Merge with main repository.

8 files changed, 28 insertions, 27 deletions

diff --git a/src/arch/arm/isa/formats/fp.isa b/src/arch/arm/isa/formats/fp.isa
index 812338c30..0cb27d7f1 100644
--- a/src/arch/arm/isa/formats/fp.isa
+++ b/src/arch/arm/isa/formats/fp.isa
@@ -561,20 +561,22 @@ let {{
                 }
             }
           case 0xa:
+            if (q)
+                return new Unknown(machInst);
             if (b) {
-                return decodeNeonUSThreeReg<VpminD, VpminQ>(
-                        q, u, size, machInst, vd, vn, vm);
+                return decodeNeonUSThreeUSReg<VpminD>(
+                        u, size, machInst, vd, vn, vm);
             } else {
-                return decodeNeonUSThreeReg<VpmaxD, VpmaxQ>(
-                        q, u, size, machInst, vd, vn, vm);
+                return decodeNeonUSThreeUSReg<VpmaxD>(
+                        u, size, machInst, vd, vn, vm);
             }
           case 0xb:
             if (b) {
-                if (u) {
+                if (u || q) {
                     return new Unknown(machInst);
                 } else {
-                    return decodeNeonUThreeReg<NVpaddD, NVpaddQ>(
-                            q, size, machInst, vd, vn, vm);
+                    return decodeNeonUThreeUSReg<NVpaddD>(
+                            size, machInst, vd, vn, vm);
                 }
             } else {
                 if (u) {
@@ -1542,7 +1544,7 @@ let {{
                 else
                     return new NVswpD<uint64_t>(machInst, vd, vm);
               case 0x1:
-                return decodeNeonUTwoMiscReg<NVtrnD, NVtrnQ>(
+                return decodeNeonUTwoMiscSReg<NVtrnD, NVtrnQ>(
                         q, size, machInst, vd, vm);
               case 0x2:
                 return decodeNeonUTwoMiscReg<NVuzpD, NVuzpQ>(
diff --git a/src/arch/arm/isa/insts/fp.isa b/src/arch/arm/isa/insts/fp.isa
index f82858bbd..b701995f4 100644
--- a/src/arch/arm/isa/insts/fp.isa
+++ b/src/arch/arm/isa/insts/fp.isa
@@ -447,7 +447,7 @@ let {{
     exec_output = ""
 
     singleSimpleCode = vfpEnabledCheckCode + '''
-        FPSCR fpscr = (FPSCR) FpscrExc;
+        FPSCR fpscr M5_VAR_USED = (FPSCR) FpscrExc;
         FpDest = %(op)s;
     '''
     singleCode = singleSimpleCode + '''
@@ -457,7 +457,7 @@ let {{
                 "%(func)s, fpscr.fz, fpscr.dn, fpscr.rMode)"
     singleUnaryOp = "unaryOp(fpscr, FpOp1, %(func)s, fpscr.fz, fpscr.rMode)"
     doubleCode = vfpEnabledCheckCode + '''
-        FPSCR fpscr = (FPSCR) FpscrExc;
+        FPSCR fpscr M5_VAR_USED = (FPSCR) FpscrExc;
         double dest = %(op)s;
         FpDestP0_uw = dblLow(dest);
         FpDestP1_uw = dblHi(dest);
diff --git a/src/arch/arm/isa/insts/macromem.isa b/src/arch/arm/isa/insts/macromem.isa
index 815d4c258..db36a3fff 100644
--- a/src/arch/arm/isa/insts/macromem.isa
+++ b/src/arch/arm/isa/insts/macromem.isa
@@ -563,15 +563,16 @@ let {{
 
 let {{
     exec_output = ''
-    for type in ('uint8_t', 'uint16_t', 'uint32_t'):
+    for typeSize in (8, 16, 32):
         for sRegs in 1, 2:
-            for dRegs in range(sRegs, 5):
+            for dRegs in range(sRegs, min(sRegs * 64 / typeSize + 1, 5)):
                 for format in ("MicroUnpackNeon%(sRegs)dto%(dRegs)dUop",
                                "MicroUnpackAllNeon%(sRegs)dto%(dRegs)dUop",
                                "MicroPackNeon%(dRegs)dto%(sRegs)dUop"):
                     Name = format % { "sRegs" : sRegs * 2,
                                       "dRegs" : dRegs * 2 }
-                    substDict = { "class_name" : Name, "targs" : type }
+                    substDict = { "class_name" : Name,
+                                  "targs" : "uint%d_t" % typeSize }
                     exec_output += MicroNeonExecDeclare.subst(substDict)
 }};
 
diff --git a/src/arch/arm/isa/insts/neon.isa b/src/arch/arm/isa/insts/neon.isa
index 9565ee14a..b1ad1eeb3 100644
--- a/src/arch/arm/isa/insts/neon.isa
+++ b/src/arch/arm/isa/insts/neon.isa
@@ -1616,10 +1616,8 @@ let {{
     threeEqualRegInst("vadd", "NVaddD", "SimdAddOp", unsignedTypes, 2, vaddCode)
     threeEqualRegInst("vadd", "NVaddQ", "SimdAddOp", unsignedTypes, 4, vaddCode)
 
