diff --git a/llvm/lib/Target/X86/X86MCInstLower.cpp b/llvm/lib/Target/X86/X86MCInstLower.cpp
--- a/llvm/lib/Target/X86/X86MCInstLower.cpp
+++ b/llvm/lib/Target/X86/X86MCInstLower.cpp
@@ -2005,6 +2005,25 @@
     break;
   }
 
+  case X86::ENDBR32:
+  case X86::ENDBR64: {
+    // CurrentPatchableFunctionEntrySym can be CurrentFnBegin only for
+    // -fpatchable-function-entry=N,0. The entry MBB is guaranteed to be
+    // non-empty. If MI is the initial ENDBR, place the
+    // __patchable_function_entries label after ENDBR.
+    if (CurrentPatchableFunctionEntrySym &&
+        CurrentPatchableFunctionEntrySym == CurrentFnBegin &&
+        MI == &MF->front().front()) {
+      MCInst Inst;
+      MCInstLowering.Lower(MI, Inst);
+      EmitAndCountInstruction(Inst);
+      CurrentPatchableFunctionEntrySym = createTempSymbol("patch");
+      OutStreamer->EmitLabel(CurrentPatchableFunctionEntrySym);
+      return;
+    }
+    break;
+  }
+
   case X86::TAILJMPr:
   case X86::TAILJMPm:
   case X86::TAILJMPd:
diff --git a/llvm/test/CodeGen/AArch64/patchable-function-entry-bti.ll b/llvm/test/CodeGen/AArch64/patchable-function-entry-bti.ll
--- a/llvm/test/CodeGen/AArch64/patchable-function-entry-bti.ll
+++ b/llvm/test/CodeGen/AArch64/patchable-function-entry-bti.ll
@@ -15,7 +15,8 @@
 define void @f1() "patchable-function-entry"="1" "branch-target-enforcement" {
 ; CHECK-LABEL: f1:
 ; CHECK-NEXT: .Lfunc_begin1:
-; CHECK:      // %bb.0:
+; CHECK-NEXT: .cfi_startproc
+; CHECK-NEXT: // %bb.0:
 ; CHECK-NEXT:  hint #34
 ; CHECK-NEXT: .Lpatch0:
 ; CHECK-NEXT:  nop
@@ -33,7 +34,8 @@
 ; CHECK-NEXT:  nop
 ; CHECK-NEXT: f2_1:
 ; CHECK-NEXT: .Lfunc_begin2:
-; CHECK:      // %bb.0:
+; CHECK-NEXT: .cfi_startproc
+; CHECK-NEXT: // %bb.0:
 ; CHECK-NEXT:  hint #34
 ; CHECK-NEXT:  nop
 ; CHECK-NEXT:  ret
diff --git a/llvm/test/CodeGen/X86/patchable-function-entry-ibt.ll b/llvm/test/CodeGen/X86/patchable-function-entry-ibt.ll
--- a/llvm/test/CodeGen/X86/patchable-function-entry-ibt.ll
+++ b/llvm/test/CodeGen/X86/patchable-function-entry-ibt.ll
@@ -1,26 +1,37 @@
-; RUN: llc -mtriple=x86_64 %s -o - | FileCheck --check-prefixes=CHECK %s
+; RUN: llc -mtriple=i686 %s -o - | FileCheck --check-prefixes=CHECK,32 %s
+; RUN: llc -mtriple=x86_64 %s -o - | FileCheck --check-prefixes=CHECK,64 %s
 
 ;; -fpatchable-function-entry=0 -fcf-protection=branch
-define void @f0() "patchable-function-entry"="0" "branch-target-enforcement" {
+define void @f0() "patchable-function-entry"="0" {
 ; CHECK-LABEL: f0:
 ; CHECK-NEXT: .Lfunc_begin0:
-; CHECK:      # %bb.0:
-; CHECK-NEXT:  endbr64
-; CHECK-NEXT:  retq
+; CHECK-NEXT: .cfi_startproc
+; CHECK-NEXT: # %bb.0:
+; 32-NEXT:     endbr32
+; 64-NEXT:     endbr64
+; CHECK-NEXT:  ret
 ; CHECK-NOT:  .section __patchable_function_entries
   ret void
 }
 
