[Sparc] Account for bias in stack readjustment

JDevlieghere · JDevlieghere · commit 865de57bde1b · 2018-01-29T12:10:32.000Z
Summary: This was broken long ago in D12208, which failed to account for the fact that 64-bit SPARC uses a stack bias of 2047, and it is the *unbiased* value which should be aligned, not the biased one. This was seen to be an issue with Rust. Patch by: jrtc27 (James Clarke) Reviewers: jyknight, venkatra Reviewed By: jyknight Subscribers: jacob_hansen, JDevlieghere, fhahn, fedor.sergeev, llvm-commits Differential Revision: https://reviews.llvm.org/D39425 llvm-svn: 323643
diff --git a/llvm/lib/Target/Sparc/SparcFrameLowering.cpp b/llvm/lib/Target/Sparc/SparcFrameLowering.cpp
@@ -88,10 +88,11 @@ void SparcFrameLowering::emitPrologue(MachineFunction &MF,
 
   assert(&MF.front() == &MBB && "Shrink-wrapping not yet supported");
   MachineFrameInfo &MFI = MF.getFrameInfo();
+  const SparcSubtarget &Subtarget = MF.getSubtarget<SparcSubtarget>();
   const SparcInstrInfo &TII =
-      *static_cast<const SparcInstrInfo *>(MF.getSubtarget().getInstrInfo());
+      *static_cast<const SparcInstrInfo *>(Subtarget.getInstrInfo());
   const SparcRegisterInfo &RegInfo =
-      *static_cast<const SparcRegisterInfo *>(MF.getSubtarget().getRegisterInfo());
+      *static_cast<const SparcRegisterInfo *>(Subtarget.getRegisterInfo());
   MachineBasicBlock::iterator MBBI = MBB.begin();
   // Debug location must be unknown since the first debug location is used
   // to determine the end of the prologue.
@@ -141,7 +142,7 @@ void SparcFrameLowering::emitPrologue(MachineFunction &MF,
 
   // Adds the SPARC subtarget-specific spill area to the stack
   // size. Also ensures target-required alignment.
-  NumBytes = MF.getSubtarget<SparcSubtarget>().getAdjustedFrameSize(NumBytes);
+  NumBytes = Subtarget.getAdjustedFrameSize(NumBytes);
 
   // Finally, ensure that the size is sufficiently aligned for the
   // data on the stack.
@@ -176,9 +177,27 @@ void SparcFrameLowering::emitPrologue(MachineFunction &MF,
       .addCFIIndex(CFIIndex);
 
   if (NeedsStackRealignment) {
-    // andn %o6, MaxAlign-1, %o6
+    int64_t Bias = Subtarget.getStackPointerBias();
+    unsigned regUnbiased;
+    if (Bias) {
+      // This clobbers G1 which we always know is available here.
+      regUnbiased = SP::G1;
+      // add %o6, BIAS, %g1
+      BuildMI(MBB, MBBI, dl, TII.get(SP::ADDri), regUnbiased)
+        .addReg(SP::O6).addImm(Bias);
+    } else
+      regUnbiased = SP::O6;
+
+    // andn %regUnbiased, MaxAlign-1, %regUnbiased
     int MaxAlign = MFI.getMaxAlignment();
-    BuildMI(MBB, MBBI, dl, TII.get(SP::ANDNri), SP::O6).addReg(SP::O6).addImm(MaxAlign - 1);
+    BuildMI(MBB, MBBI, dl, TII.get(SP::ANDNri), regUnbiased)
+      .addReg(regUnbiased).addImm(MaxAlign - 1);
+
+    if (Bias) {
+      // add %g1, -BIAS, %o6
+      BuildMI(MBB, MBBI, dl, TII.get(SP::ADDri), SP::O6)
+        .addReg(regUnbiased).addImm(-Bias);
+    }
   }
 }
 
diff --git a/llvm/test/CodeGen/SPARC/stack-align.ll b/llvm/test/CodeGen/SPARC/stack-align.ll
@@ -1,4 +1,5 @@
-; RUN: llc -march=sparc < %s | FileCheck %s
+; RUN: llc -march=sparc < %s | FileCheck %s --check-prefixes=CHECK,CHECK32
+; RUN: llc -march=sparcv9 < %s | FileCheck %s --check-prefixes=CHECK,CHECK64
 declare void @stack_realign_helper(i32 %a, i32* %b)
 
 ;; This is a function where we have a local variable of 64-byte
@@ -7,10 +8,15 @@ declare void @stack_realign_helper(i32 %a, i32* %b)
 ;; the argument is accessed via frame pointer not stack pointer (to %o0).
 
 ;; CHECK-LABEL: stack_realign:
-;; CHECK:      andn %sp, 63, %sp
-;; CHECK-NEXT: ld [%fp+92], %o0
-;; CHECK-NEXT: call stack_realign_helper
-;; CHECK-NEXT: add %sp, 128, %o1
+;; CHECK32:      andn %sp, 63, %sp
+;; CHECK32-NEXT: ld [%fp+92], %o0
+;; CHECK64:      add %sp, 2047, %g1
+;; CHECK64-NEXT: andn %g1, 63, %g1
+;; CHECK64-NEXT: add %g1, -2047, %sp
+;; CHECK64-NEXT: ld [%fp+2227], %o0
+;; CHECK-NEXT:   call stack_realign_helper
+;; CHECK32-NEXT: add %sp, 128, %o1
+;; CHECK64-NEXT: add %sp, 2239, %o1
 
 define void @stack_realign(i32 %a, i32 %b, i32 %c, i32 %d, i32 %e, i32 %f, i32 %g) {
 entry:
