Index: llvm/trunk/lib/CodeGen/AsmPrinter/AsmPrinter.cpp
===================================================================
--- llvm/trunk/lib/CodeGen/AsmPrinter/AsmPrinter.cpp
+++ llvm/trunk/lib/CodeGen/AsmPrinter/AsmPrinter.cpp
@@ -722,7 +722,9 @@
 }
 
 /// emitComments - Pretty-print comments for instructions.
-static void emitComments(const MachineInstr &MI, raw_ostream &CommentOS,
+/// It returns true iff the sched comment was emitted.
+///   Otherwise it returns false.
+static bool emitComments(const MachineInstr &MI, raw_ostream &CommentOS,
                          AsmPrinter *AP) {
   const MachineFunction *MF = MI.getMF();
   const TargetInstrInfo *TII = MF->getSubtarget().getInstrInfo();
@@ -766,12 +768,16 @@
     CommentOS << " Reload Reuse";
   }
 
-  if (Commented && AP->EnablePrintSchedInfo)
-    // If any comment was added above and we need sched info comment then
-    // add this new comment just after the above comment w/o "\n" between them.
-    CommentOS << " " << MF->getSubtarget().getSchedInfoStr(MI) << "\n";
-  else if (Commented)
+  if (Commented) {
+    if (AP->EnablePrintSchedInfo) {
+      // If any comment was added above and we need sched info comment then add
+      // this new comment just after the above comment w/o "\n" between them.
+      CommentOS << " " << MF->getSubtarget().getSchedInfoStr(MI) << "\n";
+      return true;
+    }
     CommentOS << "\n";
+  }
+  return false;
 }
 
 /// emitImplicitDef - This method emits the specified machine instruction
@@ -1013,8 +1019,10 @@
         }
       }
 
-      if (isVerbose())
-        emitComments(MI, OutStreamer->GetCommentOS(), this);
+      if (isVerbose() && emitComments(MI, OutStreamer->GetCommentOS(), this)) {
+        MachineInstr *MIP = const_cast<MachineInstr *>(&MI);
+        MIP->setAsmPrinterFlag(MachineInstr::NoSchedComment);
+      }
 
       switch (MI.getOpcode()) {
       case TargetOpcode::CFI_INSTRUCTION:
Index: llvm/trunk/lib/Target/X86/MCTargetDesc/X86BaseInfo.h
===================================================================
--- llvm/trunk/lib/Target/X86/MCTargetDesc/X86BaseInfo.h
+++ llvm/trunk/lib/Target/X86/MCTargetDesc/X86BaseInfo.h
@@ -59,7 +59,9 @@
     IP_HAS_AD_SIZE = 2,
     IP_HAS_REPEAT_NE = 4,
     IP_HAS_REPEAT = 8,
-    IP_HAS_LOCK = 16
+    IP_HAS_LOCK = 16,
+    NO_SCHED_INFO = 32 // Don't add sched comment to the current instr because
+                       // it was already added
   };
 } // end namespace X86;
 
Index: llvm/trunk/lib/Target/X86/X86MCInstLower.cpp
===================================================================
--- llvm/trunk/lib/Target/X86/X86MCInstLower.cpp
+++ llvm/trunk/lib/Target/X86/X86MCInstLower.cpp
@@ -100,7 +100,9 @@
 }
 
 void X86AsmPrinter::EmitAndCountInstruction(MCInst &Inst) {
-  OutStreamer->EmitInstruction(Inst, getSubtargetInfo(), EnablePrintSchedInfo);
+  OutStreamer->EmitInstruction(Inst, getSubtargetInfo(),
+                               EnablePrintSchedInfo &&
+                                   !(Inst.getFlags() & X86::NO_SCHED_INFO));
   SMShadowTracker.count(Inst, getSubtargetInfo(), CodeEmitter.get());
 }
 
@@ -2000,6 +2002,8 @@
 
   MCInst TmpInst;
   MCInstLowering.Lower(MI, TmpInst);
+  if (MI->getAsmPrinterFlag(MachineInstr::NoSchedComment))
+    TmpInst.setFlags(TmpInst.getFlags() | X86::NO_SCHED_INFO);
 
