[GlobalISel][X86] Add G_CTLZ_ZERO_UNDEF legalization handling

diff --git a/llvm/lib/Target/X86/X86LegalizerInfo.cpp b/llvm/lib/Target/X86/X86LegalizerInfo.cpp
index 4fbcd00..4a42492 100644 (file)
--- a/llvm/lib/Target/X86/X86LegalizerInfo.cpp
+++ b/llvm/lib/Target/X86/X86LegalizerInfo.cpp
@@ -153,10 +153,10 @@ X86LegalizerInfo::X86LegalizerInfo(const X86Subtarget &STI,
       .widenScalarToNextPow2(1, /*Min=*/16)
       .clampScalar(1, s16, sMaxScalar);
 
-  // count leading zeros (LZCNT)
-  getActionDefinitionsBuilder(G_CTLZ)
+  // count leading zeros
+  getActionDefinitionsBuilder({G_CTLZ_ZERO_UNDEF, G_CTLZ})
       .legalIf([=](const LegalityQuery &Query) -> bool {
-        return Subtarget.hasLZCNT() &&
+        return (Query.Opcode == G_CTLZ_ZERO_UNDEF || Subtarget.hasLZCNT()) &&
                (typePairInSet(0, 1, {{s16, s16}, {s32, s32}})(Query) ||
                 (Is64Bit && typePairInSet(0, 1, {{s64, s64}})(Query)));
       })

diff --git a/llvm/test/CodeGen/X86/GlobalISel/legalize-leading-zeros-undef.mir b/llvm/test/CodeGen/X86/GlobalISel/legalize-leading-zeros-undef.mir
new file mode 100644 (file)
index 0000000..3d45750
--- /dev/null
+++ b/llvm/test/CodeGen/X86/GlobalISel/legalize-leading-zeros-undef.mir
@@ -0,0 +1,143 @@
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 2
+# RUN: llc -mtriple=x86_64-linux-gnu -run-pass=legalizer %s -o - | FileCheck %s --check-prefixes=CHECK,X64
+# RUN: llc -mtriple=i386-linux-gnu -run-pass=legalizer -global-isel-abort=2 -pass-remarks-missed='gisel*'  %s 2>%t -o - | FileCheck %s --check-prefixes=CHECK,X86
+# RUN: FileCheck -check-prefix=ERR32  %s < %t
+
+# ERR32: remark: <unknown>:0:0: unable to legalize instruction: %13:_(s32), %14:_(s1) = G_USUBE %8:_, %10:_, %12:_ (in function: test_ctlz35)
+# ERR32: remark: <unknown>:0:0: unable to legalize instruction: %1:_(s64) = G_SELECT %6:_(s1), %9:_, %10:_ (in function: test_ctlz64)
+
+# test count leading zeros for s16, s32, and s64
+
+---
+name:            test_ctlz35
+alignment:       16
+legalized:       false
+regBankSelected: false
+registers:
+  - { id: 0, class: _, preferred-register: '' }
+  - { id: 1, class: _, preferred-register: '' }
+body:             |
+  bb.1:
+    ; X64-LABEL: name: test_ctlz35
+    ; X64: [[DEF:%[0-9]+]]:_(s35) = IMPLICIT_DEF
+    ; X64-NEXT: [[ZEXT:%[0-9]+]]:_(s64) = G_ZEXT [[DEF]](s35)
