update

git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@382 c046a42c-6fe2-441c-8c8c-71466251a162
finished simplifying string operations - correct TF flag handling for string operations and ss loading - simplified basic block exit code generation
2003-09-17 22:56:56 +00:00 · 2003-09-17 22:56:30 +00:00 · 2003-09-17 22:53:29 +00:00 · 2003-09-17 22:52:47 +00:00 · 2003-09-17 22:51:45 +00:00 · 2003-09-17 22:51:18 +00:00
61 changed files with 14351 additions and 2370 deletions
--- a/57
+++ b/57
@@ -1,19 +1,66 @@
+version 0.4.4:
+  
+  - full hardware level VGA emulation
+  - graphical display with SDL
+  - added PS/2 mouse and keyboard emulation
+  - popw (%esp) fix
+  - mov to/from segment data width fix
+  - added real mode support
+  - added Bochs BIOS and LGPL'ed VGA BIOS loader in vl
+  - m68k host port (Richard Zidlicky)
+  - partial soft MMU support for memory mapped I/Os
+  - multi-target build
+  - fixed: no error code in hardware interrupts
+  - fixed: pop ss, mov ss, x and sti disable hardware irqs for the next insn
+  - correct single stepping thru string operations
+
+version 0.4.3:
+
+  - x86 exception fix in case of nop instruction.
+  - gcc 3.2.2 bug workaround (RedHat 9 fix)
+  - sparc and Alpha host fixes
+  - many ARM target fixes: 'ls' and 'bash' can be launched.
+
+version 0.4.2:
+
+ - many exception handling fixes (can compile a Linux kernel inside vl)
+ - IDE emulation support
+ - initial GDB stub support
+ - deferred update support for disk images (Rusty Russell)
+ - accept User Mode Linux Copy On Write disk images
+ - SMP kernels can at least be booted
+
+version 0.4.1:
+  
+ - more accurate timer support in vl.
+ - more reliable NE2000 probe in vl.
+ - added 2.5.66 kernel in vl-test.
+ - added VLTMPDIR environment variable in vl.
+
+version 0.4:
+
+ - initial support for ring 0 x86 processor emulation
+ - fixed signal handling for correct dosemu DPMI emulation
+ - fast x86 MMU emulation with mmap()
+ - fixed popl (%esp) case
+ - Linux kernel can be executed by QEMU with the 'vl' command.
+
 version 0.3:

 - initial support for ARM emulation
 - added fnsave, frstor, fnstenv, fldenv FPU instructions
 - added FPU register save in signal emulation
- - ARM port
+ - initial ARM port
 - Sparc and Alpha ports work on the regression test
 - generic ioctl number conversion
 - fixed ioctl type conversion

 version 0.2:

- - PowerPC disassembly and ELF symbols output (Rusty Russel)
- - flock support (Rusty Russel)
- - ugetrlimit support (Rusty Russel)
- - fstat64 fix (Rusty Russel)
+ - PowerPC disassembly and ELF symbols output (Rusty Russell)
+ - flock support (Rusty Russell)
+ - ugetrlimit support (Rusty Russell)
+ - fstat64 fix (Rusty Russell)
 - initial Alpha port (Falk Hueffner)
 - initial IA64 port (Matt Wilson)
 - initial Sparc and Sparc64 port (David S. Miller)
--- a/216
+++ b/216
@@ -1,190 +1,48 @@
-include config.mak
+include config-host.mak

 CFLAGS=-Wall -O2 -g
 LDFLAGS=-g
 LIBS=
-DEFINES=-DHAVE_BYTESWAP_H
-HELPER_CFLAGS=$(CFLAGS)
-
-ifdef CONFIG_STATIC
-LDFLAGS+=-static
-endif
-
-ifeq ($(ARCH),i386)
-CFLAGS+=-fomit-frame-pointer
-OP_CFLAGS=$(CFLAGS) -mpreferred-stack-boundary=2
-ifeq ($(GCC_MAJOR),3)
-OP_CFLAGS+= -falign-functions=0
-else
-OP_CFLAGS+= -malign-functions=0
-endif
-ifdef TARGET_GPROF
-LDFLAGS+=-Wl,-T,i386.ld
-else
-# WARNING: this LDFLAGS is _very_ tricky : qemu is an ELF shared object
-# that the kernel ELF loader considers as an executable. I think this
-# is the simplest way to make it self virtualizable!
-LDFLAGS+=-Wl,-shared
-endif
-endif
-
-ifeq ($(ARCH),ppc)
-OP_CFLAGS=$(CFLAGS)
-LDFLAGS+=-Wl,-T,ppc.ld
-endif
-
-ifeq ($(ARCH),s390)
-OP_CFLAGS=$(CFLAGS)
-LDFLAGS+=-Wl,-T,s390.ld
-endif
-
-ifeq ($(ARCH),sparc)
-CFLAGS+=-m32 -ffixed-g1 -ffixed-g2 -ffixed-g3 -ffixed-g6
-LDFLAGS+=-m32
-OP_CFLAGS=$(CFLAGS) -fno-delayed-branch -ffixed-i0
-HELPER_CFLAGS=$(CFLAGS) -ffixed-i0 -mflat
-LDFLAGS+=-Wl,-T,sparc.ld
-endif
-
-ifeq ($(ARCH),sparc64)
-CFLAGS+=-m64 -ffixed-g1 -ffixed-g2 -ffixed-g3 -ffixed-g6
-LDFLAGS+=-m64
-OP_CFLAGS=$(CFLAGS) -fno-delayed-branch -ffixed-i0
-endif
-
-ifeq ($(ARCH),alpha)
-# -msmall-data is not used because we want two-instruction relocations
-# for the constant constructions
-OP_CFLAGS=-Wall -O2 -g
-# Ensure there's only a single GP
-CFLAGS += -msmall-data
-LDFLAGS+=-Wl,-T,alpha.ld
-endif
-
-ifeq ($(ARCH),ia64)
-OP_CFLAGS=$(CFLAGS)
-endif
-
-ifeq ($(ARCH),arm)
-OP_CFLAGS=$(CFLAGS) -mno-sched-prolog
-LDFLAGS+=-Wl,-T,arm.ld
-endif
-
-ifeq ($(GCC_MAJOR),3)
-# very important to generate a return at the end of every operation
-OP_CFLAGS+=-fno-reorder-blocks -fno-optimize-sibling-calls
-endif
-
-#########################################################
-
 DEFINES+=-D_GNU_SOURCE
-LIBS+=-lm
+TOOLS=vlmkcow

