Bug 1155176 - correct return types for atomics. r=luke

js/src/asmjs/AsmJSValidate.cpp

@@ -4970,7 +4970,7 @@ CheckAtomicsLoad(FunctionCompiler& f, ParseNode* call, MDefinition** def, Type*
     PrepareArrayIndex(f, &pointerDef, needsBoundsCheck, mask);
 
     *def = f.atomicLoadHeap(viewType, pointerDef, needsBoundsCheck);
-    *type = Type::Signed;
+    *type = Type::Intish;
     return true;
 }
 
@@ -5004,7 +5004,7 @@ CheckAtomicsStore(FunctionCompiler& f, ParseNode* call, MDefinition** def, Type*
     f.atomicStoreHeap(viewType, pointerDef, rhsDef, needsBoundsCheck);
 
     *def = rhsDef;
-    *type = Type::Signed;
+    *type = rhsType;
     return true;
 }
 
@@ -5036,7 +5036,7 @@ CheckAtomicsBinop(FunctionCompiler& f, ParseNode* call, MDefinition** def, Type*
     PrepareArrayIndex(f, &pointerDef, needsBoundsCheck, mask);
 
     *def = f.atomicBinopHeap(op, viewType, pointerDef, valueArgDef, needsBoundsCheck);
-    *type = Type::Signed;
+    *type = Type::Intish;
     return true;
 }
 
@@ -5095,7 +5095,7 @@ CheckAtomicsCompareExchange(FunctionCompiler& f, ParseNode* call, MDefinition**
 
     *def = f.atomicCompareExchangeHeap(viewType, pointerDef, oldValueArgDef, newValueArgDef,
                                        needsBoundsCheck);
-    *type = Type::Signed;
+    *type = Type::Intish;
     return true;
 }
 

js/src/jit-test/tests/asm.js/testAtomic-effect.js

@@ -30,6 +30,9 @@ function m(stdlib, ffi, heap)
     return {add_sharedEv:add_sharedEv};
 }
 
+if (isAsmJSCompilationAvailable())
+    assertEq(isAsmJSModule(m), true);
+
 var x;
 
 var sab = new SharedArrayBuffer(65536);

js/src/jit-test/tests/asm.js/testAtomics.js

@@ -24,7 +24,7 @@ function loadModule_int32(stdlib, foreign, heap) {
     // Load element 0
     function do_load() {
 	var v = 0;
-	v = atomic_load(i32a, 0);
+	v = atomic_load(i32a, 0)|0;
 	return v|0;
     }
 
@@ -32,14 +32,14 @@ function loadModule_int32(stdlib, foreign, heap) {
     function do_load_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_load(i32a, i>>2);
+	v = atomic_load(i32a, i>>2)|0;
 	return v|0;
     }
 
     // Store 37 in element 0
     function do_store() {
 	var v = 0;
-	v = atomic_store(i32a, 0, 37);
+	v = atomic_store(i32a, 0, 37)|0;
 	return v|0;
     }
 
@@ -47,14 +47,14 @@ function loadModule_int32(stdlib, foreign, heap) {
     function do_store_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_store(i32a, i>>2, 37);
+	v = atomic_store(i32a, i>>2, 37)|0;
 	return v|0;
     }
 
     // Add 37 to element 10
     function do_add() {
 	var v = 0;
-	v = atomic_add(i32a, 10, 37);
+	v = atomic_add(i32a, 10, 37)|0;
 	return v|0;
     }
 
@@ -62,14 +62,14 @@ function loadModule_int32(stdlib, foreign, heap) {
     function do_add_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_add(i32a, i>>2, 37);
+	v = atomic_add(i32a, i>>2, 37)|0;
 	return v|0;
     }
 
     // Subtract 148 from element 20
     function do_sub() {
 	var v = 0;
-	v = atomic_sub(i32a, 20, 148);
+	v = atomic_sub(i32a, 20, 148)|0;
 	return v|0;
     }
 
@@ -77,14 +77,14 @@ function loadModule_int32(stdlib, foreign, heap) {
     function do_sub_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_sub(i32a, i>>2, 148);
+	v = atomic_sub(i32a, i>>2, 148)|0;
 	return v|0;
     }
 
     // AND 0x33333333 into element 30
     function do_and() {
 	var v = 0;
-	v = atomic_and(i32a, 30, 0x33333333);
+	v = atomic_and(i32a, 30, 0x33333333)|0;
 	return v|0;
     }
 
@@ -92,14 +92,14 @@ function loadModule_int32(stdlib, foreign, heap) {
     function do_and_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_and(i32a, i>>2, 0x33333333);
+	v = atomic_and(i32a, i>>2, 0x33333333)|0;
 	return v|0;
     }
 