 ;; -fpatchable-function-entry=1 -fcf-protection=branch
+;; For M=0, place the label .Lpatch0 after the initial ENDBR.
+;; .cfi_startproc should be placed at the function entry.
 define void @f1() "patchable-function-entry"="1" {
 ; CHECK-LABEL: f1:
 ; CHECK-NEXT: .Lfunc_begin1:
-; CHECK:       endbr64
+; CHECK-NEXT: .cfi_startproc
+; CHECK-NEXT: # %bb.0:
+; 32-NEXT:     endbr32
+; 64-NEXT:     endbr64
+; CHECK-NEXT: .Lpatch0:
 ; CHECK-NEXT:  nop
-; CHECK-NEXT:  retq
+; CHECK-NEXT:  ret
 ; CHECK:      .section __patchable_function_entries,"awo",@progbits,f1,unique,0
-; CHECK-NEXT: .p2align 3
-; CHECK-NEXT: .quad .Lfunc_begin1
+; 32-NEXT:    .p2align 2
+; 32-NEXT:    .long .Lpatch0
+; 64-NEXT:    .p2align 3
+; 64-NEXT:    .quad .Lpatch0
   ret void
 }
 
@@ -31,15 +42,44 @@
 ; CHECK-NEXT:  nop
 ; CHECK-NEXT: f2_1:
 ; CHECK-NEXT: .Lfunc_begin2:
-; CHECK:      # %bb.0:
-; CHECK-NEXT:  endbr64
+; CHECK-NEXT: .cfi_startproc
+; CHECK-NEXT: # %bb.0:
+; 32-NEXT:     endbr32
+; 64-NEXT:     endbr64
 ; CHECK-NEXT:  nop
-; CHECK-NEXT:  retq
+; CHECK-NEXT:  ret
 ; CHECK:      .Lfunc_end2:
 ; CHECK-NEXT: .size f2_1, .Lfunc_end2-f2_1
 ; CHECK:      .section __patchable_function_entries,"awo",@progbits,f1,unique,0
-; CHECK-NEXT: .p2align 3
-; CHECK-NEXT: .quad .Ltmp0
+; 32-NEXT:    .p2align 2
+; 32-NEXT:    .long .Ltmp0
+; 64-NEXT:    .p2align 3
+; 64-NEXT:    .quad .Ltmp0
+  ret void
+}
+
+;; -fpatchable-function-entry=1 -fcf-protection=branch
+;; For M=0, don't create .Lpatch0 if the initial instruction is not ENDBR,
+;; even if other basic blocks may have ENDBR.
+@buf = internal global [5 x i8*] zeroinitializer
+declare i32 @llvm.eh.sjlj.setjmp(i8*)
+
+define internal void @f1i() "patchable-function-entry"="1" {
+; CHECK-LABEL: f1i:
+; CHECK-NEXT: .Lfunc_begin3:
+; CHECK-NEXT: .cfi_startproc
+; CHECK-NEXT: # %bb.0:
+; CHECK-NEXT:  nop
+; CHECK-NOT:  .Lpatch0:
+;; Another basic block has ENDBR, but it doesn't affect our decision to not create .Lpatch0
+; CHECK:       endbr
+; CHECK:      .section __patchable_function_entries,"awo",@progbits,f1,unique,0
+; 32-NEXT:    .p2align 2
+; 32-NEXT:    .long .Lfunc_begin3
+; 64-NEXT:    .p2align 3
+; 64-NEXT:    .quad .Lfunc_begin3
+entry:
+  tail call i32 @llvm.eh.sjlj.setjmp(i8* bitcast ([5 x i8*]* @buf to i8*))
   ret void
 }