-# profiling code
-ifdef TARGET_GPROF
-LDFLAGS+=-p
-main.o: CFLAGS+=-p
-endif
+all: dyngen $(TOOLS) qemu-doc.html
+	for d in $(TARGET_DIRS); do \
+	make -C $$d $@ || exit 1 ; \
+        done

-OBJS= elfload.o main.o syscall.o mmap.o signal.o path.o
-ifeq ($(TARGET_ARCH), i386)
-OBJS+= vm86.o
-endif
-SRCS:= $(OBJS:.o=.c)
-OBJS+= libqemu.a
+vlmkcow: vlmkcow.o
+	$(HOST_CC) -o $@ $^  $(LIBS)

-# cpu emulator library
-LIBOBJS=thunk.o exec.o translate.o cpu-exec.o 
-
-ifeq ($(TARGET_ARCH), i386)
-LIBOBJS+=translate-i386.o op-i386.o helper-i386.o
-endif
-ifeq ($(TARGET_ARCH), arm)
-LIBOBJS+=translate-arm.o op-arm.o
-endif
-
-# NOTE: the disassembler code is only needed for debugging
-LIBOBJS+=disas.o 
-ifeq ($(findstring i386, $(TARGET_ARCH) $(ARCH)),i386)
-LIBOBJS+=i386-dis.o
-endif
-ifeq ($(findstring alpha, $(TARGET_ARCH) $(ARCH)),alpha)
-LIBOBJS+=alpha-dis.o
-endif
-ifeq ($(findstring ppc, $(TARGET_ARCH) $(ARCH)),ppc)
-LIBOBJS+=ppc-dis.o
-endif
-ifeq ($(findstring sparc, $(TARGET_ARCH) $(ARCH)),sparc)
-LIBOBJS+=sparc-dis.o
-endif
-ifeq ($(findstring arm, $(TARGET_ARCH) $(ARCH)),arm)
-LIBOBJS+=arm-dis.o
-endif
-
-ifeq ($(ARCH),ia64)
-OBJS += ia64-syscall.o
-endif
-
-all: qemu qemu-doc.html
-
-qemu: $(OBJS)
-	$(CC) $(LDFLAGS) -o $@ $^  $(LIBS)
-ifeq ($(ARCH),alpha)
-# Mark as 32 bit binary, i. e. it will be mapped into the low 31 bit of
-# the address space (31 bit so sign extending doesn't matter)
-	echo -ne '\001\000\000\000' | dd of=qemu bs=1 seek=48 count=4 conv=notrunc
-endif
-
-depend: $(SRCS)
-	$(CC) -MM $(CFLAGS) $^ 1>.depend
-
-# libqemu 
-
-libqemu.a: $(LIBOBJS)
-	rm -f $@
-	$(AR) rcs $@ $(LIBOBJS)
-
-dyngen: dyngen.c
-	$(HOST_CC) -O2 -Wall -g $< -o $@
-
-translate-$(TARGET_ARCH).o: translate-$(TARGET_ARCH).c gen-op-$(TARGET_ARCH).h opc-$(TARGET_ARCH).h cpu-$(TARGET_ARCH).h
-
-translate.o: translate.c op-$(TARGET_ARCH).h opc-$(TARGET_ARCH).h cpu-$(TARGET_ARCH).h
-
-op-$(TARGET_ARCH).h: op-$(TARGET_ARCH).o dyngen
-	./dyngen -o $@ $<
-
-opc-$(TARGET_ARCH).h: op-$(TARGET_ARCH).o dyngen
-	./dyngen -c -o $@ $<
-
-gen-op-$(TARGET_ARCH).h: op-$(TARGET_ARCH).o dyngen
-	./dyngen -g -o $@ $<
-
-op-$(TARGET_ARCH).o: op-$(TARGET_ARCH).c
-	$(CC) $(OP_CFLAGS) $(DEFINES) -c -o $@ $<
-
-helper-$(TARGET_ARCH).o: helper-$(TARGET_ARCH).c
-	$(CC) $(HELPER_CFLAGS) $(DEFINES) -c -o $@ $<
-
-op-i386.o: op-i386.c opreg_template.h ops_template.h
-
-op-arm.o: op-arm.c op-arm-template.h
+dyngen: dyngen.o
+	$(HOST_CC) -o $@ $^  $(LIBS)

