diff --git a/llvm/lib/Transforms/Scalar/SeparateConstOffsetFromGEP.cpp b/llvm/lib/Transforms/Scalar/SeparateConstOffsetFromGEP.cpp
--- a/llvm/lib/Transforms/Scalar/SeparateConstOffsetFromGEP.cpp
+++ b/llvm/lib/Transforms/Scalar/SeparateConstOffsetFromGEP.cpp
@@ -211,6 +211,12 @@
                      cl::desc("Verify this pass produces no dead code"),
                      cl::Hidden);
 
+static cl::opt<bool> ForceEnableLowerGEP(
+    "seperate-const-offset-from-gep-force-enable-lower-gep", cl::init(false),
+    cl::desc("Enable the LowerGEP feature even for targets that disable it."
+	     "  This is mainly for testing purposes."),
+    cl::Hidden);
+
 namespace {
 
 /// A helper class for separating a constant offset from a GEP index.
@@ -348,7 +354,7 @@
   static char ID;
 
   SeparateConstOffsetFromGEPLegacyPass(bool LowerGEP = false)
-      : FunctionPass(ID), LowerGEP(LowerGEP) {
+      : FunctionPass(ID), LowerGEP(ForceEnableLowerGEP ? true : LowerGEP) {
     initializeSeparateConstOffsetFromGEPLegacyPassPass(
         *PassRegistry::getPassRegistry());
   }
@@ -377,7 +383,8 @@
       DominatorTree *DT, ScalarEvolution *SE, LoopInfo *LI,
       TargetLibraryInfo *TLI,
       function_ref<TargetTransformInfo &(Function &)> GetTTI, bool LowerGEP)
-      : DT(DT), SE(SE), LI(LI), TLI(TLI), GetTTI(GetTTI), LowerGEP(LowerGEP) {}
+      : DT(DT), SE(SE), LI(LI), TLI(TLI), GetTTI(GetTTI),
+        LowerGEP(ForceEnableLowerGEP ? true : LowerGEP) {}
 
   bool run(Function &F);
 
@@ -635,6 +642,30 @@
              /* ZeroExtended */ true, /* NonNegative */ false).zext(BitWidth);
   }
 
+  // If SignExtend and ZeroExtend are both true then the zext and sext
+  // operations are in this order:
+  // zext ( sext ( a op b)).
+  // We know this, because if the sext came first then the zext instruction
+  // would have changed SignExtended to false.
+  //
+  // We can't handle negative offsets that are going to be zero_extended later.
+  // For example:
+  // %1 = add i32 -15, %0
+  // %2 = zext i32 %1 to i64
+  //
+  // Will be transformed to:
+  //
+  // %1 = zext i32 -15 to i64
+  // %2 = zext i32 %0 to i64
+  // %3 = add i64 %1, %2
+  //
+  // Which is not the same.  Instead of -15 + %0, we end up with
+  // 4294967281 + %0
+  // TODO: It should be possible to fix this pass to be able to correctly handle
+  // negative offsets.
+  if (ZeroExtended && !SignExtended && !isa<SExtInst>(V) && ConstantOffset.isNegative())
+    ConstantOffset = APInt(BitWidth, 0);
+
   // If we found a non-zero constant offset, add it to the path for
   // rebuildWithoutConstOffset. Zero is a valid constant offset, but doesn't
   // help this optimization.
diff --git a/llvm/test/Transforms/SeparateConstOffsetFromGEP/pr62379-zeroext-negative.ll b/llvm/test/Transforms/SeparateConstOffsetFromGEP/pr62379-zeroext-negative.ll
new file mode 100644
--- /dev/null
+++ b/llvm/test/Transforms/SeparateConstOffsetFromGEP/pr62379-zeroext-negative.ll
@@ -0,0 +1,60 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt -S -passes=separate-const-offset-from-gep  -seperate-const-offset-from-gep-force-enable-lower-gep < %s | FileCheck %s
+
+; These tests check that this pass does not incorrectly handle negative offsets.
+
+@c = internal constant [4 x i32] [i32 0, i32 1, i32 2, i32 3]
+
+define i32 @add_negative_rhs(i32 %a, i32 %b, ptr %ptr) {
+; CHECK-LABEL: @add_negative_rhs(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[TMP0:%.*]] = add nuw nsw i32 -15, [[B:%.*]]
+; CHECK-NEXT:    [[TMP1:%.*]] = zext i32 [[TMP0]] to i64
+; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds [4 x i32], ptr [[PTR:%.*]], i64 0, i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP3:%.*]] = load i32, ptr [[TMP2]], align 4
+; CHECK-NEXT:    ret i32 [[TMP3]]
+;
+entry:
+  %0 = add nuw nsw i32 -15, %b
+  %1 = zext i32 %0 to i64
+  %2 = getelementptr inbounds [4 x i32], ptr %ptr, i64 0, i64 %1
+  %3 = load i32, ptr %2
+  ret i32 %3
+}
+
+define i32 @add_negative_lhs(i32 %a, i32 %b, ptr %ptr) {
+; CHECK-LABEL: @add_negative_lhs(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[TMP0:%.*]] = add nuw nsw i32 [[B:%.*]], -15
+; CHECK-NEXT:    [[TMP1:%.*]] = zext i32 [[TMP0]] to i64
+; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds [4 x i32], ptr [[PTR:%.*]], i64 0, i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP3:%.*]] = load i32, ptr [[TMP2]], align 4
+; CHECK-NEXT:    ret i32 [[TMP3]]
+;
+entry:
+  %0 = add nuw nsw i32 %b, -15
+  %1 = zext i32 %0 to i64
+  %2 = getelementptr inbounds [4 x i32], ptr %ptr, i64 0, i64 %1
+  %3 = load i32, ptr %2
+  ret i32 %3
+}
+
+
+define i32 @sub_positive(i32 %a, i32 %b, ptr %ptr) {
+; CHECK-LABEL: @sub_positive(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[TMP0:%.*]] = sub nuw nsw i32 15, [[A:%.*]]
+; CHECK-NEXT:    [[TMP1:%.*]] = sub nuw nsw i32 [[B:%.*]], [[TMP0]]
+; CHECK-NEXT:    [[TMP2:%.*]] = zext i32 [[TMP1]] to i64
+; CHECK-NEXT:    [[TMP3:%.*]] = getelementptr inbounds [4 x i32], ptr [[PTR:%.*]], i64 0, i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP4:%.*]] = load i32, ptr [[TMP3]], align 4
+; CHECK-NEXT:    ret i32 [[TMP4]]
+;
+entry:
+  %0 = sub nuw nsw i32 15, %a
+  %1 = sub nuw nsw i32 %b, %0
+  %2 = zext i32 %1 to i64
+  %3 = getelementptr inbounds [4 x i32], ptr %ptr, i64 0, i64 %2
+  %4 = load i32, ptr %3
+  ret i32 %4
+}