ARM: Introduce conservative load/store optimization mode

Most of the time ARM has the CCR.UNALIGN_TRP bit set to false which
means that unaligned loads/stores do not trap and even extensive testing
will not catch these bugs. However the multi/double variants are not
affected by this bit and will still trap. In effect a more aggressive
load/store optimization will break existing (bad) code.

These bugs do not necessarily manifest in the broken code where the
misaligned pointer is formed but often later in perfectly legal code
where it is accessed. This means recompiling system libraries (which
have no alignment bugs) with a newer compiler will break existing
applications (with alignment bugs) that worked before.

So (under protest) I implemented this safe mode which limits the
formation of multi/double operations to cases that are not affected by
user code (stack operations like spills/reloads) or cases where the
normal operations trap anyway (floating point load/stores). It is
disabled by default.

Differential Revision: http://reviews.llvm.org/D17015

llvm-svn: 262504

lib/Target/ARM/ARMLoadStoreOptimizer.cpp

@@ -60,6 +60,15 @@ STATISTIC(NumSTRD2STM,  "Number of strd instructions turned back into stm");
 STATISTIC(NumLDRD2LDR,  "Number of ldrd instructions turned back into ldr's");
 STATISTIC(NumSTRD2STR,  "Number of strd instructions turned back into str's");
 
+/// This switch disables formation of double/multi instructions that could
+/// potentially lead to (new) alignment traps even with CCR.UNALIGN_TRP
+/// disabled. This can be used to create libraries that are robust even when
+/// users provoke undefined behaviour by supplying misaligned pointers.
+/// \see mayCombineMisaligned()
+static cl::opt<bool>
+AssumeMisalignedLoadStores("arm-assume-misaligned-load-store", cl::Hidden,
+    cl::init(false), cl::desc("Be more conservative in ARM load/store opt"));
+
 namespace llvm {
 void initializeARMLoadStoreOptPass(PassRegistry &);
 }
@@ -916,6 +925,24 @@ static bool isValidLSDoubleOffset(int Offset) {
   return (Value % 4) == 0 && Value < 1024;
 }
 
+/// Return true for loads/stores that can be combined to a double/multi
+/// operation without increasing the requirements for alignment.
+static bool mayCombineMisaligned(const TargetSubtargetInfo &STI,
+                                 const MachineInstr &MI) {
+  // vldr/vstr trap on misaligned pointers anyway, forming vldm makes no
+  // difference.
+  unsigned Opcode = MI.getOpcode();
+  if (!isi32Load(Opcode) && !isi32Store(Opcode))
+    return true;
+
+  // Stack pointer alignment is out of the programmers control so we can trust
+  // SP-relative loads/stores.
+  if (getLoadStoreBaseOp(MI).getReg() == ARM::SP &&
+      STI.getFrameLowering()->getTransientStackAlignment() >= 4)
+    return true;
+  return false;
+}
+
 /// Find candidates for load/store multiple merge in list of MemOpQueueEntries.
 void ARMLoadStoreOpt::FormCandidates(const MemOpQueue &MemOps) {
   const MachineInstr *FirstMI = MemOps[0].MI;
@@ -954,6 +981,10 @@ void ARMLoadStoreOpt::FormCandidates(const MemOpQueue &MemOps) {
     if (PReg == ARM::SP || PReg == ARM::PC)
       CanMergeToLSMulti = CanMergeToLSDouble = false;
 
+    // Should we be conservative?
+    if (AssumeMisalignedLoadStores && !mayCombineMisaligned(*STI, *MI))
+      CanMergeToLSMulti = CanMergeToLSDouble = false;
+
     // Merge following instructions where possible.
     for (unsigned I = SIndex+1; I < EIndex; ++I, ++Count) {
       int NewOffset = MemOps[I].Offset;
@@ -1926,6 +1957,9 @@ INITIALIZE_PASS(ARMPreAllocLoadStoreOpt, "arm-prera-load-store-opt",
                 ARM_PREALLOC_LOAD_STORE_OPT_NAME, false, false)
 
 bool ARMPreAllocLoadStoreOpt::runOnMachineFunction(MachineFunction &Fn) {
+  if (AssumeMisalignedLoadStores)
+    return false;
+
   TD = &Fn.getDataLayout();
   STI = &static_cast<const ARMSubtarget &>(Fn.getSubtarget());
   TII = STI->getInstrInfo();