 %.o: %.c
-	$(CC) $(CFLAGS) $(DEFINES) -c -o $@ $<
+	$(HOST_CC) $(CFLAGS) $(DEFINES) -c -o $@ $<

 clean:
-	$(MAKE) -C tests clean
-	rm -f *.o  *.a *~ qemu dyngen TAGS
+# avoid old build problems by removing potentially incorrect old files
+	rm -f config.mak config.h op-i386.h opc-i386.h gen-op-i386.h op-arm.h opc-arm.h gen-op-arm.h 
+	rm -f *.o *.a $(TOOLS) dyngen TAGS
+	for d in $(TARGET_DIRS); do \
+	make -C $$d $@ || exit 1 ; \
+        done

 distclean: clean
-	rm -f config.mak config.h
+	rm -f config-host.mak config-host.h
+	for d in $(TARGET_DIRS); do \
+	rm -f $$d/config.h $$d/config.mak || exit 1 ; \
+        done

-install: qemu
-	install -m 755 -s qemu $(prefix)/bin
+install: all 
+	mkdir -p $(prefix)/bin
+	install -m 755 -s $(TOOLS) $(prefix)/bin
+	for d in $(TARGET_DIRS); do \
+	make -C $$d $@ || exit 1 ; \
+        done

 # various test targets
-test speed: qemu
+test speed: all
 	make -C tests $@

 TAGS: 
@@ -196,22 +54,24 @@ qemu-doc.html: qemu-doc.texi

 FILES= \
 README README.distrib COPYING COPYING.LIB TODO Changelog VERSION \
-configure \
+configure Makefile Makefile.target \
 dyngen.c dyngen.h dyngen-exec.h ioctls.h syscall_types.h \
-Makefile elf.h elfload.c main.c signal.c qemu.h \
+elf.h elfload.c main.c signal.c qemu.h \
 syscall.c syscall_defs.h vm86.c path.c mmap.c \
-ppc.ld alpha.ld s390.ld sparc.ld arm.ld\
-thunk.c cpu-exec.c translate.c cpu-all.h thunk.h exec.h\
-exec.c cpu-exec.c\
-cpu-i386.h op-i386.c helper-i386.c syscall-i386.h translate-i386.c \
-exec-i386.h ops_template.h op_string.h opreg_template.h \
+i386.ld ppc.ld alpha.ld s390.ld sparc.ld arm.ld m68k.ld \
+vl.c i386-vl.ld vl.h block.c vlmkcow.c vga.c vga_template.h sdl.c \
+thunk.c cpu-exec.c translate.c cpu-all.h cpu-defs.h thunk.h exec.h\
+exec.c cpu-exec.c gdbstub.c bswap.h \
+cpu-i386.h op-i386.c helper-i386.c helper2-i386.c syscall-i386.h translate-i386.c \
+exec-i386.h ops_template.h ops_template_mem.h op_string.h opreg_template.h \
+ops_mem.h softmmu_template.h softmmu_header.h \
 cpu-arm.h syscall-arm.h exec-arm.h op-arm.c translate-arm.c op-arm-template.h \
 dis-asm.h disas.c disas.h alpha-dis.c ppc-dis.c i386-dis.c sparc-dis.c \
 arm-dis.c \
 tests/Makefile \
 tests/test-i386.c tests/test-i386-shift.h tests/test-i386.h \
 tests/test-i386-muldiv.h tests/test-i386-code16.S tests/test-i386-vm86.S \
-tests/hello.c tests/hello \
+tests/hello-i386.c tests/hello-i386 \
 tests/hello-arm.c tests/hello-arm \
 tests/sha1.c \
 tests/testsig.c tests/testclone.c tests/testthread.c \
--- a/Makefile.target
+++ b/Makefile.target
@@ -0,0 +1,206 @@
+include config.mak
+
+VPATH=$(SRC_PATH)
+CFLAGS=-Wall -O2 -g
+LDFLAGS=-g
+LIBS=
+DEFINES=-I.
+HELPER_CFLAGS=$(CFLAGS)
+DYNGEN=../dyngen
+ifndef CONFIG_SOFTMMU
+PROGS=qemu
+endif
+
+ifdef CONFIG_STATIC
+LDFLAGS+=-static
+endif
+
+ifeq ($(ARCH),i386)
+CFLAGS+=-fomit-frame-pointer
+OP_CFLAGS=$(CFLAGS) -mpreferred-stack-boundary=2
+ifeq ($(HAVE_GCC3_OPTIONS),yes)
+OP_CFLAGS+= -falign-functions=0
+else
+OP_CFLAGS+= -malign-functions=0
+endif
+ifdef TARGET_GPROF
+LDFLAGS+=-Wl,-T,$(SRC_PATH)/i386.ld
+else
+# WARNING: this LDFLAGS is _very_ tricky : qemu is an ELF shared object
+# that the kernel ELF loader considers as an executable. I think this
+# is the simplest way to make it self virtualizable!
+LDFLAGS+=-Wl,-shared
+endif
+ifeq ($(TARGET_ARCH), i386)
+PROGS+=vl
+endif
+endif
+
+ifeq ($(ARCH),ppc)
+OP_CFLAGS=$(CFLAGS)
+LDFLAGS+=-Wl,-T,$(SRC_PATH)/ppc.ld
+endif
+
+ifeq ($(ARCH),s390)
+OP_CFLAGS=$(CFLAGS)
+LDFLAGS+=-Wl,-T,$(SRC_PATH)/s390.ld
+endif
+
+ifeq ($(ARCH),sparc)
+CFLAGS+=-m32 -ffixed-g1 -ffixed-g2 -ffixed-g3 -ffixed-g6
+LDFLAGS+=-m32
+OP_CFLAGS=$(CFLAGS) -fno-delayed-branch -ffixed-i0
+HELPER_CFLAGS=$(CFLAGS) -ffixed-i0 -mflat
+# -static is used to avoid g1/g3 usage by the dynamic linker
+LDFLAGS+=-Wl,-T,$(SRC_PATH)/sparc.ld -static
+endif
+
+ifeq ($(ARCH),sparc64)
+CFLAGS+=-m64 -ffixed-g1 -ffixed-g2 -ffixed-g3 -ffixed-g6