     // OR 0x33333333 into element 40
     function do_or() {
 	var v = 0;
-	v = atomic_or(i32a, 40, 0x33333333);
+	v = atomic_or(i32a, 40, 0x33333333)|0;
 	return v|0;
     }
 
@@ -107,14 +107,14 @@ function loadModule_int32(stdlib, foreign, heap) {
     function do_or_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_or(i32a, i>>2, 0x33333333);
+	v = atomic_or(i32a, i>>2, 0x33333333)|0;
 	return v|0;
     }
 
     // XOR 0x33333333 into element 50
     function do_xor() {
 	var v = 0;
-	v = atomic_xor(i32a, 50, 0x33333333);
+	v = atomic_xor(i32a, 50, 0x33333333)|0;
 	return v|0;
     }
 
@@ -122,21 +122,21 @@ function loadModule_int32(stdlib, foreign, heap) {
     function do_xor_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_xor(i32a, i>>2, 0x33333333);
+	v = atomic_xor(i32a, i>>2, 0x33333333)|0;
 	return v|0;
     }
 
     // CAS element 100: 0 -> -1
     function do_cas1() {
 	var v = 0;
-	v = atomic_cmpxchg(i32a, 100, 0, -1);
+	v = atomic_cmpxchg(i32a, 100, 0, -1)|0;
 	return v|0;
     }
 
     // CAS element 100: -1 -> 0x5A5A5A5A
     function do_cas2() {
 	var v = 0;
-	v = atomic_cmpxchg(i32a, 100, -1, 0x5A5A5A5A);
+	v = atomic_cmpxchg(i32a, 100, -1, 0x5A5A5A5A)|0;
 	return v|0;
     }
 
@@ -144,7 +144,7 @@ function loadModule_int32(stdlib, foreign, heap) {
     function do_cas1_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_cmpxchg(i32a, i>>2, 0, -1);
+	v = atomic_cmpxchg(i32a, i>>2, 0, -1)|0;
 	return v|0;
     }
 
@@ -152,7 +152,7 @@ function loadModule_int32(stdlib, foreign, heap) {
     function do_cas2_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_cmpxchg(i32a, i>>2, -1, 0x5A5A5A5A);
+	v = atomic_cmpxchg(i32a, i>>2, -1, 0x5A5A5A5A)|0;
 	return v|0;
     }
 
@@ -177,6 +177,9 @@ function loadModule_int32(stdlib, foreign, heap) {
 	     cas2_i: do_cas2_i };
 }
 
+if (isAsmJSCompilationAvailable())
+    assertEq(isAsmJSModule(loadModule_int32), true);
+
 function loadModule_int8(stdlib, foreign, heap) {
     "use asm";
 
@@ -194,7 +197,7 @@ function loadModule_int8(stdlib, foreign, heap) {
     // Load element 0
     function do_load() {
 	var v = 0;
-	v = atomic_load(i8a, 0);
+	v = atomic_load(i8a, 0)|0;
 	return v|0;
     }
 
@@ -202,14 +205,14 @@ function loadModule_int8(stdlib, foreign, heap) {
     function do_load_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_load(i8a, i);
+	v = atomic_load(i8a, i)|0;
 	return v|0;
     }
 
     // Store 37 in element 0
     function do_store() {
 	var v = 0;
-	v = atomic_store(i8a, 0, 37);
+	v = atomic_store(i8a, 0, 37)|0;
 	return v|0;
     }
 
@@ -217,14 +220,14 @@ function loadModule_int8(stdlib, foreign, heap) {
     function do_store_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_store(i8a, i, 37);
+	v = atomic_store(i8a, i, 37)|0;
 	return v|0;
     }
 
     // Add 37 to element 10
     function do_add() {
 	var v = 0;
-	v = atomic_add(i8a, 10, 37);
+	v = atomic_add(i8a, 10, 37)|0;
 	return v|0;
     }
 
@@ -232,14 +235,14 @@ function loadModule_int8(stdlib, foreign, heap) {
     function do_add_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_add(i8a, i, 37);
+	v = atomic_add(i8a, i, 37)|0;
 	return v|0;
     }
 
     // Subtract 108 from element 20
     function do_sub() {
 	var v = 0;
-	v = atomic_sub(i8a, 20, 108);
+	v = atomic_sub(i8a, 20, 108)|0;
 	return v|0;
     }
 
@@ -247,14 +250,14 @@ function loadModule_int8(stdlib, foreign, heap) {
     function do_sub_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_sub(i8a, i, 108);
+	v = atomic_sub(i8a, i, 108)|0;
 	return v|0;
     }
 
     // AND 0x33 into element 30
     function do_and() {
 	var v = 0;
-	v = atomic_and(i8a, 30, 0x33);
+	v = atomic_and(i8a, 30, 0x33)|0;
 	return v|0;
     }
 
