Merge

make/solaris/makefiles/amd64.make

@@ -33,14 +33,8 @@ Obj_Files += solaris_x86_64.o
 #
 ifeq ("${Platform_compiler}", "sparcWorks")
 
-# Temporary until C++ compiler is fixed
-
-# _lwp_create_interpose must have a frame
-OPT_CFLAGS/os_solaris_x86_64.o = -xO1
-
 # Temporary until SS10 C++ compiler is fixed
 OPT_CFLAGS/generateOptoStub.o = -xO2
-OPT_CFLAGS/thread.o = -xO2
 
 else
 

make/solaris/makefiles/fastdebug.make

@@ -36,15 +36,15 @@ OPT_CFLAGS/BYFILE = $(OPT_CFLAGS/$@)$(OPT_CFLAGS/DEFAULT$(OPT_CFLAGS/$@))
 ifeq ("${Platform_compiler}", "sparcWorks")
 OPT_CFLAGS/SLOWER = -xO2
 
-# Problem with SS12 compiler, dtrace doesn't like the .o files  (bug 6693876)
 ifeq ($(COMPILER_REV_NUMERIC), 509)
-  # To avoid jvm98 crash
-  OPT_CFLAGS/instanceKlass.o = $(OPT_CFLAGS/SLOWER)
-  # Not clear this workaround could be skipped in some cases.
-  OPT_CFLAGS/vmGCOperations.o = $(OPT_CFLAGS/SLOWER)
-  OPT_CFLAGS/java.o = $(OPT_CFLAGS/SLOWER)
-  OPT_CFLAGS/jni.o = $(OPT_CFLAGS/SLOWER)
-endif
+# To avoid jvm98 crash
+OPT_CFLAGS/instanceKlass.o = $(OPT_CFLAGS/SLOWER)
+endif # COMPILER_NUMERIC_REV == 509
+
+ifeq ($(shell expr $(COMPILER_REV_NUMERIC) \>= 509), 1)
+# dtrace cannot handle tail call optimization (6672627, 6693876)
+OPT_CFLAGS/jni.o = $(OPT_CFLAGS/DEFAULT) $(OPT_CCFLAGS/NO_TAIL_CALL_OPT)
+endif # COMPILER_NUMERIC_REV >= 509
 
 ifeq ($(COMPILER_REV_NUMERIC), 505)
 # CC 5.5 has bug 4908364 with -xO4  (Fixed in 5.6)

make/solaris/makefiles/i486.make

@@ -32,25 +32,6 @@ Obj_Files += solaris_x86_32.o
 #
 # Special case flags for compilers and compiler versions on i486.
 #
-ifeq ("${Platform_compiler}", "sparcWorks")
-
-# _lwp_create_interpose must have a frame
-OPT_CFLAGS/os_solaris_x86.o = -xO1
-else
-
-ifeq ("${Platform_compiler}", "gcc")
-# gcc
-# _lwp_create_interpose must have a frame
-OPT_CFLAGS/os_solaris_x86.o = -fno-omit-frame-pointer
-#
-else
-# error
-_JUNK2_ := $(shell echo >&2 \
-       "*** ERROR: this compiler is not yet supported by this code base!")
-	@exit 1
-endif
-endif
-
 ifeq ("${Platform_compiler}", "sparcWorks")
 # ILD is gone as of SS11 (5.8), not supported in SS10 (5.7)
 ifeq ($(shell expr $(COMPILER_REV_NUMERIC) \< 507), 1)

make/solaris/makefiles/launcher.make

@@ -80,15 +80,12 @@ launcher.c:
 	} > $@
 
 $(LAUNCHER): $(LAUNCHER.o) $(LIBJVM) $(LAUNCHER_MAPFILE)
+ifeq ($(filter -sbfast -xsbfast, $(CFLAGS_BROWSE)),)
+	@echo Linking launcher...
+	$(QUIETLY) $(LINK_LAUNCHER/PRE_HOOK)
 	$(QUIETLY) \
-	case "$(CFLAGS_BROWSE)" in \
-	-sbfast|-xsbfast) \
-	    ;; \
-	*) \
-	    echo Linking launcher...; \
-	    $(LINK_LAUNCHER/PRE_HOOK) \
-	    $(LINK_LAUNCHER) $(LFLAGS_LAUNCHER) -o $@ $(LAUNCHER.o) $(LIBS_LAUNCHER); \
-	    $(LINK_LAUNCHER/POST_HOOK) \
-	    [ -f $(LAUNCHER_G) ] || { ln -s $@ $(LAUNCHER_G); }; \
-	    ;; \
-	esac
+	$(LINK_LAUNCHER) $(LFLAGS_LAUNCHER) -o $@ $(LAUNCHER.o) $(LIBS_LAUNCHER)
+	$(QUIETLY) $(LINK_LAUNCHER/POST_HOOK)
+	[ -f $(LAUNCHER_G) ] || ln -s $@ $(LAUNCHER_G)
+endif # filter -sbfast -xsbfast
+

make/solaris/makefiles/optimized.make

@@ -32,13 +32,10 @@ OPT_CFLAGS/BYFILE = $(OPT_CFLAGS/$@)$(OPT_CFLAGS/DEFAULT$(OPT_CFLAGS/$@))
 # (OPT_CFLAGS/SLOWER is also available, to alter compilation of buggy files)
 ifeq ("${Platform_compiler}", "sparcWorks")
 