+LDFLAGS+=-m64
+OP_CFLAGS=$(CFLAGS) -fno-delayed-branch -ffixed-i0
+endif
+
+ifeq ($(ARCH),alpha)
+# -msmall-data is not used because we want two-instruction relocations
+# for the constant constructions
+OP_CFLAGS=-Wall -O2 -g
+# Ensure there's only a single GP
+CFLAGS += -msmall-data
+LDFLAGS+=-Wl,-T,$(SRC_PATH)/alpha.ld
+endif
+
+ifeq ($(ARCH),ia64)
+OP_CFLAGS=$(CFLAGS)
+endif
+
+ifeq ($(ARCH),arm)
+OP_CFLAGS=$(CFLAGS) -mno-sched-prolog
+LDFLAGS+=-Wl,-T,$(SRC_PATH)/arm.ld
+endif
+
+ifeq ($(ARCH),m68k)
+OP_CFLAGS=$(CFLAGS) -fomit-frame-pointer
+LDFLAGS+=-Wl,-T,m68k.ld
+endif
+
+ifeq ($(HAVE_GCC3_OPTIONS),yes)
+# very important to generate a return at the end of every operation
+OP_CFLAGS+=-fno-reorder-blocks -fno-optimize-sibling-calls
+endif
+
+#########################################################
+
+DEFINES+=-D_GNU_SOURCE
+LIBS+=-lm
+
+# profiling code
+ifdef TARGET_GPROF
+LDFLAGS+=-p
+main.o: CFLAGS+=-p
+endif
+
+OBJS= elfload.o main.o syscall.o mmap.o signal.o path.o
+ifeq ($(TARGET_ARCH), i386)
+OBJS+= vm86.o
+endif
+SRCS:= $(OBJS:.o=.c)
+OBJS+= libqemu.a
+
+# cpu emulator library
+LIBOBJS=thunk.o exec.o translate.o cpu-exec.o gdbstub.o
+
+ifeq ($(TARGET_ARCH), i386)
+LIBOBJS+=translate-i386.o op-i386.o helper-i386.o helper2-i386.o
+endif
+ifeq ($(TARGET_ARCH), arm)
+LIBOBJS+=translate-arm.o op-arm.o
+endif
+
+# NOTE: the disassembler code is only needed for debugging
+LIBOBJS+=disas.o 
+ifeq ($(findstring i386, $(TARGET_ARCH) $(ARCH)),i386)
+LIBOBJS+=i386-dis.o
+endif
+ifeq ($(findstring alpha, $(TARGET_ARCH) $(ARCH)),alpha)
+LIBOBJS+=alpha-dis.o
+endif
+ifeq ($(findstring ppc, $(TARGET_ARCH) $(ARCH)),ppc)
+LIBOBJS+=ppc-dis.o
+endif
+ifeq ($(findstring sparc, $(TARGET_ARCH) $(ARCH)),sparc)
+LIBOBJS+=sparc-dis.o
+endif
+ifeq ($(findstring arm, $(TARGET_ARCH) $(ARCH)),arm)
+LIBOBJS+=arm-dis.o
+endif
+
+ifeq ($(ARCH),ia64)
+OBJS += ia64-syscall.o
+endif
+
+all: $(PROGS) qemu-doc.html
+
+qemu: $(OBJS)
+	$(CC) $(LDFLAGS) -o $@ $^  $(LIBS)
+ifeq ($(ARCH),alpha)
+# Mark as 32 bit binary, i. e. it will be mapped into the low 31 bit of
+# the address space (31 bit so sign extending doesn't matter)
+	echo -ne '\001\000\000\000' | dd of=qemu bs=1 seek=48 count=4 conv=notrunc
+endif
+
+# must use static linking to avoid leaving stuff in virtual address space
+VL_OBJS=vl.o block.o vga.o
+ifdef CONFIG_SDL
+VL_OBJS+=sdl.o
+SDL_LIBS+=-L/usr/X11R6/lib -lX11 -lXext -lXv -ldl
+endif
+
+vl: $(VL_OBJS) libqemu.a
+	$(CC) -static -Wl,-T,$(SRC_PATH)/i386-vl.ld -o $@ $^ $(LIBS) $(SDL_LIBS)
+
+sdl.o: sdl.c
+	$(CC) $(CFLAGS) $(DEFINES) $(SDL_CFLAGS) -c -o $@ $<
+
+depend: $(SRCS)
+	$(CC) -MM $(CFLAGS) $^ 1>.depend
+
+# libqemu 
+
+libqemu.a: $(LIBOBJS)
+	rm -f $@
+	$(AR) rcs $@ $(LIBOBJS)
+
+translate-$(TARGET_ARCH).o: translate-$(TARGET_ARCH).c gen-op-$(TARGET_ARCH).h opc-$(TARGET_ARCH).h cpu-$(TARGET_ARCH).h
+
+translate.o: translate.c op-$(TARGET_ARCH).h opc-$(TARGET_ARCH).h cpu-$(TARGET_ARCH).h
+
+op-$(TARGET_ARCH).h: op-$(TARGET_ARCH).o $(DYNGEN)
+	$(DYNGEN) -o $@ $<
+
+opc-$(TARGET_ARCH).h: op-$(TARGET_ARCH).o $(DYNGEN)
+	$(DYNGEN) -c -o $@ $<
+
+gen-op-$(TARGET_ARCH).h: op-$(TARGET_ARCH).o $(DYNGEN)
+	$(DYNGEN) -g -o $@ $<
+
+op-$(TARGET_ARCH).o: op-$(TARGET_ARCH).c
+	$(CC) $(OP_CFLAGS) $(DEFINES) -c -o $@ $<
+
+helper-$(TARGET_ARCH).o: helper-$(TARGET_ARCH).c
+	$(CC) $(HELPER_CFLAGS) $(DEFINES) -c -o $@ $<
+
+op-i386.o: op-i386.c opreg_template.h ops_template.h ops_template_mem.h ops_mem.h
+
+op-arm.o: op-arm.c op-arm-template.h
+
+%.o: %.c
+	$(CC) $(CFLAGS) $(DEFINES) -c -o $@ $<
+
+clean:
+	rm -f *.o  *.a *~ $(PROGS) \
+           gen-op-$(TARGET_ARCH).h opc-$(TARGET_ARCH).h op-$(TARGET_ARCH).h
+
+ifneq ($(wildcard .depend),)
+include .depend
+endif
--- a/22
+++ b/22
@@ -45,22 +45,30 @@ that QEMU works if you do not use a tested gcc version. Look at
 'configure' and 'Makefile' if you want to make a different gcc
 version work.

