Index: ELF/Arch/AArch64.cpp
===================================================================
--- ELF/Arch/AArch64.cpp
+++ ELF/Arch/AArch64.cpp
@@ -251,15 +251,19 @@
     or32AArch64Imm(Loc, getBits(Val, 0, 11));
     break;
   case R_AARCH64_LDST16_ABS_LO12_NC:
+    checkAlignment<2>(Loc, Val, Type);
     or32AArch64Imm(Loc, getBits(Val, 1, 11));
     break;
   case R_AARCH64_LDST32_ABS_LO12_NC:
+    checkAlignment<4>(Loc, Val, Type);
     or32AArch64Imm(Loc, getBits(Val, 2, 11));
     break;
   case R_AARCH64_LDST64_ABS_LO12_NC:
+    checkAlignment<8>(Loc, Val, Type);
     or32AArch64Imm(Loc, getBits(Val, 3, 11));
     break;
   case R_AARCH64_LDST128_ABS_LO12_NC:
+    checkAlignment<16>(Loc, Val, Type);
     or32AArch64Imm(Loc, getBits(Val, 4, 11));
     break;
   case R_AARCH64_MOVW_UABS_G0_NC:
Index: test/ELF/aarch64-lo12-alignment.s
===================================================================
--- /dev/null
+++ test/ELF/aarch64-lo12-alignment.s
@@ -0,0 +1,44 @@
+// RUN: llvm-mc -filetype=obj -triple=aarch64-none-linux %s -o %t
+// RUN: not ld.lld %t -o %t2 2>&1 | FileCheck %s
+
+// Test derived from a typical ODR violation where a global is declared
+// extern int but defined as a half or byte sized type.
+ .section .text
+ .globl _start
+ .type _start, %function
+// Access foo2 as if it were an aligned 32-bit int, expect an error as
+// foo is not aligned
+
+_start:
+ ldrb w2, [x0, #:lo12:foo1]  // Ok as no shift involved
+ ldrh w2, [x0, #:lo12:foo1]  // Error foo1 is not 2-byte aligned
+ ldrh w2, [x0, #:lo12:foo2]  // Ok as foo2 is 2-byte aligned
+ ldr  w2, [x0, #:lo12:foo2]  // Error foo2 is not 4-byte aligned
+ ldr  w2, [x0, #:lo12:foo4]  // Ok as foo4 is 4-byte aligned
+ ldr  x3, [x0, #:lo12:foo4]  // Error foo4 is not 8-byte aligned
+ ldr  x3, [x0, #:lo12:foo8]  // Ok as foo8 is 8-byte aligned
+ ldr  q0, [x0, #:lo12:foo8]  // Error foo8 is not 16-byte aligned
+ ldr  q0, [x0, #:lo12:foo16] // Ok as foo16 is 16-byte aligned
+
+ .section .data.bool, "a", @nobits
+ .balign 16
+ .globl foo16
+ .globl foo1
+ .globl foo2
+ .globl foo4
+ .globl foo8
+foo16:
+ .space 1
+foo1:
+ .space 1
+foo2:
+ .space 2
+foo4:
+ .space 4
+foo8:
+ .space 8
+
+// CHECK: improper alignment for relocation R_AARCH64_LDST16_ABS_LO12_NC
+// CHECK-NEXT: improper alignment for relocation R_AARCH64_LDST32_ABS_LO12_NC
+// CHECK-NEXT: improper alignment for relocation R_AARCH64_LDST64_ABS_LO12_NC
+// CHECK-NEXT: improper alignment for relocation R_AARCH64_LDST128_ABS_LO12_NC