-# Problem with SS12 compiler, dtrace doesn't like the .o files  (bug 6693876)
-ifeq ($(COMPILER_REV_NUMERIC),509)
-  # Not clear this workaround could be skipped in some cases.
-  OPT_CFLAGS/vmGCOperations.o = $(OPT_CFLAGS/SLOWER) -g
-  OPT_CFLAGS/java.o = $(OPT_CFLAGS/SLOWER) -g
-  OPT_CFLAGS/jni.o = $(OPT_CFLAGS/SLOWER) -g
-endif
+ifeq ($(shell expr $(COMPILER_REV_NUMERIC) \>= 509), 1)
+# dtrace cannot handle tail call optimization (6672627, 6693876)
+OPT_CFLAGS/jni.o = $(OPT_CFLAGS/DEFAULT) $(OPT_CCFLAGS/NO_TAIL_CALL_OPT)
+endif # COMPILER_NUMERIC_REV >= 509
 
 # Workaround SS11 bug 6345274 (all platforms) (Fixed in SS11 patch and SS12)
 ifeq ($(COMPILER_REV_NUMERIC),508)

make/solaris/makefiles/product.make

@@ -40,13 +40,10 @@ endif
 # (OPT_CFLAGS/SLOWER is also available, to alter compilation of buggy files)
 ifeq ("${Platform_compiler}", "sparcWorks")
 
-# Problem with SS12 compiler, dtrace doesn't like the .o files  (bug 6693876)
-ifeq ($(COMPILER_REV_NUMERIC),509)
-  # Not clear this workaround could be skipped in some cases.
-  OPT_CFLAGS/vmGCOperations.o = $(OPT_CFLAGS/SLOWER) -g
-  OPT_CFLAGS/java.o = $(OPT_CFLAGS/SLOWER) -g
-  OPT_CFLAGS/jni.o = $(OPT_CFLAGS/SLOWER) -g
-endif
+ifeq ($(shell expr $(COMPILER_REV_NUMERIC) \>= 509), 1)
+# dtrace cannot handle tail call optimization (6672627, 6693876)
+OPT_CFLAGS/jni.o = $(OPT_CFLAGS/DEFAULT) $(OPT_CCFLAGS/NO_TAIL_CALL_OPT)
+endif # COMPILER_NUMERIC_REV >= 509
 
 # Workaround SS11 bug 6345274 (all platforms) (Fixed in SS11 patch and SS12)
 ifeq ($(COMPILER_REV_NUMERIC),508)

make/solaris/makefiles/sparcWorks.make

@@ -48,27 +48,33 @@ $(shell $(CC) -V 2>&1 | sed -n 's/^.*[ ,\t]C[ ,\t]\([1-9]\.[0-9][0-9]*\).*/\1/p'
 # Pick which compiler is validated
 ifeq ($(JRE_RELEASE_VER),1.6.0)
   # Validated compiler for JDK6 is SS11 (5.8)
-  VALIDATED_COMPILER_REV   := 5.8
-  VALIDATED_C_COMPILER_REV := 5.8
+  VALIDATED_COMPILER_REVS   := 5.8
+  VALIDATED_C_COMPILER_REVS := 5.8
 else
-  # Validated compiler for JDK7 is SS12 (5.9)
-  VALIDATED_COMPILER_REV   := 5.9
-  VALIDATED_C_COMPILER_REV := 5.9
+  # Validated compilers for JDK7 are SS12 (5.9) or SS12 update 1 (5.10)
+  VALIDATED_COMPILER_REVS   := 5.9 5.10
+  VALIDATED_C_COMPILER_REVS := 5.9 5.10
 endif
 
-# Warning messages about not using the above validated version
-ENFORCE_COMPILER_REV${ENFORCE_COMPILER_REV} := ${VALIDATED_COMPILER_REV}
-ifneq (${COMPILER_REV},${ENFORCE_COMPILER_REV})
-dummy_target_to_enforce_compiler_rev:=\
-$(shell echo >&2 WARNING: You are using CC version ${COMPILER_REV} \
-and should be using version ${ENFORCE_COMPILER_REV}. Set ENFORCE_COMPILER_REV=${COMPILER_REV} to avoid this warning.)
+# Warning messages about not using the above validated versions
+ENFORCE_COMPILER_REV${ENFORCE_COMPILER_REV} := $(strip ${VALIDATED_COMPILER_REVS})
+ifeq ($(filter ${ENFORCE_COMPILER_REV},${COMPILER_REV}),)
+PRINTABLE_CC_REVS := $(subst $(shell echo ' '), or ,${ENFORCE_COMPILER_REV})
+dummy_var_to_enforce_compiler_rev := $(shell \
+	echo >&2 WARNING: You are using CC version ${COMPILER_REV} and \
+	should be using version ${PRINTABLE_CC_REVS}.; \
+	echo >&2 Set ENFORCE_COMPILER_REV=${COMPILER_REV} to avoid this \
+	warning.)
 endif
 