-host      gcc      binutils      glibc    linux
-------------------------------------------------------
-x86       2.95.2   2.13.2        2.1.3    2.4.18
+host      gcc      binutils      glibc    linux       distribution
+----------------------------------------------------------------------
+x86       2.95.2   2.13.2        2.1.3    2.4.18           
+          3.2      2.13.2        2.1.3    2.4.18
+          2.96     2.11.93.0.2   2.2.5    2.4.18      Red Hat 7.3
+          3.2.2    2.13.90.0.18  2.3.2    2.4.20      Red Hat 9

-PowerPC   2.95.4   2.12.90.0.1   2.2.5    2.4.20-pre2
+PowerPC   3.3 [4]  2.13.90.0.18  2.3.1    2.4.20briq
+          3.2

-Alpha     3.3 [1]  2.14.90.0.4   2.2.5    2.2.20 [2]
+Alpha     3.3 [1]  2.14.90.0.4   2.2.5    2.2.20 [2]  Debian 3.0

-Sparc32   2.95.4   2.12.90.0.1   2.2.5    2.4.18
+Sparc32   2.95.4   2.12.90.0.1   2.2.5    2.4.18      Debian 3.0

-ARM       2.95.4   2.12.90.0.1   2.2.5    2.4.9-ac10-rmk2-np1-cerf2
+ARM       2.95.4   2.12.90.0.1   2.2.5    2.4.9 [3]   Debian 3.0

 [1] On Alpha, QEMU needs the gcc 'visibility' attribute only available
    for gcc version >= 3.3.
 [2] Linux >= 2.4.20 is necessary for precise exception support
    (untested).
+[3] 2.4.9-ac10-rmk2-np1-cerf2
+
+[4] gcc 2.95.x generates invalid code when using too many register
+variables. You must use gcc 3.x on PowerPC.

 Documentation
 -------------
--- a/13
+++ b/13
@@ -1,16 +1,17 @@
-
- fix arm fpu rounding (at least for float->integer conversions)
+- optimize FPU operations (evaluate x87 stack pointer statically) and
+  fix cr0.TS emulation
 - fix CCOP optimisation
- optimize FPU operations (evaluate x87 stack pointer statically) 
- fpush not before mem load restarting
 - fix all remaining thread lock issues (must put TBs in a specific invalid
  state, find a solution for tb_flush()).
- add gcc 2.96 test configure (some gcc3 flags are needed)
+- cpu loop optimisation (optimise ret case as the cpu state does not change)
+- fix arm fpu rounding (at least for float->integer conversions)
 - add IPC syscalls
- submit a patch to fix DOSEMU coopthreads

 lower priority:
 --------------
+- sysenter/sysexit emulation
+- SMP support
+- finish segment ops (call far, ret far, load_seg suppressed)
 - use -msoft-float on ARM
 - use kernel traps for unaligned accesses on ARM ?
 - handle rare page fault cases (in particular if page fault in heplers or
--- a/2
+++ b/2
@@ -1 +1 @@
-0.3
+0.4.4
--- a/arm-dis.c
+++ b/arm-dis.c
@@ -560,8 +560,8 @@ static arm_regname regnames[] =
    { "a1", "a2", "a3", "a4", "v1", "v2", "v3", "WR", "v5", "SB", "SL",  "FP",  "IP",  "SP",  "LR",  "PC" }}
 };

