[llvm] de3cb95 - Fix a bug in memset formation with vectors of non-integral pointers

[Philip Reames via llvm-commits]

Author: Philip Reames
Date: 2020-10-01T16:11:11-07:00
New Revision: de3cb9548d77726186db2d384193e0565cb0afc5

URL: https://github.com/llvm/llvm-project/commit/de3cb9548d77726186db2d384193e0565cb0afc5
DIFF: https://github.com/llvm/llvm-project/commit/de3cb9548d77726186db2d384193e0565cb0afc5.diff

LOG: Fix a bug in memset formation with vectors of non-integral pointers

We were converting the non-integral store into a integer store which is not legal.

Added: 
    

Modified: 
    llvm/lib/Transforms/Scalar/LoopIdiomRecognize.cpp
    llvm/test/Transforms/LoopIdiom/non-integral-pointers.ll

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Transforms/Scalar/LoopIdiomRecognize.cpp b/llvm/lib/Transforms/Scalar/LoopIdiomRecognize.cpp
index 837ef869ccf0..38094377141f 100644
--- a/llvm/lib/Transforms/Scalar/LoopIdiomRecognize.cpp
+++ b/llvm/lib/Transforms/Scalar/LoopIdiomRecognize.cpp
@@ -455,11 +455,6 @@ LoopIdiomRecognize::isLegalStore(StoreInst *SI) {
   if (!SI->isUnordered())
     return LegalStoreKind::None;
 
-  // Don't convert stores of non-integral pointer types to memsets (which stores
-  // integers).
-  if (DL->isNonIntegralPointerType(SI->getValueOperand()->getType()))
-    return LegalStoreKind::None;
-
   // Avoid merging nontemporal stores.
   if (SI->getMetadata(LLVMContext::MD_nontemporal))
     return LegalStoreKind::None;
@@ -467,6 +462,11 @@ LoopIdiomRecognize::isLegalStore(StoreInst *SI) {
   Value *StoredVal = SI->getValueOperand();
   Value *StorePtr = SI->getPointerOperand();
 
+  // Don't convert stores of non-integral pointer types to memsets (which stores
+  // integers).
+  if (DL->isNonIntegralPointerType(StoredVal->getType()->getScalarType()))
+    return LegalStoreKind::None;
+
   // Reject stores that are so large that they overflow an unsigned.
   // When storing out scalable vectors we bail out for now, since the code
   // below currently only works for constant strides.

diff  --git a/llvm/test/Transforms/LoopIdiom/non-integral-pointers.ll b/llvm/test/Transforms/LoopIdiom/non-integral-pointers.ll
index 6846e8825394..c8271306535b 100644
--- a/llvm/test/Transforms/LoopIdiom/non-integral-pointers.ll
+++ b/llvm/test/Transforms/LoopIdiom/non-integral-pointers.ll
@@ -3,13 +3,12 @@
 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128-ni:4"
 target triple = "x86_64-unknown-linux-gnu"
 
+; LIR'ing stores of pointers with address space 3 is fine, since
+; they're integral pointers.
 define void @f_0(i8 addrspace(3)** %ptr) {
 ; CHECK-LABEL: @f_0(
 ; CHECK: call{{.*}}memset
 
-; LIR'ing stores of pointers with address space 3 is fine, since
-; they're integral pointers.
-
 entry:
   br label %for.body
 
@@ -25,13 +24,14 @@ for.end:
   ret void
 }
 
+; LIR'ing stores of pointers with address space 4 is not ok, since
+; they're non-integral pointers. NOTE: Zero is special value which
+; can be converted, if we add said handling here, convert this test
+; to use any non-null pointer.
 define void @f_1(i8 addrspace(4)** %ptr) {
 ; CHECK-LABEL: @f_1(
 ; CHECK-NOT: call{{.*}}memset
 
-; LIR'ing stores of pointers with address space 4 is not ok, since
-; they're non-integral pointers.
-
 entry:
   br label %for.body
 
@@ -46,3 +46,23 @@ for.body:
 for.end:
   ret void
 }
+
+; Same as previous case, but vector of non-integral pointers
+define void @f_2(i8 addrspace(4)** %ptr) {
+; CHECK-LABEL: @f_2(
+; CHECK-NOT: call{{.*}}memset
+entry:
+  br label %for.body
+
+for.body:
+  %indvar = phi i64 [ 0, %entry ], [ %indvar.next, %for.body ]
+  %arrayidx = getelementptr i8 addrspace(4)*, i8 addrspace(4)** %ptr, i64 %indvar
+  %addr = bitcast i8 addrspace(4)** %arrayidx to <2 x i8 addrspace(4)*>*
+  store <2 x i8 addrspace(4)*> zeroinitializer, <2 x i8 addrspace(4)*>* %addr, align 8
+  %indvar.next = add i64 %indvar, 2
+  %exitcond = icmp eq i64 %indvar.next, 10000
+  br i1 %exitcond, label %for.end, label %for.body
+
+for.end:
+  ret void
+}