-ENFORCE_C_COMPILER_REV${ENFORCE_C_COMPILER_REV} := ${VALIDATED_C_COMPILER_REV}
-ifneq (${C_COMPILER_REV},${ENFORCE_C_COMPILER_REV})
-dummy_target_to_enforce_c_compiler_rev:=\
-$(shell echo >&2 WARNING: You are using cc version ${C_COMPILER_REV} \
-and should be using version ${ENFORCE_C_COMPILER_REV}. Set ENFORCE_C_COMPILER_REV=${C_COMPILER_REV} to avoid this warning.)
+ENFORCE_C_COMPILER_REV${ENFORCE_C_COMPILER_REV} := $(strip ${VALIDATED_C_COMPILER_REVS})
+ifeq ($(filter ${ENFORCE_C_COMPILER_REV},${C_COMPILER_REV}),)
+PRINTABLE_C_REVS := $(subst $(shell echo ' '), or ,${ENFORCE_C_COMPILER_REV})
+dummy_var_to_enforce_c_compiler_rev := $(shell \
+	echo >&2 WARNING: You are using cc version ${C_COMPILER_REV} and \
+	should be using version ${PRINTABLE_C_REVS}.; \
+	echo >&2 Set ENFORCE_C_COMPILER_REV=${C_COMPILER_REV} to avoid this \
+	warning.)
 endif
 
 COMPILER_REV_NUMERIC := $(shell echo $(COMPILER_REV) | awk -F. '{ print $$1 * 100 + $$2 }')
@@ -139,6 +145,13 @@ OPT_CFLAGS/SLOWER=-xO3
 OPT_CFLAGS/O2=-xO2
 OPT_CFLAGS/NOOPT=-xO1
 
+ifeq ($(shell expr $(COMPILER_REV_NUMERIC) \>= 509), 1)
+ifeq ($(Platform_arch), x86)
+OPT_CFLAGS/NO_TAIL_CALL_OPT  = -Wu,-O~yz
+OPT_CCFLAGS/NO_TAIL_CALL_OPT = -Qoption ube -O~yz
+endif # Platform_arch == x86
+endif # COMPILER_REV_NUMERIC >= 509
+
 #################################################
 # Begin current (>=5.6) Forte compiler options #
 #################################################
@@ -181,10 +194,7 @@ endif # sparc
 
 ifeq ("${Platform_arch_model}", "x86_32")
 
-OPT_CFLAGS=-xtarget=pentium $(EXTRA_OPT_CFLAGS)
-
-# UBE (CC 5.5) has bug 4923569 with -xO4
-OPT_CFLAGS+=-xO3
+OPT_CFLAGS=-xtarget=pentium -xO4 $(EXTRA_OPT_CFLAGS)
 
 endif # 32bit x86
 
@@ -461,7 +471,7 @@ FASTDEBUG_CFLAGS = -g0
 # The -g0 setting allows the C++ frontend to inline, which is a big win.
 
 # Special global options for SS12
-ifeq ($(COMPILER_REV_NUMERIC),509)
+ifeq ($(shell expr $(COMPILER_REV_NUMERIC) \>= 509), 1)
   # There appears to be multiple issues with the new Dwarf2 debug format, so
   #   we tell the compiler to use the older 'stabs' debug format all the time.
   #   Note that this needs to be used in optimized compiles too to be 100%.

make/solaris/makefiles/vm.make

@@ -174,19 +174,16 @@ LINK_VM = $(LINK_LIB.CC)
 endif
 # making the library:
 $(LIBJVM): $(LIBJVM.o) $(LIBJVM_MAPFILE) 
-	$(QUIETLY) \
-	case "$(CFLAGS_BROWSE)" in \
-	-sbfast|-xsbfast) \
-	    ;; \
-	*) \
-	    echo Linking vm...; \
-	    $(LINK_LIB.CC/PRE_HOOK) \
-	    $(LINK_VM) $(LFLAGS_VM) -o $@ $(LIBJVM.o) $(LIBS_VM); \
-	    $(LINK_LIB.CC/POST_HOOK) \
-	    rm -f $@.1; ln -s $@ $@.1; \
-	    [ -f $(LIBJVM_G) ] || { ln -s $@ $(LIBJVM_G); ln -s $@.1 $(LIBJVM_G).1; }; \
-	    ;; \
-	esac
+ifeq ($(filter -sbfast -xsbfast, $(CFLAGS_BROWSE)),)
+	@echo Linking vm...
+	$(QUIETLY) $(LINK_LIB.CC/PRE_HOOK)
+	$(QUIETLY) $(LINK_VM) $(LFLAGS_VM) -o $@ $(LIBJVM.o) $(LIBS_VM)
+	$(QUIETLY) $(LINK_LIB.CC/POST_HOOK)
+	$(QUIETLY) rm -f $@.1 && ln -s $@ $@.1
+	$(QUIETLY) [ -f $(LIBJVM_G) ] || ln -s $@ $(LIBJVM_G)
+	$(QUIETLY) [ -f $(LIBJVM_G).1 ] || ln -s $@.1 $(LIBJVM_G).1
+endif # filter -sbfast -xsbfast
+
 
 DEST_JVM = $(JDK_LIBDIR)/$(VM_SUBDIR)/$(LIBJVM)
 

src/cpu/sparc/vm/assembler_sparc.hpp

@@ -2233,7 +2233,7 @@ public:
   AddressLiteral constant_oop_address(jobject obj);                          // find_index
   inline void    set_oop             (jobject obj, Register d);              // uses allocate_oop_address
   inline void    set_oop_constant    (jobject obj, Register d);              // uses constant_oop_address
-  inline void    set_oop             (AddressLiteral& obj_addr, Register d); // same as load_address
+  inline void    set_oop             (const AddressLiteral& obj_addr, Register d); // same as load_address
 
   void set_narrow_oop( jobject obj, Register d );
 

src/cpu/sparc/vm/assembler_sparc.inline.hpp

@@ -712,7 +712,7 @@ inline void MacroAssembler::set_oop_constant(jobject obj, Register d) {
 }
 
 
-inline void MacroAssembler::set_oop(AddressLiteral& obj_addr, Register d) {
+inline void MacroAssembler::set_oop(const AddressLiteral& obj_addr, Register d) {
   assert(obj_addr.rspec().type() == relocInfo::oop_type, "must be an oop reloc");
   set(obj_addr, d);
 }

src/os_cpu/solaris_x86/vm/atomic_solaris_x86.inline.hpp

@@ -47,40 +47,56 @@ inline void Atomic::dec_ptr(volatile void*     dest) { (void)add_ptr(-1, dest);
 // For Sun Studio - implementation is in solaris_x86_[32/64].il.
 // For gcc - implementation is just below.
 