-/* Default to GCC register name set.  */
-static unsigned int regname_selected = 1;
+/* Default to STD register name set.  */
+static unsigned int regname_selected = 2;

 #define NUM_ARM_REGNAMES  NUM_ELEM (regnames)
 #define arm_regnames      regnames[regname_selected].reg_names
--- a/block.c
+++ b/block.c
@@ -0,0 +1,326 @@
+/*
+ * QEMU System Emulator block driver
+ * 
+ * Copyright (c) 2003 Fabrice Bellard
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+#include <stdlib.h>
+#include <stdio.h>
+#include <stdarg.h>
+#include <string.h>
+#include <getopt.h>
+#include <inttypes.h>
+#include <unistd.h>
+#include <sys/mman.h>
+#include <fcntl.h>
+#include <signal.h>
+#include <time.h>
+#include <sys/time.h>
+#include <malloc.h>
+#include <termios.h>
+#include <sys/poll.h>
+#include <errno.h>
+#include <sys/wait.h>
+#include <netinet/in.h>
+
+#include "vl.h"
+
+#define NO_THUNK_TYPE_SIZE
+#include "thunk.h"
+
+struct BlockDriverState {
+    int fd; /* if -1, only COW mappings */
+    int64_t total_sectors;
+    int read_only;
+
+    uint8_t *cow_bitmap; /* if non NULL, COW mappings are used first */
+    uint8_t *cow_bitmap_addr; /* mmap address of cow_bitmap */
+    int cow_bitmap_size;
+    int cow_fd;
+    int64_t cow_sectors_offset;
+    char filename[1024];
+};
+
+BlockDriverState *bdrv_open(const char *filename, int snapshot)
+{
+    BlockDriverState *bs;
+    int fd, cow_fd;
+    int64_t size;
+    char template[] = "/tmp/vl.XXXXXX";
+    struct cow_header_v2 cow_header;
+    struct stat st;
+
+    bs = malloc(sizeof(BlockDriverState));
+    if(!bs)
+        return NULL;
+    bs->read_only = 0;
+    bs->fd = -1;
+    bs->cow_fd = -1;
+    bs->cow_bitmap = NULL;
+    strcpy(bs->filename, filename);
+
+    /* open standard HD image */
+    fd = open(filename, O_RDWR | O_LARGEFILE);
+    if (fd < 0) {
+        /* read only image on disk */
+        fd = open(filename, O_RDONLY | O_LARGEFILE);
+        if (fd < 0) {
+            perror(filename);
+            goto fail;
+        }
+        if (!snapshot)
+            bs->read_only = 1;
+    }
+    bs->fd = fd;
+
+    /* see if it is a cow image */
+    if (read(fd, &cow_header, sizeof(cow_header)) != sizeof(cow_header)) {
+        fprintf(stderr, "%s: could not read header\n", filename);
+        goto fail;
+    }
+    if (cow_header.magic == htonl(COW_MAGIC) &&
+        cow_header.version == htonl(COW_VERSION)) {
+        /* cow image found */
+        size = cow_header.size;
+#ifndef WORDS_BIGENDIAN
+        size = bswap64(size);
+#endif    
+        bs->total_sectors = size / 512;
+
+        bs->cow_fd = fd;
+        bs->fd = -1;
+        if (cow_header.backing_file[0] != '\0') {
+            if (stat(cow_header.backing_file, &st) != 0) {
+                fprintf(stderr, "%s: could not find original disk image '%s'\n", filename, cow_header.backing_file);
+                goto fail;
+            }
+            if (st.st_mtime != htonl(cow_header.mtime)) {
+                fprintf(stderr, "%s: original raw disk image '%s' does not match saved timestamp\n", filename, cow_header.backing_file);
+                goto fail;
+            }
+            fd = open(cow_header.backing_file, O_RDONLY | O_LARGEFILE);
+            if (fd < 0)
+                goto fail;
+            bs->fd = fd;
+        }
+        /* mmap the bitmap */
+        bs->cow_bitmap_size = ((bs->total_sectors + 7) >> 3) + sizeof(cow_header);
+        bs->cow_bitmap_addr = mmap(get_mmap_addr(bs->cow_bitmap_size), 
+                                   bs->cow_bitmap_size, 
+                                   PROT_READ | PROT_WRITE,
+                                   MAP_SHARED, bs->cow_fd, 0);
+        if (bs->cow_bitmap_addr == MAP_FAILED)
+            goto fail;
+        bs->cow_bitmap = bs->cow_bitmap_addr + sizeof(cow_header);
+        bs->cow_sectors_offset = (bs->cow_bitmap_size + 511) & ~511;
+        snapshot = 0;
+    } else {
