk = (Imm >> Shift) & Mask;
535     if (Chunk == Mask)
536       OneChunks++;
537     else if (Chunk == 0)
538       ZeroChunks++;
539   }
540 
541   // Prefer MOVZ/MOVN over ORR because of the rules for the "mov" alias.
542   if ((BitSize / 16) - OneChunks <= 1 || (BitSize / 16) - ZeroChunks <= 1) {
543     expandMOVImmSimple(Imm, BitSize, OneChunks, ZeroChunks, Insn);
544     return;
545   }
546 
547   // Try a single ORR.
548   uint64_t UImm = Imm << (64 - BitSize) >> (64 - BitSize);
549   uint64_t Encoding;
550   if (AArch64_AM::processLogicalImmediate(UImm, BitSize, Encoding)) {
551     unsigned Opc = (BitSize == 32 ? AArch64::ORRWri : AArch64::ORRXri);
552     Insn.push_back({ Opc, 0, Encoding });
553     return;
554   }
555 
556   // One to up three instruction sequences.
557   //
558   // Prefer MOVZ/MOVN followed by MOVK; it's more readable, and possibly the
559   // fastest sequence with fast literal generation.
560   if (OneChunks >= (BitSize / 16) - 2 || ZeroChunks >= (BitSize / 16) - 2) {
561     expandMOVImmSimple(Imm, BitSize, OneChunks, ZeroChunks, Insn);
562     return;
563   }
564 
565   assert(BitSize == 64 && "All 32-bit immediates can be expanded with a"
566                           "MOVZ/MOVK pair");
567 
568   // Try other two-instruction sequences.
569 
570   // 64-bit ORR followed by MOVK.
571   // We try to construct the ORR immediate in three different ways: either we
572   // zero out the chunk which will be replaced, we fill the chunk which will
573   // be replaced with ones, or we take the bit pattern from the other half of
574   // the 64-bit immediate. This is comprehensive because of the way ORR
575   // immediates are constructed.
576   for (unsigned Shift = 0; Shift < BitSize; Shift += 16) {
577     uint64_t ShiftedMask = (0xFFFFULL << Shift);
578     uint64_t ZeroChunk = UImm & ~ShiftedMask;
579     uint64_t OneChunk = UImm | ShiftedMask;
580     uint64_t RotatedImm = (UImm << 32) | (UImm >> 32);
581     uint64_t ReplicateChunk = ZeroChunk | (RotatedImm & ShiftedMask);
582     if (AArch64_AM::processLogicalImmediate(ZeroChunk, BitSize, Encoding) ||
583         AArch64_AM::processLogicalImmediate(OneChunk, BitSize, Encoding) ||
584         AArch64_AM::processLogicalImmediate(ReplicateChunk, BitSize,
585                                             Encoding)) {
586       // Create the ORR-immediate instruction.
587       Insn.push_back({ AArch64::ORRXri, 0, Encoding });
588 
589       // Create the MOVK instruction.
590       const unsigned Imm16 = getChunk(UImm, Shift / 16);
591       Insn.push_back({ AArch64::MOVKXi, Imm16,
592 		       AArch64_AM::getShifterImm(AArch64_AM::LSL, Shift) });
593       return;
594     }
595   }
596 
597   // Attempt to use a sequence of two ORR-immediate instructions.
598   if (tryOrrOfLogicalImmediates(Imm, Insn))
599     return;
600 
601   // Attempt to use a sequence of ORR-immediate followed by AND-immediate.
602   if (tryAndOfLogicalImmediates(Imm, Insn))
603     return;
604 
605   // Attempt to use a sequence of ORR-immediate followed by EOR-immediate.
606   if (tryEorOfLogicalImmediates(UImm, Insn))
607     return;
608 
609   // FIXME: Add more two-instruction sequences.
610 
611   // Three instruction sequences.
612   //
613   // Prefer MOVZ/MOVN followed by two MOVK; it's more readable, and possibly
614   // the fastest sequence with fast literal generation. (If neither MOVK is
615   // part of a fast literal generation pair, it could be slower than the
616   // four-instruction sequence, but we won't worry about that for now.)
617   if (OneChunks || ZeroChunks) {
618     expandMOVImmSimple(Imm, BitSize, OneChunks, ZeroChunks, Insn);
619     return;
620   }
621 
622   // Check for identical 16-bit chunks within the constant and if so materialize
623   // them with a single ORR instruction. The remaining one or two 16-bit chunks
624   // will be materialized with MOVK instructions.
625   if (BitSize == 64 && tryToreplicateChunks(UImm, Insn))
626     return;
627 
628   // Check whether the constant contains a sequence of contiguous ones, which
629   // might be interrupted by one or two chunks. If so, materialize the sequence
630   // of contiguous ones with an ORR instruction. Materialize the chunks which
631   // are either interrupting the sequence or outside of the sequence with a
632   // MOVK instruction.
633   if (BitSize == 64 && trySequenceOfOnes(UImm, Insn))
634     return;
635 
636   // We found no possible two or three instruction sequence; use the general
637   // four-instruction sequence.
638   expandMOVImmSimple(Imm, BitSize, OneChunks, ZeroChunks, Insn);
639 }
640