-extern "C" jint _Atomic_add(jint add_value, volatile jint* dest, int mp);
-extern "C" jint _Atomic_xchg(jint exchange_value, volatile jint* dest);
-extern "C" jint _Atomic_cmpxchg(jint exchange_value, volatile jint* dest, jint compare_value, int mp);
-extern "C" jlong _Atomic_cmpxchg_long(jlong exchange_value, volatile jlong* dest, jlong compare_value, int mp);
+// The lock prefix can be omitted for certain instructions on uniprocessors; to
+// facilitate this, os::is_MP() is passed as an additional argument.  64-bit
+// processors are assumed to be multi-threaded and/or multi-core, so the extra
+// argument is unnecessary.
+#ifndef _LP64
+#define IS_MP_DECL() , int is_mp
+#define IS_MP_ARG()  , (int) os::is_MP()
+#else
+#define IS_MP_DECL()
+#define IS_MP_ARG()
+#endif // _LP64
+
+extern "C" {
+  jint _Atomic_add(jint add_value, volatile jint* dest IS_MP_DECL());
+  jint _Atomic_xchg(jint exchange_value, volatile jint* dest);
+  jint _Atomic_cmpxchg(jint exchange_value, volatile jint* dest,
+                       jint compare_value IS_MP_DECL());
+  jlong _Atomic_cmpxchg_long(jlong exchange_value, volatile jlong* dest,
+                             jlong compare_value IS_MP_DECL());
+}
 
 inline jint     Atomic::add    (jint     add_value, volatile jint*     dest) {
-  return _Atomic_add(add_value, dest, (int) os::is_MP());
+  return _Atomic_add(add_value, dest IS_MP_ARG());
+}
+
+inline jint     Atomic::xchg       (jint     exchange_value, volatile jint*     dest) {
+  return _Atomic_xchg(exchange_value, dest);
 }
 
 inline jint     Atomic::cmpxchg    (jint     exchange_value, volatile jint*     dest, jint     compare_value) {
-  return _Atomic_cmpxchg(exchange_value, dest, compare_value, (int) os::is_MP());
+  return _Atomic_cmpxchg(exchange_value, dest, compare_value IS_MP_ARG());
 }
 
 inline jlong    Atomic::cmpxchg    (jlong    exchange_value, volatile jlong*    dest, jlong    compare_value) {
-  return _Atomic_cmpxchg_long(exchange_value, dest, compare_value, (int) os::is_MP());
+  return _Atomic_cmpxchg_long(exchange_value, dest, compare_value IS_MP_ARG());
 }
 
 
 #ifdef AMD64
 inline void Atomic::store    (jlong    store_value, jlong*             dest) { *dest = store_value; }
 inline void Atomic::store    (jlong    store_value, volatile jlong*    dest) { *dest = store_value; }
-extern "C" jlong _Atomic_add_long(jlong add_value, volatile jlong* dest, int mp);
+extern "C" jlong _Atomic_add_long(jlong add_value, volatile jlong* dest);
 extern "C" jlong _Atomic_xchg_long(jlong exchange_value, volatile jlong* dest);
 
 inline intptr_t Atomic::add_ptr(intptr_t add_value, volatile intptr_t* dest) {
-  return (intptr_t)_Atomic_add_long((jlong)add_value, (volatile jlong*)dest, (int) os::is_MP());
+  return (intptr_t)_Atomic_add_long((jlong)add_value, (volatile jlong*)dest);
 }
 
 inline void*    Atomic::add_ptr(intptr_t add_value, volatile void*     dest) {
-  return (void*)_Atomic_add_long((jlong)add_value, (volatile jlong*)dest, (int) os::is_MP());
-}
-
-inline jint     Atomic::xchg    (jint     exchange_value, volatile jint*     dest) {
-  return _Atomic_xchg(exchange_value, dest);
+  return (void*)_Atomic_add_long((jlong)add_value, (volatile jlong*)dest);
 }
 
 inline intptr_t Atomic::xchg_ptr(intptr_t exchange_value, volatile intptr_t* dest) {
@@ -92,11 +108,11 @@ inline void*    Atomic::xchg_ptr(void*    exchange_value, volatile void*     des
 }
 
 inline intptr_t Atomic::cmpxchg_ptr(intptr_t exchange_value, volatile intptr_t* dest, intptr_t compare_value) {
-  return (intptr_t)_Atomic_cmpxchg_long((jlong)exchange_value, (volatile jlong*)dest, (jlong)compare_value, (int) os::is_MP());
+  return (intptr_t)_Atomic_cmpxchg_long((jlong)exchange_value, (volatile jlong*)dest, (jlong)compare_value);
 }
 
 inline void*    Atomic::cmpxchg_ptr(void*    exchange_value, volatile void*     dest, void*    compare_value) {
-  return (void*)_Atomic_cmpxchg_long((jlong)exchange_value, (volatile jlong*)dest, (jlong)compare_value, (int) os::is_MP());
+  return (void*)_Atomic_cmpxchg_long((jlong)exchange_value, (volatile jlong*)dest, (jlong)compare_value);
 }
 
 inline jlong Atomic::load(volatile jlong* src) { return *src; }
@@ -111,13 +127,6 @@ inline void*    Atomic::add_ptr(intptr_t add_value, volatile void*     dest) {
   return (void*)add((jint)add_value, (volatile jint*)dest);
 }
 