+        /* standard raw image */
+        size = lseek64(fd, 0, SEEK_END);
+        bs->total_sectors = size / 512;
+        bs->fd = fd;
+    }
+
+    if (snapshot) {
+        /* create a temporary COW file */
+        cow_fd = mkstemp(template);
+        if (cow_fd < 0)
+            goto fail;
+        bs->cow_fd = cow_fd;
+	unlink(template);
+        
+        /* just need to allocate bitmap */
+        bs->cow_bitmap_size = (bs->total_sectors + 7) >> 3;
+        bs->cow_bitmap_addr = mmap(get_mmap_addr(bs->cow_bitmap_size), 
+                                   bs->cow_bitmap_size, 
+                                   PROT_READ | PROT_WRITE,
+                                   MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
+        if (bs->cow_bitmap_addr == MAP_FAILED)
+            goto fail;
+        bs->cow_bitmap = bs->cow_bitmap_addr;
+        bs->cow_sectors_offset = 0;
+    }
+    
+    return bs;
+ fail:
+    bdrv_close(bs);
+    return NULL;
+}
+
+void bdrv_close(BlockDriverState *bs)
+{
+    /* we unmap the mapping so that it is written to the COW file */
+    if (bs->cow_bitmap_addr)
+        munmap(bs->cow_bitmap_addr, bs->cow_bitmap_size);
+    if (bs->cow_fd >= 0)
+        close(bs->cow_fd);
+    if (bs->fd >= 0)
+        close(bs->fd);
+    free(bs);
+}
+
+static inline void set_bit(uint8_t *bitmap, int64_t bitnum)
+{
+    bitmap[bitnum / 8] |= (1 << (bitnum%8));
+}
+
+static inline int is_bit_set(const uint8_t *bitmap, int64_t bitnum)
+{
+    return !!(bitmap[bitnum / 8] & (1 << (bitnum%8)));
+}
+
+
+/* Return true if first block has been changed (ie. current version is
+ * in COW file).  Set the number of continuous blocks for which that
+ * is true. */
+static int is_changed(uint8_t *bitmap,
+                      int64_t sector_num, int nb_sectors,
+                      int *num_same)
+{
+    int changed;
+
+    if (!bitmap || nb_sectors == 0) {
+	*num_same = nb_sectors;
+	return 0;
+    }
+
+    changed = is_bit_set(bitmap, sector_num);
+    for (*num_same = 1; *num_same < nb_sectors; (*num_same)++) {
+	if (is_bit_set(bitmap, sector_num + *num_same) != changed)
+	    break;
+    }
+
+    return changed;
+}
+
+/* commit COW file into the raw image */
+int bdrv_commit(BlockDriverState *bs)
+{
+    int64_t i;
+    uint8_t *cow_bitmap;
+
+    if (!bs->cow_bitmap) {
+	fprintf(stderr, "Already committed to %s\n", bs->filename);
+	return 0;
+    }
+
+    if (bs->read_only) {
+	fprintf(stderr, "Can't commit to %s: read-only\n", bs->filename);
+	return -1;
+    }
+
+    cow_bitmap = bs->cow_bitmap;
+    for (i = 0; i < bs->total_sectors; i++) {
+	if (is_bit_set(cow_bitmap, i)) {
+	    unsigned char sector[512];
+	    if (bdrv_read(bs, i, sector, 1) != 0) {
+		fprintf(stderr, "Error reading sector %lli: aborting commit\n",
+			(long long)i);
+		return -1;
+	    }
+
+	    /* Make bdrv_write write to real file for a moment. */
+	    bs->cow_bitmap = NULL;
+	    if (bdrv_write(bs, i, sector, 1) != 0) {
+		fprintf(stderr, "Error writing sector %lli: aborting commit\n",
+			(long long)i);
+		bs->cow_bitmap = cow_bitmap;
+		return -1;
+	    }
+	    bs->cow_bitmap = cow_bitmap;
+	}
+    }
+    fprintf(stderr, "Committed snapshot to %s\n", bs->filename);
+    return 0;
+}
+
+/* return -1 if error */
+int bdrv_read(BlockDriverState *bs, int64_t sector_num, 
+              uint8_t *buf, int nb_sectors)
+{
+    int ret, n, fd;
+    int64_t offset;
+    
+    while (nb_sectors > 0) {
+        if (is_changed(bs->cow_bitmap, sector_num, nb_sectors, &n)) {
+            fd = bs->cow_fd;
+            offset = bs->cow_sectors_offset;
+        } else {
+            fd = bs->fd;
+            offset = 0;
+        }
+
+        if (fd < 0) {
+            /* no file, just return empty sectors */
+            memset(buf, 0, n * 512);
+        } else {
+            offset += sector_num * 512;
+            lseek64(fd, offset, SEEK_SET);
+            ret = read(fd, buf, n * 512);
