[llvm] [AArch64] Fix correctness of FORM_TRANSPOSED_REG_TUPLE expansions (PR #205528)
Benjamin Maxwell via llvm-commits
llvm-commits at lists.llvm.org
Mon Jun 29 05:54:33 PDT 2026
================
@@ -1282,26 +1282,114 @@ bool AArch64ExpandPseudoImpl::expandMultiVecPseudo(
return true;
}
+struct Copy {
+ Register Dst;
+ Register Src;
+ int SrcTupleIdx = -1;
+};
+
bool AArch64ExpandPseudoImpl::expandFormTuplePseudo(
MachineBasicBlock &MBB, MachineBasicBlock::iterator MBBI,
MachineBasicBlock::iterator &NextMBBI, unsigned Size) {
assert((Size == 2 || Size == 4) && "Invalid Tuple Size");
- MachineInstr &MI = *MBBI;
- Register ReturnTuple = MI.getOperand(0).getReg();
const TargetRegisterInfo *TRI =
MBB.getParent()->getSubtarget().getRegisterInfo();
+
+ MachineInstr &MI = *MBBI;
+ DebugLoc DL = MI.getDebugLoc();
+ Register Tuple = MI.getOperand(0).getReg();
+
+ // Collect the copies required to form the tuple. Count if any tuple members
+ // appear as operands to the FORM_TRANSPOSED_REG_TUPLE. If a destination
+ // also appears as a source operand, we can't write to that register until
+ // we've handled the other copies that use it.
+
+ Copy Copies[4] = {};
+ unsigned Users[4] = {};
+
+ for (unsigned I = 0; I < Size; ++I) {
+ Register Dst = TRI->getSubReg(Tuple, AArch64::zsub0 + I);
+ Register Src = MI.getOperand(I + 1).getReg();
+
+ Copies[I] = {Dst, Src};
+
+ if (Dst != Src && TRI->isSubRegister(Tuple, Src)) {
+ int SrcTupleIdx = TRI->getSubRegIndex(Tuple, Src) - AArch64::zsub0;
+ Copies[I].SrcTupleIdx = SrcTupleIdx;
+ ++Users[SrcTupleIdx];
----------------
MacDue wrote:
Yes, `z19, z20, z21, z22 = pseudo z17, z19, z20, z21`.
This can be scheduled as:
```
z22 = copy z21
z21 = copy z20
z20 = copy z19
z19 = copy z17
```
Any example, where _all_ source operands are within the tuple is going to be a permutation or identity. Any example with at least one free register may be possible to schedule (though not in all cases, it could still contain a permutation).
https://github.com/llvm/llvm-project/pull/205528
More information about the llvm-commits
mailing list