@@ -262,14 +265,14 @@ function loadModule_int8(stdlib, foreign, heap) {
     function do_and_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_and(i8a, i, 0x33);
+	v = atomic_and(i8a, i, 0x33)|0;
 	return v|0;
     }
 
     // OR 0x33 into element 40
     function do_or() {
 	var v = 0;
-	v = atomic_or(i8a, 40, 0x33);
+	v = atomic_or(i8a, 40, 0x33)|0;
 	return v|0;
     }
 
@@ -277,14 +280,14 @@ function loadModule_int8(stdlib, foreign, heap) {
     function do_or_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_or(i8a, i, 0x33);
+	v = atomic_or(i8a, i, 0x33)|0;
 	return v|0;
     }
 
     // XOR 0x33 into element 50
     function do_xor() {
 	var v = 0;
-	v = atomic_xor(i8a, 50, 0x33);
+	v = atomic_xor(i8a, 50, 0x33)|0;
 	return v|0;
     }
 
@@ -292,21 +295,21 @@ function loadModule_int8(stdlib, foreign, heap) {
     function do_xor_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_xor(i8a, i, 0x33);
+	v = atomic_xor(i8a, i, 0x33)|0;
 	return v|0;
     }
 
     // CAS element 100: 0 -> -1
     function do_cas1() {
 	var v = 0;
-	v = atomic_cmpxchg(i8a, 100, 0, -1);
+	v = atomic_cmpxchg(i8a, 100, 0, -1)|0;
 	return v|0;
     }
 
     // CAS element 100: -1 -> 0x5A
     function do_cas2() {
 	var v = 0;
-	v = atomic_cmpxchg(i8a, 100, -1, 0x5A);
+	v = atomic_cmpxchg(i8a, 100, -1, 0x5A)|0;
 	return v|0;
     }
 
@@ -314,7 +317,7 @@ function loadModule_int8(stdlib, foreign, heap) {
     function do_cas1_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_cmpxchg(i8a, i, 0, -1);
+	v = atomic_cmpxchg(i8a, i, 0, -1)|0;
 	return v|0;
     }
 
@@ -322,7 +325,7 @@ function loadModule_int8(stdlib, foreign, heap) {
     function do_cas2_i(i) {
 	i = i|0;
 	var v = 0;
-	v = atomic_cmpxchg(i8a, i, -1, 0x5A);
+	v = atomic_cmpxchg(i8a, i, -1, 0x5A)|0;
 	return v|0;
     }
 
@@ -346,6 +349,9 @@ function loadModule_int8(stdlib, foreign, heap) {
 	     cas2_i: do_cas2_i };
 }
 
+if (isAsmJSCompilationAvailable())
+    assertEq(isAsmJSModule(loadModule_int8), true);
+
 // TODO: halfword arrays
 // TODO: signed vs unsigned; negative results
 

js/src/jit-test/tests/asm.js/testBug1155176.js

@@ -0,0 +1,49 @@
+if (!this.SharedArrayBuffer || !isAsmJSCompilationAvailable())
+    quit(0);
+
+// The way this is constructed, either the first module does not
+// verify as asm.js (if the >>>0 is left off, which was legal prior to
+// bug 1155176), or the results of the two modules have to be equal.
+
+function m(stdlib, ffi, heap) {
+    "use asm";
+
+    var view = new stdlib.SharedUint32Array(heap);
+    var cas = stdlib.Atomics.compareExchange;
+    var hi = ffi.hi;
+
+    function run() {
+	hi(+(cas(view, 37, 0, 0)>>>0));
+    }
+
+    return run;
+}
+
+assertEq(isAsmJSModule(m), true);
+
+function nonm(stdlib, ffi, heap) {
+
+    var view = new stdlib.SharedUint32Array(heap);
+    var cas = stdlib.Atomics.compareExchange;
+    var hi = ffi.hi;
+
+    function run() {
+	hi(+cas(view, 37, 0, 0));
+    }
+
+    return run;
+}
+
+var sab = new SharedArrayBuffer(65536);
+var ua = new SharedUint32Array(sab);
+var results = [];
+var mrun = m(this, {hi: function (x) { results.push(x) }}, sab);
+var nonmrun = nonm(this, {hi: function (x) { results.push(x) }}, sab);
+
+ua[37] = 0x80000001;
+
+mrun();
+nonmrun();
+
+assertEq(results[0], ua[37]);
+assertEq(results[0], results[1]);

js/src/jit-test/tests/asm.js/testBug1164391.js

@@ -14,7 +14,10 @@ function m(stdlib, ffi, heap) {
     }
     return {add_sharedEv:add_sharedEv};
 }
+
+if (isAsmJSCompilationAvailable())
+    assertEq(isAsmJSModule(m), true);
+
 var sab = new SharedArrayBuffer(65536);
 var {add_sharedEv} = m(this, {}, sab);
 add_sharedEv(sab.byteLength);
-