+    ; X64-NEXT: [[CTLZ_ZERO_UNDEF:%[0-9]+]]:_(s64) = G_CTLZ_ZERO_UNDEF [[ZEXT]](s64)
+    ; X64-NEXT: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 29
+    ; X64-NEXT: [[SUB:%[0-9]+]]:_(s64) = G_SUB [[CTLZ_ZERO_UNDEF]], [[C]]
+    ; X64-NEXT: [[TRUNC:%[0-9]+]]:_(s35) = G_TRUNC [[SUB]](s64)
+    ; X64-NEXT: [[COPY:%[0-9]+]]:_(s35) = COPY [[TRUNC]](s35)
+    ; X64-NEXT: RET 0, implicit [[COPY]](s35)
+    ; X86-LABEL: name: test_ctlz35
+    ; X86: [[DEF:%[0-9]+]]:_(s35) = IMPLICIT_DEF
+    ; X86-NEXT: [[ZEXT:%[0-9]+]]:_(s64) = G_ZEXT [[DEF]](s35)
+    ; X86-NEXT: [[CTLZ_ZERO_UNDEF:%[0-9]+]]:_(s64) = G_CTLZ_ZERO_UNDEF [[ZEXT]](s64)
+    ; X86-NEXT: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 29
+    ; X86-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[CTLZ_ZERO_UNDEF]](s64)
+    ; X86-NEXT: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[C]](s64)
+    ; X86-NEXT: [[USUBO:%[0-9]+]]:_(s32), [[USUBO1:%[0-9]+]]:_(s1) = G_USUBO [[UV]], [[UV2]]
+    ; X86-NEXT: [[USUBE:%[0-9]+]]:_(s32), [[USUBE1:%[0-9]+]]:_(s1) = G_USUBE [[UV1]], [[UV3]], [[USUBO1]]
+    ; X86-NEXT: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[USUBO]](s32), [[USUBE]](s32)
+    ; X86-NEXT: [[TRUNC:%[0-9]+]]:_(s35) = G_TRUNC [[MV]](s64)
+    ; X86-NEXT: [[COPY:%[0-9]+]]:_(s35) = COPY [[TRUNC]](s35)
+    ; X86-NEXT: RET 0, implicit [[COPY]](s35)
+    %0:_(s35) = IMPLICIT_DEF
+    %1:_(s35) = G_CTLZ_ZERO_UNDEF %0
+    %2:_(s35) = COPY %1(s35)
+    RET 0, implicit %2
+...
+---
+name:            test_ctlz8
+alignment:       16
+legalized:       false
+regBankSelected: false
+registers:
+  - { id: 0, class: _, preferred-register: '' }
+  - { id: 1, class: _, preferred-register: '' }
+body:             |
+  bb.1:
+    ; CHECK-LABEL: name: test_ctlz8
+    ; CHECK: [[DEF:%[0-9]+]]:_(s8) = IMPLICIT_DEF
+    ; CHECK-NEXT: [[ZEXT:%[0-9]+]]:_(s16) = G_ZEXT [[DEF]](s8)
+    ; CHECK-NEXT: [[CTLZ_ZERO_UNDEF:%[0-9]+]]:_(s16) = G_CTLZ_ZERO_UNDEF [[ZEXT]](s16)
+    ; CHECK-NEXT: [[C:%[0-9]+]]:_(s16) = G_CONSTANT i16 8
+    ; CHECK-NEXT: [[SUB:%[0-9]+]]:_(s16) = G_SUB [[CTLZ_ZERO_UNDEF]], [[C]]
+    ; CHECK-NEXT: [[TRUNC:%[0-9]+]]:_(s8) = G_TRUNC [[SUB]](s16)
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:_(s8) = COPY [[TRUNC]](s8)
+    ; CHECK-NEXT: RET 0, implicit [[COPY]](s8)
+    %0:_(s8) = IMPLICIT_DEF
+    %1:_(s8) = G_CTLZ_ZERO_UNDEF %0
+    %2:_(s8) = COPY %1(s8)
+    RET 0, implicit %2
+...
