diff --git a/llvm/lib/Transforms/InstCombine/InstCombineAtomicRMW.cpp b/llvm/lib/Transforms/InstCombine/InstCombineAtomicRMW.cpp
--- a/llvm/lib/Transforms/InstCombine/InstCombineAtomicRMW.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineAtomicRMW.cpp
@@ -152,13 +152,5 @@
     return replaceOperand(RMWI, 1, ConstantFP::getNegativeZero(RMWI.getType()));
   }
 
-  // Check if the required ordering is compatible with an atomic load.
-  if (Ordering != AtomicOrdering::Acquire &&
-      Ordering != AtomicOrdering::Monotonic)
-    return nullptr;
-
-  LoadInst *Load = new LoadInst(RMWI.getType(), RMWI.getPointerOperand(), "",
-                                false, DL.getABITypeAlign(RMWI.getType()),
-                                Ordering, RMWI.getSyncScopeID());
-  return Load;
+  return nullptr;
 }
diff --git a/llvm/test/Transforms/InstCombine/atomicrmw.ll b/llvm/test/Transforms/InstCombine/atomicrmw.ll
--- a/llvm/test/Transforms/InstCombine/atomicrmw.ll
+++ b/llvm/test/Transforms/InstCombine/atomicrmw.ll
@@ -1,11 +1,12 @@
 ; RUN: opt -passes=instcombine -S -o - %s | FileCheck %s
-; Check that we can replace `atomicrmw <op> LHS, 0` with `load atomic LHS`.
-; This is possible when:
-; - <op> LHS, 0 == LHS
-; - the ordering of atomicrmw is compatible with a load (i.e., no release semantic)
+; Check that we don't replace `atomicrmw <op> LHS, 0` with `load atomic LHS`.
+; Doing that would lose the store semantic of the `atomicrmw` operation.
+; This may enable some other optimizations that would otherwise be illegal when
+; the store semantic was present (e.g., like dropping a fence).
 
 ; CHECK-LABEL: atomic_add_zero
-; CHECK-NEXT: %res = load atomic i32, ptr %addr monotonic, align 4
+; Idempotent atomicrmw are still canonicalized.
+; CHECK-NEXT: %res = atomicrmw or ptr %addr, i32 0 monotonic
 ; CHECK-NEXT: ret i32 %res
 define i32 @atomic_add_zero(ptr %addr) {
   %res = atomicrmw add ptr %addr, i32 0 monotonic
@@ -13,15 +14,16 @@
 }
 
 ; CHECK-LABEL: atomic_or_zero
-; CHECK-NEXT: %res = load atomic i32, ptr %addr monotonic, align 4
+; CHECK-NEXT: %res = atomicrmw or ptr %addr, i32 0 monotonic
 ; CHECK-NEXT: ret i32 %res
 define i32 @atomic_or_zero(ptr %addr) {
-  %res = atomicrmw add ptr %addr, i32 0 monotonic
+  %res = atomicrmw or ptr %addr, i32 0 monotonic
   ret i32 %res
 }
 
 ; CHECK-LABEL: atomic_sub_zero
-; CHECK-NEXT: %res = load atomic i32, ptr %addr monotonic, align 4
+; Idempotent atomicrmw are still canonicalized.
+; CHECK-NEXT: %res = atomicrmw or ptr %addr, i32 0 monotonic
 ; CHECK-NEXT: ret i32 %res
 define i32 @atomic_sub_zero(ptr %addr) {
   %res = atomicrmw sub ptr %addr, i32 0 monotonic
@@ -29,14 +31,16 @@
 }
 
 ; CHECK-LABEL: atomic_and_allones
-; CHECK-NEXT: %res = load atomic i32, ptr %addr monotonic, align 4
+; Idempotent atomicrmw are still canonicalized.
+; CHECK-NEXT: %res = atomicrmw or ptr %addr, i32 0 monotonic
 ; CHECK-NEXT: ret i32 %res
 define i32 @atomic_and_allones(ptr %addr) {
   %res = atomicrmw and ptr %addr, i32 -1 monotonic
   ret i32 %res
 }
 ; CHECK-LABEL: atomic_umin_uint_max
-; CHECK-NEXT: %res = load atomic i32, ptr %addr monotonic, align 4
+; Idempotent atomicrmw are still canonicalized.
+; CHECK-NEXT: %res = atomicrmw or ptr %addr, i32 0 monotonic
 ; CHECK-NEXT: ret i32 %res
 define i32 @atomic_umin_uint_max(ptr %addr) {
   %res = atomicrmw umin ptr %addr, i32 -1 monotonic
@@ -44,7 +48,8 @@
 }
 