-    threeEqualRegInst("vpadd", "NVpaddD", "SimdAddOp", unsignedTypes,
+    threeEqualRegInst("vpadd", "NVpaddD", "SimdAddOp", smallUnsignedTypes,
                       2, vaddCode, pairwise=True)
-    threeEqualRegInst("vpadd", "NVpaddQ", "SimdAddOp", unsignedTypes,
-                      4, vaddCode, pairwise=True)
     vaddlwCode = '''
         destElem = (BigElement)srcElem1 + (BigElement)srcElem2;
     '''
@@ -2110,11 +2108,9 @@ let {{
     '''
     threeRegLongInst("vmull", "Vmullp", "SimdMultOp", smallUnsignedTypes, vmullpCode)
 
-    threeEqualRegInst("vpmax", "VpmaxD", "SimdCmpOp", allTypes, 2, vmaxCode, pairwise=True)
-    threeEqualRegInst("vpmax", "VpmaxQ", "SimdCmpOp", allTypes, 4, vmaxCode, pairwise=True)
+    threeEqualRegInst("vpmax", "VpmaxD", "SimdCmpOp", smallTypes, 2, vmaxCode, pairwise=True)
 
-    threeEqualRegInst("vpmin", "VpminD", "SimdCmpOp", allTypes, 2, vminCode, pairwise=True)
-    threeEqualRegInst("vpmin", "VpminQ", "SimdCmpOp", allTypes, 4, vminCode, pairwise=True)
+    threeEqualRegInst("vpmin", "VpminD", "SimdCmpOp", smallTypes, 2, vminCode, pairwise=True)
 
     vqdmulhCode = '''
         FPSCR fpscr = (FPSCR) FpscrQc;
@@ -3137,8 +3133,10 @@ let {{
             destReg.elements[i + 1] = mid;
         }
     '''
-    twoRegMiscScramble("vtrn", "NVtrnD", "SimdAluOp", unsignedTypes, 2, vtrnCode)
-    twoRegMiscScramble("vtrn", "NVtrnQ", "SimdAluOp", unsignedTypes, 4, vtrnCode)
+    twoRegMiscScramble("vtrn", "NVtrnD", "SimdAluOp",
+            smallUnsignedTypes, 2, vtrnCode)
+    twoRegMiscScramble("vtrn", "NVtrnQ", "SimdAluOp",
+            smallUnsignedTypes, 4, vtrnCode)
 
     vuzpCode = '''
         Element mid[eCount];
diff --git a/src/arch/arm/isa/templates/mem.isa b/src/arch/arm/isa/templates/mem.isa
index a00114409..a4a740f89 100644
--- a/src/arch/arm/isa/templates/mem.isa
+++ b/src/arch/arm/isa/templates/mem.isa
@@ -1112,7 +1112,7 @@ def template LoadRegConstructor {{
                  (IntRegIndex)_index)
     {
         %(constructor)s;
-        bool conditional = false;
+        bool conditional M5_VAR_USED = false;
         if (!(condCode == COND_AL || condCode == COND_UC)) {
             conditional = true;
             for (int x = 0; x < _numDestRegs; x++) {
@@ -1166,7 +1166,7 @@ def template LoadImmConstructor {{
                  (IntRegIndex)_dest, (IntRegIndex)_base, _add, _imm)
     {
         %(constructor)s;
-        bool conditional = false;
+        bool conditional M5_VAR_USED = false;
         if (!(condCode == COND_AL || condCode == COND_UC)) {
             conditional = true;
             for (int x = 0; x < _numDestRegs; x++) {
diff --git a/src/arch/arm/linux/linux.hh b/src/arch/arm/linux/linux.hh
index 33e48fc93..40d586aaf 100644
--- a/src/arch/arm/linux/linux.hh
+++ b/src/arch/arm/linux/linux.hh
@@ -91,6 +91,7 @@ class ArmLinux : public Linux
 
     /// For mmap().
     static const unsigned TGT_MAP_ANONYMOUS = 0x20;
+    static const unsigned TGT_MAP_FIXED     = 0x10;
 
     //@{
     /// For getrusage().
diff --git a/src/arch/arm/linux/process.cc b/src/arch/arm/linux/process.cc
index f17749252..c65962d00 100644
--- a/src/arch/arm/linux/process.cc
+++ b/src/arch/arm/linux/process.cc
@@ -503,7 +503,7 @@ void
 ArmLinuxProcess::initState()
 {
     ArmLiveProcess::initState();
-    pTable->allocate(commPage, PageBytes);
+    allocateMem(commPage, PageBytes);
     ThreadContext *tc = system->getThreadContext(contextIds[0]);
 
     uint8_t swiNeg1[] = {
diff --git a/src/arch/arm/process.cc b/src/arch/arm/process.cc
index c3b02744e..aa5d7dfce 100644
--- a/src/arch/arm/process.cc
+++ b/src/arch/arm/process.cc
@@ -251,8 +251,7 @@ ArmLiveProcess::argsInit(int intSize, int pageSize)
     stack_size = stack_base - stack_min;
 
     // map memory
-    pTable->allocate(roundDown(stack_min, pageSize),
-                     roundUp(stack_size, pageSize));
+    allocateMem(roundDown(stack_min, pageSize), roundUp(stack_size, pageSize));
 
     // map out initial stack contents
     uint32_t sentry_base = stack_base - sentry_size;