test/CodeGen/ARM/ldrd.ll

@@ -1,9 +1,11 @@
-; RUN: llc < %s -mtriple=thumbv7-apple-ios -mcpu=cortex-a8 -regalloc=fast -optimize-regalloc=0 -verify-machineinstrs | FileCheck %s -check-prefix=A8 -check-prefix=CHECK
+; RUN: llc < %s -mtriple=thumbv7-apple-ios -mcpu=cortex-a8 -regalloc=fast -optimize-regalloc=0 -verify-machineinstrs | FileCheck %s -check-prefix=A8 -check-prefix=CHECK -check-prefix=NORMAL
-; RUN: llc < %s -mtriple=thumbv7-apple-ios -mcpu=cortex-m3 -regalloc=fast -optimize-regalloc=0 | FileCheck %s -check-prefix=M3 -check-prefix=CHECK
+; RUN: llc < %s -mtriple=thumbv7-apple-ios -mcpu=cortex-m3 -regalloc=fast -optimize-regalloc=0 | FileCheck %s -check-prefix=M3 -check-prefix=CHECK -check-prefix=NORMAL
 ; rdar://6949835
-; RUN: llc < %s -mtriple=thumbv7-apple-ios -mcpu=cortex-a8 -regalloc=basic | FileCheck %s -check-prefix=BASIC -check-prefix=CHECK
+; RUN: llc < %s -mtriple=thumbv7-apple-ios -mcpu=cortex-a8 -regalloc=basic | FileCheck %s -check-prefix=BASIC -check-prefix=CHECK -check-prefix=NORMAL
-; RUN: llc < %s -mtriple=thumbv7-apple-ios -mcpu=cortex-a8 -regalloc=greedy | FileCheck %s -check-prefix=GREEDY -check-prefix=CHECK
+; RUN: llc < %s -mtriple=thumbv7-apple-ios -mcpu=cortex-a8 -regalloc=greedy | FileCheck %s -check-prefix=GREEDY -check-prefix=CHECK -check-prefix=NORMAL
-; RUN: llc < %s -mtriple=thumbv7-apple-ios -mcpu=swift | FileCheck %s -check-prefix=SWIFT -check-prefix=CHECK
+; RUN: llc < %s -mtriple=thumbv7-apple-ios -mcpu=swift | FileCheck %s -check-prefix=SWIFT -check-prefix=CHECK -check-prefix=NORMAL
+
+; RUN: llc < %s -mtriple=thumbv7-apple-ios -arm-assume-misaligned-load-store | FileCheck %s -check-prefix=CHECK -check-prefix=CONSERVATIVE
 
 ; Magic ARM pair hints works best with linearscan / fast.
 
@@ -15,12 +17,13 @@ declare void @use_i64(i64 %v)
 
 define void @test_ldrd(i64 %a) nounwind readonly {
 ; CHECK-LABEL: test_ldrd:
-; CHECK: bl{{x?}} _get_ptr
+; NORMAL: bl{{x?}} _get_ptr
 ; A8: ldrd r0, r1, [r0]
 ; Cortex-M3 errata 602117: LDRD with base in list may result in incorrect base
 ; register when interrupted or faulted.
 ; M3-NOT: ldrd r[[REGNUM:[0-9]+]], {{r[0-9]+}}, [r[[REGNUM]]]
-; CHECK: bl{{x?}} _use_i64
+; CONSERVATIVE-NOT: ldrd
+; NORMAL: bl{{x?}} _use_i64
   %ptr = call i64* @get_ptr()
   %v = load i64, i64* %ptr, align 8
   call void @use_i64(i64 %v)
@@ -39,11 +42,10 @@ define void @test_ldrd(i64 %a) nounwind readonly {
 ; evict another live range or use callee saved regs. Sorry if the test
 ; is sensitive to Regalloc changes, but it is an interesting case.
 ;
-; BASIC: @f
+; CHECK-LABEL: f:
 ; BASIC: %bb
 ; BASIC: ldrd
 ; BASIC: str
-; GREEDY: @f
 ; GREEDY: %bb
 ; GREEDY: ldrd
 ; GREEDY: str
@@ -76,14 +78,15 @@ return:                                           ; preds = %bb, %entry
 
 @TestVar = external global %struct.Test
 
+; CHECK-LABEL: Func1:
 define void @Func1() nounwind ssp {
-; CHECK: @Func1
 entry: 
 ; A8: movw [[BASE:r[0-9]+]], :lower16:{{.*}}TestVar{{.*}}
 ; A8: movt [[BASE]], :upper16:{{.*}}TestVar{{.*}}
 ; A8: ldrd [[FIELD1:r[0-9]+]], [[FIELD2:r[0-9]+]], {{\[}}[[BASE]], #4]
 ; A8-NEXT: add [[FIELD1]], [[FIELD2]]
 ; A8-NEXT: str [[FIELD1]], {{\[}}[[BASE]]{{\]}}
+; CONSERVATIVE-NOT: ldrd
   %orig_blocks = alloca [256 x i16], align 2
   %0 = bitcast [256 x i16]* %orig_blocks to i8*call void @llvm.lifetime.start(i64 512, i8* %0) nounwind
   %tmp1 = load i32, i32* getelementptr inbounds (%struct.Test, %struct.Test* @TestVar, i32 0, i32 1), align 4
@@ -97,8 +100,9 @@ entry:
 declare void @extfunc(i32, i32, i32, i32)
 