 ; CHECK-LABEL: atomic_umax_zero
-; CHECK-NEXT: %res = load atomic i32, ptr %addr monotonic, align 4
+; Idempotent atomicrmw are still canonicalized.
+; CHECK-NEXT: %res = atomicrmw or ptr %addr, i32 0 monotonic
 ; CHECK-NEXT: ret i32 %res
 define i32 @atomic_umax_zero(ptr %addr) {
   %res = atomicrmw umax ptr %addr, i32 0 monotonic
@@ -52,7 +57,8 @@
 }
 
 ; CHECK-LABEL: atomic_min_smax_char
-; CHECK-NEXT: %res = load atomic i8, ptr %addr monotonic, align 1
+; Idempotent atomicrmw are still canonicalized.
+; CHECK-NEXT: %res = atomicrmw or ptr %addr, i8 0 monotonic
 ; CHECK-NEXT: ret i8 %res
 define i8 @atomic_min_smax_char(ptr %addr) {
   %res = atomicrmw min ptr %addr, i8 127 monotonic
@@ -60,7 +66,8 @@
 }
 
 ; CHECK-LABEL: atomic_max_smin_char
-; CHECK-NEXT: %res = load atomic i8, ptr %addr monotonic, align 1
+; Idempotent atomicrmw are still canonicalized.
+; CHECK-NEXT: %res = atomicrmw or ptr %addr, i8 0 monotonic
 ; CHECK-NEXT: ret i8 %res
 define i8 @atomic_max_smin_char(ptr %addr) {
   %res = atomicrmw max ptr %addr, i8 -128 monotonic
@@ -68,7 +75,8 @@
 }
 
 ; CHECK-LABEL: atomic_fsub
-; CHECK-NEXT: %res = load atomic float, ptr %addr monotonic, align 4
+; Idempotent atomicrmw are still canonicalized.
+; CHECK-NEXT: %res = atomicrmw fadd ptr %addr, float -0.000000e+00 monotonic
 ; CHECK-NEXT: ret float %res
 define float @atomic_fsub_zero(ptr %addr) {
   %res = atomicrmw fsub ptr %addr, float 0.0 monotonic
@@ -76,7 +84,7 @@
 }
 
 ; CHECK-LABEL: atomic_fadd
-; CHECK-NEXT: %res = load atomic float, ptr %addr monotonic, align 4
+; CHECK-NEXT: %res = atomicrmw fadd ptr %addr, float -0.000000e+00 monotonic
 ; CHECK-NEXT: ret float %res
 define float @atomic_fadd_zero(ptr %addr) {
   %res = atomicrmw fadd ptr %addr, float -0.0 monotonic
@@ -110,7 +118,7 @@
 
 ; Check that the transformation properly preserve the syncscope.
 ; CHECK-LABEL: atomic_syncscope
-; CHECK-NEXT: %res = load atomic i16, ptr %addr syncscope("some_syncscope") acquire, align 2
+; CHECK-NEXT:   %res = atomicrmw or ptr %addr, i16 0 syncscope("some_syncscope") acquire
 ; CHECK-NEXT: ret i16 %res
 define i16 @atomic_syncscope(ptr %addr) {
   %res = atomicrmw or ptr %addr, i16 0 syncscope("some_syncscope") acquire
@@ -138,7 +146,8 @@
 }
 
 ; CHECK-LABEL: atomic_xor_zero
-; CHECK-NEXT: %res = load atomic i16, ptr %addr monotonic, align 2
+; Idempotent atomicrmw are still canonicalized.
+; CHECK-NEXT: %res = atomicrmw or ptr %addr, i16 0 monotonic
 ; CHECK-NEXT: ret i16 %res
 define i16 @atomic_xor_zero(ptr %addr) {
   %res = atomicrmw xor ptr %addr, i16 0 monotonic