Fix register subclass handling in PPCInstrInfo::insertSelect

PPCInstrInfo::insertSelect and PPCInstrInfo::canInsertSelect were computing the
common subclass of the true and false inputs, and then selecting either the
32-bit or the 64-bit isel variant based on the result of calling
PPC::GPRCRegClass.hasSubClassEq(RC) and PPC::G8RCRegClass.hasSubClassEq(RC)
(where RC is the common subclass). Unfortunately, this is not quite right: if
we have something like this:

  %vreg8<def> = SELECT_CC_I8 %vreg4<kill>, %vreg7<kill>, %vreg6<kill>, 76;
    G8RC_and_G8RC_NOX0:%vreg8 CRRC:%vreg4 G8RC_NOX0:%vreg7,%vreg6

then the common subclass of G8RC_and_G8RC_NOX0 and G8RC_NOX0 is G8RC_NOX0, and
G8RC_NOX0 is not a subclass of G8RC (because it also contains the ZERO8
pseudo-register). As a result, we also need to check the common subclass
against GPRC_NOR0 and G8RC_NOX0 explicitly.

This had not been a problem for clients of insertSelect that called
canInsertSelect first (because it had a compensating mistake), but insertSelect
is also used by the PPC pseudo-instruction expander, and this error was causing
a problem in that context.

This problem was found by csmith.

llvm-svn: 186343

diff --git a/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp b/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp
index 1ad879d..375daee 100644 (file)
--- a/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp
+++ b/llvm/lib/Target/PowerPC/PPCInstrInfo.cpp
@@ -448,7 +448,9 @@ bool PPCInstrInfo::canInsertSelect(const MachineBasicBlock &MBB,
 
   // isel is for regular integer GPRs only.
   if (!PPC::GPRCRegClass.hasSubClassEq(RC) &&
-      !PPC::G8RCRegClass.hasSubClassEq(RC))
+      !PPC::GPRC_NOR0RegClass.hasSubClassEq(RC) &&
+      !PPC::G8RCRegClass.hasSubClassEq(RC) &&
+      !PPC::G8RC_NOX0RegClass.hasSubClassEq(RC))
     return false;
 
   // FIXME: These numbers are for the A2, how well they work for other cores is
@@ -478,12 +480,15 @@ void PPCInstrInfo::insertSelect(MachineBasicBlock &MBB,
   const TargetRegisterClass *RC =
     RI.getCommonSubClass(MRI.getRegClass(TrueReg), MRI.getRegClass(FalseReg));
   assert(RC && "TrueReg and FalseReg must have overlapping register classes");
-  assert((PPC::GPRCRegClass.hasSubClassEq(RC) ||
-          PPC::G8RCRegClass.hasSubClassEq(RC)) &&
+
+  bool Is64Bit = PPC::G8RCRegClass.hasSubClassEq(RC) ||
+                 PPC::G8RC_NOX0RegClass.hasSubClassEq(RC);
+  assert((Is64Bit ||
+          PPC::GPRCRegClass.hasSubClassEq(RC) ||
+          PPC::GPRC_NOR0RegClass.hasSubClassEq(RC)) &&
          "isel is for regular integer GPRs only");
 
-  unsigned OpCode =
-    PPC::GPRCRegClass.hasSubClassEq(RC) ? PPC::ISEL : PPC::ISEL8;
+  unsigned OpCode = Is64Bit ? PPC::ISEL8 : PPC::ISEL;
   unsigned SelectPred = Cond[0].getImm();
 
   unsigned SubIdx;

diff --git a/llvm/test/CodeGen/PowerPC/isel-rc-nox0.ll b/llvm/test/CodeGen/PowerPC/isel-rc-nox0.ll
new file mode 100644 (file)
index 0000000..7d425cc
--- /dev/null
+++ b/llvm/test/CodeGen/PowerPC/isel-rc-nox0.ll
@@ -0,0 +1,50 @@
+; RUN: llc -mtriple=powerpc64-unknown-linux-gnu -mcpu=pwr7 < %s
+target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-f128:128:128-v128:128:128-n32:64"
+target triple = "powerpc64-unknown-linux-gnu"
+
+@g_62 = external global [1 x [9 x i32]], align 4
+
+; Function Attrs: nounwind
+define void @main() #0 {
+entry:
+  br i1 undef, label %cond.true, label %for.cond1.preheader.i
+
+cond.true:                                        ; preds = %entry
+  br label %for.cond1.preheader.i
+
+for.cond1.preheader.i:                            ; preds = %for.cond1.preheader.i, %cond.true, %entry
+  br i1 undef, label %crc32_gentab.exit, label %for.cond1.preheader.i
+
+crc32_gentab.exit:                                ; preds = %for.cond1.preheader.i
+  %tobool.i19.i.i = icmp eq i32 undef, 0
+  %retval.0.i.i.i = select i1 %tobool.i19.i.i, i32* getelementptr inbounds ([1 x [9 x i32]]* @g_62, i64 0, i64 0, i64 6), i32* getelementptr inbounds ([1 x [9 x i32]]* @g_62, i64 0, i64 0, i64 8)
+  br label %for.cond1.preheader.i2961.i
+
+for.cond1.preheader.i2961.i:                      ; preds = %for.inc44.i2977.i, %crc32_gentab.exit
+  call void @llvm.memset.p0i8.i64(i8* bitcast ([1 x [9 x i32]]* @g_62 to i8*), i8 -1, i64 36, i32 4, i1 false) #1
+  %0 = load i32* %retval.0.i.i.i, align 4, !tbaa !0
+  %tobool.i2967.i = icmp eq i32 %0, 0
+  br label %for.body21.i2968.i
+
+for.body21.i2968.i:                               ; preds = %safe_mod_func_int32_t_s_s.exit.i2974.i, %for.cond1.preheader.i2961.i
+  br i1 %tobool.i2967.i, label %safe_mod_func_int32_t_s_s.exit.i2974.i, label %for.inc44.i2977.i
+
+safe_mod_func_int32_t_s_s.exit.i2974.i:           ; preds = %for.body21.i2968.i
+  br i1 undef, label %for.body21.i2968.i, label %for.inc44.i2977.i
+
+for.inc44.i2977.i:                                ; preds = %safe_mod_func_int32_t_s_s.exit.i2974.i, %for.body21.i2968.i
+  br i1 undef, label %func_80.exit2978.i, label %for.cond1.preheader.i2961.i
+
+func_80.exit2978.i:                               ; preds = %for.inc44.i2977.i
+  unreachable
+}
+
+; Function Attrs: nounwind
+declare void @llvm.memset.p0i8.i64(i8* nocapture, i8, i64, i32, i1) #1
+
+attributes #0 = { nounwind "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf"="true" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "ssp-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
+attributes #1 = { nounwind }
+
+!0 = metadata !{metadata !"int", metadata !1}
+!1 = metadata !{metadata !"omnipotent char", metadata !2}
+!2 = metadata !{metadata !"Simple C/C++ TBAA"}