 ; CHECK-LABEL: Func2:
+; CONSERVATIVE-NOT: ldrd
 ; A8: ldrd
-; A8: blx
+; CHECK: bl{{x?}} _extfunc
 ; A8: pop
 define void @Func2(i32* %p) {
 entry:
@@ -116,12 +120,14 @@ entry:
 ; M3: strd r1, r0, [sp, #-8]!
 ; BASIC: strd r1, r0, [sp, #-8]!
 ; GREEDY: strd r0, r1, [sp, #-8]!
-; CHECK: @ InlineAsm Start
+; CONSERVATIVE: strd r0, r1, [sp, #-8]!
-; CHECK: @ InlineAsm End
+; NORMAL: @ InlineAsm Start
+; NORMAL: @ InlineAsm End
 ; A8: ldrd r2, r1, [sp]
 ; M3: ldrd r2, r1, [sp]
 ; BASIC: ldrd r2, r1, [sp]
 ; GREEDY: ldrd r1, r2, [sp]
+; CONSERVATIVE: ldrd r1, r2, [sp]
 ; CHECK: bl{{x?}} _extfunc
 define void @strd_spill_ldrd_reload(i32 %v0, i32 %v1) {
   ; force %v0 and %v1 to be spilled
@@ -134,8 +140,9 @@ define void @strd_spill_ldrd_reload(i32 %v0, i32 %v1) {
 declare void @extfunc2(i32*, i32, i32)
 
 ; CHECK-LABEL: ldrd_postupdate_dec:
-; CHECK: ldrd r1, r2, [r0], #-8
+; NORMAL: ldrd r1, r2, [r0], #-8
-; CHECK-NEXT: bl{{x?}} _extfunc
+; CONSERVATIVE-NOT: ldrd
+; CHECK: bl{{x?}} _extfunc
 define void @ldrd_postupdate_dec(i32* %p0) {
   %p0.1 = getelementptr i32, i32* %p0, i32 1
   %v0 = load i32, i32* %p0
@@ -146,8 +153,9 @@ define void @ldrd_postupdate_dec(i32* %p0) {
 }
 
 ; CHECK-LABEL: ldrd_postupdate_inc:
-; CHECK: ldrd r1, r2, [r0], #8
+; NORMAL: ldrd r1, r2, [r0], #8
-; CHECK-NEXT: bl{{x?}} _extfunc
+; CONSERVATIVE-NOT: ldrd
+; CHECK: bl{{x?}} _extfunc
 define void @ldrd_postupdate_inc(i32* %p0) {
   %p0.1 = getelementptr i32, i32* %p0, i32 1
   %v0 = load i32, i32* %p0
@@ -158,8 +166,9 @@ define void @ldrd_postupdate_inc(i32* %p0) {
 }
 
 ; CHECK-LABEL: strd_postupdate_dec:
-; CHECK: strd r1, r2, [r0], #-8
+; NORMAL: strd r1, r2, [r0], #-8
-; CHECK-NEXT: bx lr
+; CONSERVATIVE-NOT: strd
+; CHECK: bx lr
 define i32* @strd_postupdate_dec(i32* %p0, i32 %v0, i32 %v1) {
   %p0.1 = getelementptr i32, i32* %p0, i32 1
   store i32 %v0, i32* %p0
@@ -169,8 +178,9 @@ define i32* @strd_postupdate_dec(i32* %p0, i32 %v0, i32 %v1) {
 }
 
 ; CHECK-LABEL: strd_postupdate_inc:
-; CHECK: strd r1, r2, [r0], #8
+; NORMAL: strd r1, r2, [r0], #8
-; CHECK-NEXT: bx lr
+; CONSERVATIVE-NOT: strd
+; CHECK: bx lr
 define i32* @strd_postupdate_inc(i32* %p0, i32 %v0, i32 %v1) {
   %p0.1 = getelementptr i32, i32* %p0, i32 1
   store i32 %v0, i32* %p0

test/CodeGen/ARM/swift-vldm.ll

@@ -1,4 +1,5 @@
 ; RUN: llc < %s -mcpu=swift -mtriple=armv7s-apple-ios | FileCheck %s
+; RUN: llc < %s -arm-assume-misaligned-load-store -mcpu=swift -mtriple=armv7s-apple-ios | FileCheck %s
 
 ; Check that we avoid producing vldm instructions using d registers that
 ; begin in the most-significant half of a q register. These require more

test/CodeGen/Thumb2/thumb2-ldm.ll

@@ -1,12 +1,15 @@
-; RUN: llc < %s -mtriple=thumbv7-apple-ios -mattr=+thumb2 | FileCheck %s
+; RUN: llc < %s -mtriple=thumbv7-apple-ios -mattr=+thumb2 | FileCheck %s -check-prefix=ALL -check-prefix=CHECK
+; RUN: llc < %s -mtriple=thumbv7-apple-ios -mattr=+thumb2 -arm-assume-misaligned-load-store | FileCheck %s -check-prefix=ALL -check-prefix=CONSERVATIVE
 
 @X = external global [0 x i32]          ; <[0 x i32]*> [#uses=5]
 
 define i32 @t1() {
-; CHECK-LABEL: t1:
+; ALL-LABEL: t1:
-; CHECK: push {r7, lr}
+; ALL: push {r7, lr}
 ; CHECK: ldrd
-; CHECK: pop {r7, pc}
+; CONSERVATIVE-NOT: ldrd
+; CONSERVATIVE-NOT: ldm
+; ALL: pop {r7, pc}
         %tmp = load i32, i32* getelementptr ([0 x i32], [0 x i32]* @X, i32 0, i32 0)            ; <i32> [#uses=1]
         %tmp3 = load i32, i32* getelementptr ([0 x i32], [0 x i32]* @X, i32 0, i32 1)           ; <i32> [#uses=1]
         %tmp4 = call i32 @f1( i32 %tmp, i32 %tmp3 )                ; <i32> [#uses=1]
@@ -14,10 +17,12 @@ define i32 @t1() {
 }
 
 define i32 @t2() {
-; CHECK-LABEL: t2:
+; ALL-LABEL: t2:
-; CHECK: push {r7, lr}
+; ALL: push {r7, lr}
 ; CHECK: ldm
-; CHECK: pop {r7, pc}
+; CONSERVATIVE-NOT: ldrd
+; CONSERVATIVE-NOT: ldm
+; ALL: pop {r7, pc}
         %tmp = load i32, i32* getelementptr ([0 x i32], [0 x i32]* @X, i32 0, i32 2)            ; <i32> [#uses=1]
         %tmp3 = load i32, i32* getelementptr ([0 x i32], [0 x i32]* @X, i32 0, i32 3)           ; <i32> [#uses=1]
         %tmp5 = load i32, i32* getelementptr ([0 x i32], [0 x i32]* @X, i32 0, i32 4)           ; <i32> [#uses=1]
@@ -26,10 +31,12 @@ define i32 @t2() {
 }
 
 define i32 @t3() {
-; CHECK-LABEL: t3:
+; ALL-LABEL: t3:
-; CHECK: push {r7, lr}
+; ALL: push {r7, lr}
 ; CHECK: ldm
-; CHECK: pop {r7, pc}
+; CONSERVATIVE-NOT: ldrd
+; CONSERVATIVE-NOT: ldm
+; ALL: pop {r7, pc}
         %tmp = load i32, i32* getelementptr ([0 x i32], [0 x i32]* @X, i32 0, i32 1)            ; <i32> [#uses=1]
         %tmp3 = load i32, i32* getelementptr ([0 x i32], [0 x i32]* @X, i32 0, i32 2)           ; <i32> [#uses=1]
         %tmp5 = load i32, i32* getelementptr ([0 x i32], [0 x i32]* @X, i32 0, i32 3)           ; <i32> [#uses=1]
@@ -37,6 +44,34 @@ define i32 @t3() {
         ret i32 %tmp6
 }
 
+@g = common global i32* null
+
+define void @t4(i32 %a0, i32 %a1, i32 %a2) {
+; ALL-LABEL: t4:
+; ALL: stm.w sp, {r0, r1, r2}
+; ALL: blx _ext
+; ALL: ldm.w sp, {r0, r1, r2}
+; ALL: blx _f2
+  %arr = alloca [4 x i32], align 4
+  %p0 = getelementptr inbounds [4 x i32], [4 x i32]* %arr, i64 0, i64 0
+  %p1 = getelementptr inbounds [4 x i32], [4 x i32]* %arr, i64 0, i64 1
+  %p2 = getelementptr inbounds [4 x i32], [4 x i32]* %arr, i64 0, i64 2
+  store i32* %p0, i32** @g, align 8
+
+  store i32 %a0, i32* %p0, align 4
+  store i32 %a1, i32* %p1, align 4
+  store i32 %a2, i32* %p2, align 4
+  call void @ext()
+
+  %v0 = load i32, i32* %p0, align 4
+  %v1 = load i32, i32* %p1, align 4
+  %v2 = load i32, i32* %p2, align 4
+  call i32 @f2(i32 %v0, i32 %v1, i32 %v2)
+  ret void
+}
+
 declare i32 @f1(i32, i32)
 
 declare i32 @f2(i32, i32, i32)
+
+declare void @ext()