-inline jint     Atomic::xchg    (jint     exchange_value, volatile jint*     dest) {
-  // We noticed a CC5.5 bug (4894807), so keep calling the stub just to be safe.
-  // Will use the inline template version after 4894807 is fixed.
-  // return _Atomic_xchg(exchange_value, dest);
-  return (*os::atomic_xchg_func)(exchange_value, dest);
-}
-
 inline intptr_t Atomic::xchg_ptr(intptr_t exchange_value, volatile intptr_t* dest) {
   return (intptr_t)xchg((jint)exchange_value, (volatile jint*)dest);
 }
@@ -179,9 +188,6 @@ extern "C" {
 #endif // AMD64
 
   inline jint _Atomic_xchg(jint exchange_value, volatile jint* dest) {
-
-    // 32bit version originally did nothing!!
-
     __asm__ __volatile__ ("xchgl (%2),%0"
                           : "=r" (exchange_value)
                         : "0" (exchange_value), "r" (dest)

src/os_cpu/solaris_x86/vm/solaris_x86_32.il

@@ -50,10 +50,12 @@
       movl     4(%esp), %edx   // dest
       movl     %eax, %ecx
       cmpl     $0, 8(%esp)     // MP test
-      je       1f
-      lock
-1:    xaddl    %eax, (%edx)
-      addl     %ecx, %eax
+      jne      1f
+      xaddl    %eax, (%edx)
+      jmp      2f
+1:    lock
+      xaddl    %eax, (%edx)
+2:    addl     %ecx, %eax
       .end
 
   // Support for jint Atomic::xchg(jint exchange_value, volatile jint* dest).
@@ -72,9 +74,12 @@
       movl     0(%esp), %ecx   // exchange_value
       movl     4(%esp), %edx   // dest
       cmp      $0, 12(%esp)    // MP test
-      je       1f
-      lock
-1:    cmpxchgl %ecx, (%edx)
+      jne      1f
+      cmpxchgl %ecx, (%edx)
+      jmp      2f
+1:    lock
+      cmpxchgl %ecx, (%edx)
+2:
       .end
 
   // Support for jlong Atomic::cmpxchg(jlong exchange_value,
@@ -90,10 +95,12 @@
       movl     8(%esp), %ebx   // exchange_value (low)
       movl     12(%esp), %ecx  // exchange_high (high)
       cmp      $0, 28(%esp)    // MP test
-      je       1f
-      lock
-1:    cmpxchg8b (%edi)
-      popl     %edi
+      jne      1f
+      cmpxchg8b (%edi)
+      jmp      2f
+1:    lock
+      cmpxchg8b (%edi)
+2:    popl     %edi
       popl     %ebx
       .end
 

src/os_cpu/solaris_x86/vm/solaris_x86_64.il

@@ -37,24 +37,18 @@
       .end
 
   // Support for jint Atomic::add(jint add_value, volatile jint* dest)
-  // An additional bool (os::is_MP()) is passed as the last argument.
-      .inline _Atomic_add,3
+      .inline _Atomic_add,2
       movl     %edi, %eax      // save add_value for return
-      testl    %edx, %edx      // MP test
-      je       1f
       lock
-1:    xaddl    %edi, (%rsi)
+      xaddl    %edi, (%rsi)
       addl     %edi, %eax
       .end
 
   // Support for jlong Atomic::add(jlong add_value, volatile jlong* dest)
-  // An additional bool (os::is_MP()) is passed as the last argument.
-      .inline _Atomic_add_long,3
+      .inline _Atomic_add_long,2
       movq     %rdi, %rax      // save add_value for return
-      testq    %rdx, %rdx      // MP test
-      je       1f
       lock
-1:    xaddq    %rdi, (%rsi)
+      xaddq    %rdi, (%rsi)
       addq     %rdi, %rax
       .end
 
@@ -73,25 +67,19 @@
   // Support for jint Atomic::cmpxchg(jint exchange_value, 
   //                                  volatile jint *dest, 
   //                                  jint compare_value)
-  // An additional bool (os::is_MP()) is passed as the last argument.
-      .inline _Atomic_cmpxchg,4
+      .inline _Atomic_cmpxchg,3
       movl     %edx, %eax      // compare_value
-      testl    %ecx, %ecx      // MP test
-      je       1f
       lock
-1:    cmpxchgl %edi, (%rsi)
+      cmpxchgl %edi, (%rsi)
       .end
 
   // Support for jlong Atomic::cmpxchg(jlong exchange_value,
   //                                   volatile jlong* dest,
   //                                   jlong compare_value)
-  // An additional bool (os::is_MP()) is passed as the last argument.
-      .inline _Atomic_cmpxchg_long,6
+      .inline _Atomic_cmpxchg_long,3
       movq     %rdx, %rax      // compare_value
-      testq    %rcx, %rcx      // MP test
-      je       1f
       lock
-1:    cmpxchgq %rdi, (%rsi)
+      cmpxchgq %rdi, (%rsi)
       .end
 