   // Stackmap shadows cannot include branch targets, so we can count the bytes
   // in a call towards the shadow, but must ensure that the no thread returns
Index: llvm/trunk/test/CodeGen/X86/avx512-schedule.ll
===================================================================
--- llvm/trunk/test/CodeGen/X86/avx512-schedule.ll
+++ llvm/trunk/test/CodeGen/X86/avx512-schedule.ll
@@ -8811,10 +8811,8 @@
 ; GENERIC-NEXT:    .cfi_def_cfa_offset 32
 ; GENERIC-NEXT:    vaddss %xmm0, %xmm0, %xmm0 # sched: [3:1.00]
 ; GENERIC-NEXT:    vmovaps %xmm0, (%rsp) # 16-byte Spill sched: [5:1.00]
-; GENERIC-NEXT:    # sched: [5:1.00]
 ; GENERIC-NEXT:    callq func_f32
 ; GENERIC-NEXT:    vbroadcastss (%rsp), %zmm0 # 16-byte Folded Reload sched: [5:1.00]
-; GENERIC-NEXT:    # sched: [5:1.00]
 ; GENERIC-NEXT:    addq $24, %rsp # sched: [1:0.33]
 ; GENERIC-NEXT:    retq # sched: [1:1.00]
 ;
@@ -8824,10 +8822,8 @@
 ; SKX-NEXT:    .cfi_def_cfa_offset 32
 ; SKX-NEXT:    vaddss %xmm0, %xmm0, %xmm0 # sched: [4:0.33]
 ; SKX-NEXT:    vmovaps %xmm0, (%rsp) # 16-byte Spill sched: [1:1.00]
-; SKX-NEXT:    # sched: [1:1.00]
 ; SKX-NEXT:    callq func_f32
 ; SKX-NEXT:    vbroadcastss (%rsp), %zmm0 # 16-byte Folded Reload sched: [8:0.50]
-; SKX-NEXT:    # sched: [8:0.50]
 ; SKX-NEXT:    addq $24, %rsp # sched: [1:0.25]
 ; SKX-NEXT:    retq # sched: [7:1.00]
   %a  = fadd float %x, %x
@@ -8845,10 +8841,8 @@
 ; GENERIC-NEXT:    .cfi_def_cfa_offset 32
 ; GENERIC-NEXT:    vaddsd %xmm0, %xmm0, %xmm0 # sched: [3:1.00]
 ; GENERIC-NEXT:    vmovapd %xmm0, (%rsp) # 16-byte Spill sched: [5:1.00]
-; GENERIC-NEXT:    # sched: [5:1.00]
 ; GENERIC-NEXT:    callq func_f64
 ; GENERIC-NEXT:    vbroadcastsd (%rsp), %zmm0 # 16-byte Folded Reload sched: [5:1.00]
-; GENERIC-NEXT:    # sched: [5:1.00]
 ; GENERIC-NEXT:    addq $24, %rsp # sched: [1:0.33]
 ; GENERIC-NEXT:    retq # sched: [1:1.00]
 ;
@@ -8858,10 +8852,8 @@
 ; SKX-NEXT:    .cfi_def_cfa_offset 32
 ; SKX-NEXT:    vaddsd %xmm0, %xmm0, %xmm0 # sched: [4:0.33]
 ; SKX-NEXT:    vmovapd %xmm0, (%rsp) # 16-byte Spill sched: [1:1.00]
-; SKX-NEXT:    # sched: [1:1.00]
 ; SKX-NEXT:    callq func_f64
 ; SKX-NEXT:    vbroadcastsd (%rsp), %zmm0 # 16-byte Folded Reload sched: [8:0.50]
-; SKX-NEXT:    # sched: [8:0.50]
 ; SKX-NEXT:    addq $24, %rsp # sched: [1:0.25]
 ; SKX-NEXT:    retq # sched: [7:1.00]
   %a  = fadd double %x, %x