+---
+name:            test_ctlz64
+alignment:       16
+legalized:       false
+regBankSelected: false
+registers:
+  - { id: 0, class: _, preferred-register: '' }
+  - { id: 1, class: _, preferred-register: '' }
+body:             |
+  bb.1:
+    ; X64-LABEL: name: test_ctlz64
+    ; X64: [[DEF:%[0-9]+]]:_(s64) = IMPLICIT_DEF
+    ; X64-NEXT: [[CTLZ_ZERO_UNDEF:%[0-9]+]]:_(s64) = G_CTLZ_ZERO_UNDEF [[DEF]](s64)
+    ; X64-NEXT: [[COPY:%[0-9]+]]:_(s64) = COPY [[CTLZ_ZERO_UNDEF]](s64)
+    ; X64-NEXT: RET 0, implicit [[COPY]](s64)
+    ; X86-LABEL: name: test_ctlz64
+    ; X86: [[DEF:%[0-9]+]]:_(s64) = IMPLICIT_DEF
+    ; X86-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[DEF]](s64)
+    ; X86-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
+    ; X86-NEXT: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(eq), [[UV1]](s32), [[C]]
+    ; X86-NEXT: [[CTLZ_ZERO_UNDEF:%[0-9]+]]:_(s64) = G_CTLZ_ZERO_UNDEF [[UV]](s32)
+    ; X86-NEXT: [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 32
+    ; X86-NEXT: [[ADD:%[0-9]+]]:_(s64) = G_ADD [[CTLZ_ZERO_UNDEF]], [[C1]]
+    ; X86-NEXT: [[CTLZ_ZERO_UNDEF1:%[0-9]+]]:_(s64) = G_CTLZ_ZERO_UNDEF [[UV1]](s32)
+    ; X86-NEXT: [[SELECT:%[0-9]+]]:_(s64) = G_SELECT [[ICMP]](s1), [[ADD]], [[CTLZ_ZERO_UNDEF1]]
+    ; X86-NEXT: [[COPY:%[0-9]+]]:_(s64) = COPY [[SELECT]](s64)
+    ; X86-NEXT: RET 0, implicit [[COPY]](s64)
+    %0:_(s64) = IMPLICIT_DEF
+    %1:_(s64) = G_CTLZ_ZERO_UNDEF %0
+    %2:_(s64) = COPY %1(s64)
+    RET 0, implicit %2
+...
+---
+name:            test_ctlz32
+alignment:       16
+legalized:       false
+regBankSelected: false
+registers:
+  - { id: 0, class: _, preferred-register: '' }
+  - { id: 1, class: _, preferred-register: '' }
+body:             |
+  bb.1:
+    ; CHECK-LABEL: name: test_ctlz32
+    ; CHECK: [[DEF:%[0-9]+]]:_(s32) = IMPLICIT_DEF
+    ; CHECK-NEXT: [[CTLZ_ZERO_UNDEF:%[0-9]+]]:_(s32) = G_CTLZ_ZERO_UNDEF [[DEF]](s32)
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY [[CTLZ_ZERO_UNDEF]](s32)
+    ; CHECK-NEXT: RET 0, implicit [[COPY]](s32)
+    %0:_(s32) = IMPLICIT_DEF
+    %1:_(s32) = G_CTLZ_ZERO_UNDEF %0
+    %2:_(s32) = COPY %1(s32)
+    RET 0, implicit %2
+...
+---
+name:            test_ctlz16
+alignment:       16
+legalized:       false
+regBankSelected: false
+registers:
+  - { id: 0, class: _, preferred-register: '' }
+  - { id: 1, class: _, preferred-register: '' }
+body:             |
+  bb.1:
+    ; CHECK-LABEL: name: test_ctlz16
+    ; CHECK: [[DEF:%[0-9]+]]:_(s16) = IMPLICIT_DEF
+    ; CHECK-NEXT: [[CTLZ_ZERO_UNDEF:%[0-9]+]]:_(s16) = G_CTLZ_ZERO_UNDEF [[DEF]](s16)
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:_(s16) = COPY [[CTLZ_ZERO_UNDEF]](s16)
+    ; CHECK-NEXT: RET 0, implicit [[COPY]](s16)
+    %0:_(s16) = IMPLICIT_DEF
+    %1:_(s16) = G_CTLZ_ZERO_UNDEF %0
+    %2:_(s16) = COPY %1(s16)
+    RET 0, implicit %2
+...