   // Support for OrderAccess::acquire()

src/share/vm/gc_implementation/concurrentMarkSweep/promotionInfo.cpp

 /*
- * Copyright 2001-2009 Sun Microsystems, Inc.  All Rights Reserved.
+ * Copyright 2010 Sun Microsystems, Inc.  All Rights Reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
  * This code is free software; you can redistribute it and/or modify it

src/share/vm/gc_implementation/concurrentMarkSweep/promotionInfo.hpp

 /*
- * Copyright 2001-2009 Sun Microsystems, Inc.  All Rights Reserved.
+ * Copyright 2010 Sun Microsystems, Inc.  All Rights Reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
  * This code is free software; you can redistribute it and/or modify it
@@ -32,30 +32,75 @@ class PromotedObject VALUE_OBJ_CLASS_SPEC {
     displaced_mark = nth_bit(2),        // i.e. 0x4
     next_mask      = ~(right_n_bits(3)) // i.e. ~(0x7)
   };
-  intptr_t _next;
+
+  // Below, we want _narrow_next in the "higher" 32 bit slot,
+  // whose position will depend on endian-ness of the platform.
+  // This is so that there is no interference with the
+  // cms_free_bit occupying bit position 7 (lsb == 0)
+  // when we are using compressed oops; see FreeChunk::isFree().
+  // We cannot move the cms_free_bit down because currently
+  // biased locking code assumes that age bits are contiguous
+  // with the lock bits. Even if that assumption were relaxed,
+  // the least position we could move this bit to would be
+  // to bit position 3, which would require 16 byte alignment.
+  typedef struct {
+#ifdef VM_LITTLE_ENDIAN
+    LP64_ONLY(narrowOop _pad;)
+              narrowOop _narrow_next;
+#else
+              narrowOop _narrow_next;
+    LP64_ONLY(narrowOop _pad;)
+#endif
+  } Data;
+
+  union {
+    intptr_t _next;
+    Data     _data;
+  };
  public:
   inline PromotedObject* next() const {
-    return (PromotedObject*)(_next & next_mask);
+    assert(!((FreeChunk*)this)->isFree(), "Error");
+    PromotedObject* res;
+    if (UseCompressedOops) {
+      // The next pointer is a compressed oop stored in the top 32 bits
+      res = (PromotedObject*)oopDesc::decode_heap_oop(_data._narrow_next);
+    } else {
+      res = (PromotedObject*)(_next & next_mask);
+    }
+    assert(oop(res)->is_oop_or_null(true /* ignore mark word */), "Not an oop?");
+    return res;
   }
   inline void setNext(PromotedObject* x) {
-    assert(((intptr_t)x & ~next_mask) == 0,
-           "Conflict in bit usage, "
-           " or insufficient alignment of objects");
-    _next |= (intptr_t)x;
+    assert(((intptr_t)x & ~next_mask) == 0, "Conflict in bit usage, "
+           "or insufficient alignment of objects");
+    if (UseCompressedOops) {
+      assert(_data._narrow_next == 0, "Overwrite?");
+      _data._narrow_next = oopDesc::encode_heap_oop(oop(x));
+    } else {
+      _next |= (intptr_t)x;
+    }
+    assert(!((FreeChunk*)this)->isFree(), "Error");
   }
   inline void setPromotedMark() {
     _next |= promoted_mask;
+    assert(!((FreeChunk*)this)->isFree(), "Error");
   }
   inline bool hasPromotedMark() const {
+    assert(!((FreeChunk*)this)->isFree(), "Error");
     return (_next & promoted_mask) == promoted_mask;
   }
   inline void setDisplacedMark() {
     _next |= displaced_mark;
+    assert(!((FreeChunk*)this)->isFree(), "Error");
   }
   inline bool hasDisplacedMark() const {
+    assert(!((FreeChunk*)this)->isFree(), "Error");
     return (_next & displaced_mark) != 0;
   }
-  inline void clearNext()        { _next = 0; }
+  inline void clearNext()        {
+    _next = 0;
+    assert(!((FreeChunk*)this)->isFree(), "Error");
+  }
   debug_only(void *next_addr() { return (void *) &_next; })
 };
 

src/share/vm/gc_implementation/g1/concurrentMark.cpp

@@ -766,10 +766,12 @@ void ConcurrentMark::checkpointRootsInitialPre() {
 
   _has_aborted = false;
 
+#ifndef PRODUCT
   if (G1PrintReachableAtInitialMark) {
     print_reachable("at-cycle-start",
                     true /* use_prev_marking */, true /* all */);
   }
+#endif
 
   // Initialise marking structures. This has to be done in a STW phase.
   reset();

src/share/vm/gc_implementation/g1/g1CollectedHeap.cpp

@@ -471,21 +471,23 @@ HeapRegion* G1CollectedHeap::newAllocRegion_work(size_t word_size,
              res->zero_fill_state() == HeapRegion::Allocated)),
            "Alloc Regions must be zero filled (and non-H)");
   }
-  if (res != NULL && res->is_empty()) _free_regions--;
-  assert(res == NULL ||
-         (!res->isHumongous() &&
-          (!zero_filled ||
-           res->zero_fill_state() == HeapRegion::Allocated)),
-         "Non-young alloc Regions must be zero filled (and non-H)");
-
-  if (G1PrintHeapRegions) {
-    if (res != NULL) {
+  if (res != NULL) {
+    if (res->is_empty()) {
+      _free_regions--;
+    }
+    assert(!res->isHumongous() &&
+           (!zero_filled || res->zero_fill_state() == HeapRegion::Allocated),
+           err_msg("Non-young alloc Regions must be zero filled (and non-H):"
+                   " res->isHumongous()=%d, zero_filled=%d, res->zero_fill_state()=%d",
+                   res->isHumongous(), zero_filled, res->zero_fill_state()));
+    assert(!res->is_on_unclean_list(),
+           "Alloc Regions must not be on the unclean list");
+    if (G1PrintHeapRegions) {
       gclog_or_tty->print_cr("new alloc region %d:["PTR_FORMAT", "PTR_FORMAT"], "
                              "top "PTR_FORMAT,
                              res->hrs_index(), res->bottom(), res->end(), res->top());
     }
   }
-
   return res;
 }
 
@@ -2338,10 +2340,12 @@ void G1CollectedHeap::verify(bool allow_dirty,
       gclog_or_tty->print_cr("Heap:");
       print_on(gclog_or_tty, true /* extended */);
       gclog_or_tty->print_cr("");
+#ifndef PRODUCT
       if (VerifyDuringGC && G1VerifyDuringGCPrintReachable) {
         concurrent_mark()->print_reachable("at-verification-failure",
                                            use_prev_marking, false /* all */);
       }
+#endif
       gclog_or_tty->flush();
     }
     guarantee(!failures, "there should not have been any failures");
@@ -4600,6 +4604,15 @@ void G1CollectedHeap::wait_for_cleanup_complete_locked() {
 void
 G1CollectedHeap::put_region_on_unclean_list_locked(HeapRegion* r) {
   assert(ZF_mon->owned_by_self(), "precondition.");
+#ifdef ASSERT
+  if (r->is_gc_alloc_region()) {
+    ResourceMark rm;
+    stringStream region_str;
+    print_on(&region_str);
+    assert(!r->is_gc_alloc_region(), err_msg("Unexpected GC allocation region: %s",
+                                             region_str.as_string()));
+  }
+#endif
   _unclean_region_list.insert_before_head(r);
 }
 

src/share/vm/gc_implementation/g1/heapRegion.cpp

@@ -554,11 +554,19 @@ HeapWord* HeapRegion::allocate(size_t size) {
 #endif
 
 void HeapRegion::set_zero_fill_state_work(ZeroFillState zfs) {
-  assert(top() == bottom() || zfs == Allocated,
-         "Region must be empty, or we must be setting it to allocated.");
   assert(ZF_mon->owned_by_self() ||
          Universe::heap()->is_gc_active(),
          "Must hold the lock or be a full GC to modify.");
+#ifdef ASSERT
+  if (top() != bottom() && zfs != Allocated) {
+    ResourceMark rm;
+    stringStream region_str;
+    print_on(&region_str);
+    assert(top() == bottom() || zfs == Allocated,
+           err_msg("Region must be empty, or we must be setting it to allocated. "
+                   "_zfs=%d, zfs=%d, region: %s", _zfs, zfs, region_str.as_string()));
+  }
+#endif
   _zfs = zfs;
 }
 

src/share/vm/gc_implementation/shared/spaceDecorator.hpp

@@ -109,7 +109,7 @@ class SpaceMangler: public CHeapObj {
   // is fully constructed.  Also is used when a generation is expanded
   // and possibly before the spaces have been reshaped to to the new
   // size of the generation.
-  static void mangle_region(MemRegion mr);
+  static void mangle_region(MemRegion mr) PRODUCT_RETURN;
 };
 
 class ContiguousSpace;

src/share/vm/gc_implementation/shared/vmGCOperations.cpp

@@ -32,10 +32,12 @@ HS_DTRACE_PROBE_DECL(hotspot, gc__end);
 // for the other file anymore.   The dtrace probes have to remain stable.
 void VM_GC_Operation::notify_gc_begin(bool full) {
   HS_DTRACE_PROBE1(hotspot, gc__begin, full);
+  HS_DTRACE_WORKAROUND_TAIL_CALL_BUG();
 }
 
 void VM_GC_Operation::notify_gc_end() {
   HS_DTRACE_PROBE(hotspot, gc__end);
+  HS_DTRACE_WORKAROUND_TAIL_CALL_BUG();
 }
 
 void VM_GC_Operation::acquire_pending_list_lock() {

src/share/vm/gc_interface/collectedHeap.cpp

@@ -65,7 +65,7 @@ CollectedHeap::CollectedHeap()
 void CollectedHeap::pre_initialize() {
   // Used for ReduceInitialCardMarks (when COMPILER2 is used);
   // otherwise remains unused.
-#ifdef COMPLER2
+#ifdef COMPILER2
   _defer_initial_card_mark =    ReduceInitialCardMarks && can_elide_tlab_store_barriers()
                              && (DeferInitialCardMark || card_mark_must_follow_store());
 #else
@@ -309,7 +309,7 @@ void CollectedHeap::fill_with_objects(HeapWord* start, size_t words, bool zap)
   DEBUG_ONLY(fill_args_check(start, words);)
   HandleMark hm;  // Free handles before leaving.
 
-#ifdef LP64
+#ifdef _LP64
   // A single array can fill ~8G, so multiple objects are needed only in 64-bit.
   // First fill with arrays, ensuring that any remaining space is big enough to
   // fill.  The remainder is filled with a single object.

src/share/vm/oops/markOop.hpp

@@ -27,12 +27,26 @@
 // Note that the mark is not a real oop but just a word.
 // It is placed in the oop hierarchy for historical reasons.
 //
-// Bit-format of an object header (most significant first):
+// Bit-format of an object header (most significant first, big endian layout below):
 //
-//  32 bits: unused:0  hash:25 age:4 biased_lock:1 lock:2
-//  64 bits: unused:24 hash:31 cms:2 age:4 biased_lock:1 lock:2
-//           unused:20 size:35 cms:2 age:4 biased_lock:1 lock:2 (if cms
-//                                                               free chunk)
+//  32 bits:
+//  --------
+//             hash:25 ------------>| age:4    biased_lock:1 lock:2 (normal object)
+//             JavaThread*:23 epoch:2 age:4    biased_lock:1 lock:2 (biased object)
+//             size:32 ------------------------------------------>| (CMS free block)
+//             PromotedObject*:29 ---------->| promo_bits:3 ----->| (CMS promoted object)
+//
+//  64 bits:
+//  --------
+//  unused:25 hash:31 -->| unused:1   age:4    biased_lock:1 lock:2 (normal object)
+//  JavaThread*:54 epoch:2 unused:1   age:4    biased_lock:1 lock:2 (biased object)
+//  PromotedObject*:61 --------------------->| promo_bits:3 ----->| (CMS promoted object)
+//  size:64 ----------------------------------------------------->| (CMS free block)
+//
+//  unused:25 hash:31 -->| cms_free:1 age:4    biased_lock:1 lock:2 (COOPs && normal object)
+//  JavaThread*:54 epoch:2 cms_free:1 age:4    biased_lock:1 lock:2 (COOPs && biased object)
+//  narrowOop:32 unused:24 cms_free:1 unused:4 promo_bits:3 ----->| (COOPs && CMS promoted object)
+//  unused:21 size:35 -->| cms_free:1 unused:7 ------------------>| (COOPs && CMS free block)
 //
 //  - hash contains the identity hash value: largest value is
 //    31 bits, see os::random().  Also, 64-bit vm's require
@@ -61,8 +75,9 @@
 //    significant fraction of the eden semispaces and were not
 //    promoted promptly, causing an increase in the amount of copying
 //    performed. The runtime system aligns all JavaThread* pointers to
-//    a very large value (currently 128 bytes) to make room for the
-//    age bits when biased locking is enabled.
+//    a very large value (currently 128 bytes (32bVM) or 256 bytes (64bVM))
+//    to make room for the age bits & the epoch bits (used in support of
+//    biased locking), and for the CMS "freeness" bit in the 64bVM (+COOPs).
 //
 //    [JavaThread* | epoch | age | 1 | 01]       lock is biased toward given thread
 //    [0           | epoch | age | 1 | 01]       lock is anonymously biased

src/share/vm/runtime/java.cpp

@@ -470,6 +470,7 @@ void vm_exit(int code) {
 void notify_vm_shutdown() {
   // For now, just a dtrace probe.
   HS_DTRACE_PROBE(hotspot, vm__shutdown);
+  HS_DTRACE_WORKAROUND_TAIL_CALL_BUG();
 }
 
 void vm_direct_exit(int code) {

src/share/vm/runtime/vframe.cpp

@@ -101,8 +101,8 @@ GrowableArray<MonitorInfo*>* javaVFrame::locked_monitors() {
   bool found_first_monitor = false;
   ObjectMonitor *pending_monitor = thread()->current_pending_monitor();
   ObjectMonitor *waiting_monitor = thread()->current_waiting_monitor();
-  oop pending_obj = (pending_monitor != NULL ? (oop) pending_monitor->object() : NULL);
-  oop waiting_obj = (waiting_monitor != NULL ? (oop) waiting_monitor->object() : NULL);
+  oop pending_obj = (pending_monitor != NULL ? (oop) pending_monitor->object() : (oop) NULL);
+  oop waiting_obj = (waiting_monitor != NULL ? (oop) waiting_monitor->object() : (oop) NULL);
 
   for (int index = (mons->length()-1); index >= 0; index--) {
     MonitorInfo* monitor = mons->at(index);

src/share/vm/runtime/vm_version.cpp

@@ -190,6 +190,8 @@ const char* Abstract_VM_Version::internal_vm_info_string() {
         #define HOTSPOT_BUILD_COMPILER "Workshop 5.8"
       #elif __SUNPRO_CC == 0x590
         #define HOTSPOT_BUILD_COMPILER "Workshop 5.9"
+      #elif __SUNPRO_CC == 0x5100
+        #define HOTSPOT_BUILD_COMPILER "Sun Studio 12u1"
       #else
         #define HOTSPOT_BUILD_COMPILER "unknown Workshop:" XSTR(__SUNPRO_CC)
       #endif

src/share/vm/utilities/dtrace.hpp

@@ -29,6 +29,10 @@
 #define DTRACE_ONLY(x) x
 #define NOT_DTRACE(x)
 
+// Work around dtrace tail call bug 6672627 until it is fixed in solaris 10.
+#define HS_DTRACE_WORKAROUND_TAIL_CALL_BUG() \
+  do { volatile size_t dtrace_workaround_tail_call_bug = 1; } while (0)
+
 #else // ndef SOLARIS || ndef DTRACE_ENABLED
 
 #define DTRACE_ONLY(x)
@@ -41,6 +45,8 @@
 #define DTRACE_PROBE4(a,b,c,d,e,f) {;}
 #define DTRACE_PROBE5(a,b,c,d,e,f,g) {;}
 
+#define HS_DTRACE_WORKAROUND_TAIL_CALL_BUG()
+
 #endif
 
 #define HS_DTRACE_PROBE_FN(provider,name)\