+            if (ret != n * 512) {
+                return -1;
+            }
+        }
+        nb_sectors -= n;
+        sector_num += n;
+        buf += n * 512;
+    }
+    return 0;
+}
+
+/* return -1 if error */
+int bdrv_write(BlockDriverState *bs, int64_t sector_num, 
+               const uint8_t *buf, int nb_sectors)
+{
+    int ret, fd, i;
+    int64_t offset, retl;
+
+    if (bs->read_only)
+        return -1;
+
+    if (bs->cow_bitmap) {
+        fd = bs->cow_fd;
+        offset = bs->cow_sectors_offset;
+    } else {
+        fd = bs->fd;
+        offset = 0;
+    }
+    
+    offset += sector_num * 512;
+    retl = lseek64(fd, offset, SEEK_SET);
+    if (retl == -1) {
+        return -1;
+    }
+    ret = write(fd, buf, nb_sectors * 512);
+    if (ret != nb_sectors * 512) {
+        return -1;
+    }
+
+    if (bs->cow_bitmap) {
+	for (i = 0; i < nb_sectors; i++)
+	    set_bit(bs->cow_bitmap, sector_num + i);
+    }
+    return 0;
+}
+
+void bdrv_get_geometry(BlockDriverState *bs, int64_t *nb_sectors_ptr)
+{
+    *nb_sectors_ptr = bs->total_sectors;
+}
--- a/bswap.h
+++ b/bswap.h
@@ -0,0 +1,84 @@
+#ifndef BSWAP_H
+#define BSWAP_H
+
+#include "config-host.h"
+
+#include <inttypes.h>
+
+#ifdef HAVE_BYTESWAP_H
+#include <byteswap.h>
+#else
+
+#define bswap_16(x) \
+({ \
+	uint16_t __x = (x); \
+	((uint16_t)( \
+		(((uint16_t)(__x) & (uint16_t)0x00ffU) << 8) | \
+		(((uint16_t)(__x) & (uint16_t)0xff00U) >> 8) )); \
+})
+
+#define bswap_32(x) \
+({ \
+	uint32_t __x = (x); \
+	((uint32_t)( \
+		(((uint32_t)(__x) & (uint32_t)0x000000ffUL) << 24) | \
+		(((uint32_t)(__x) & (uint32_t)0x0000ff00UL) <<  8) | \
+		(((uint32_t)(__x) & (uint32_t)0x00ff0000UL) >>  8) | \
+		(((uint32_t)(__x) & (uint32_t)0xff000000UL) >> 24) )); \
+})
+
+#define bswap_64(x) \
+({ \
+	uint64_t __x = (x); \
+	((uint64_t)( \
+		(uint64_t)(((uint64_t)(__x) & (uint64_t)0x00000000000000ffULL) << 56) | \
+		(uint64_t)(((uint64_t)(__x) & (uint64_t)0x000000000000ff00ULL) << 40) | \
+		(uint64_t)(((uint64_t)(__x) & (uint64_t)0x0000000000ff0000ULL) << 24) | \
+		(uint64_t)(((uint64_t)(__x) & (uint64_t)0x00000000ff000000ULL) <<  8) | \
+	        (uint64_t)(((uint64_t)(__x) & (uint64_t)0x000000ff00000000ULL) >>  8) | \
+		(uint64_t)(((uint64_t)(__x) & (uint64_t)0x0000ff0000000000ULL) >> 24) | \
+		(uint64_t)(((uint64_t)(__x) & (uint64_t)0x00ff000000000000ULL) >> 40) | \
+		(uint64_t)(((uint64_t)(__x) & (uint64_t)0xff00000000000000ULL) >> 56) )); \
+})
+
+#endif /* !HAVE_BYTESWAP_H */
+
+#if defined(__alpha__) || defined (__ia64__)
+#define HOST_LONG_BITS 64
+#else
+#define HOST_LONG_BITS 32
+#endif
+
+#define HOST_LONG_SIZE (HOST_LONG_BITS / 8)
+
+static inline uint16_t bswap16(uint16_t x)
+{
+    return bswap_16(x);
+}
+
+static inline uint32_t bswap32(uint32_t x) 
+{
+    return bswap_32(x);
+}
+
+static inline uint64_t bswap64(uint64_t x) 
+{
+    return bswap_64(x);
+}
+
+static inline void bswap16s(uint16_t *s)
+{
+    *s = bswap16(*s);
+}
+
+static inline void bswap32s(uint32_t *s)
+{
+    *s = bswap32(*s);
+}
+
+static inline void bswap64s(uint64_t *s)
+{
+    *s = bswap64(*s);
+}
+
+#endif /* BSWAP_H */
--- a/240
+++ b/240
@@ -15,7 +15,6 @@ TMPC="${TMPDIR1}/qemu-conf-${RANDOM}-$$-${RANDOM}.c"
 TMPO="${TMPDIR1}/qemu-conf-${RANDOM}-$$-${RANDOM}.o"
 TMPE="${TMPDIR1}/qemu-conf-${RANDOM}-$$-${RANDOM}"
 TMPS="${TMPDIR1}/qemu-conf-${RANDOM}-$$-${RANDOM}.S"
-TMPH="${TMPDIR1}/qemu-conf-${RANDOM}-$$-${RANDOM}.h"

 # default parameters
 prefix="/usr/local"
@@ -27,12 +26,11 @@ host_cc="gcc"
 ar="ar"
 make="make"
 strip="strip"
-target_cpu="x86"
-target_bigendian="default"
 cpu=`uname -m`
+target_list="i386 i386-softmmu arm"
 case "$cpu" in
  i386|i486|i586|i686|i86pc|BePC)
-    cpu="x86"
+    cpu="i386"
  ;;
  armv4l)
    cpu="armv4l"
@@ -58,6 +56,9 @@ case "$cpu" in
  ia64)
    cpu="ia64"
  ;;
+  m68k)
+    cpu="m68k"
+  ;;
  *)
    cpu="unknown"
  ;;
@@ -71,6 +72,26 @@ case $targetos in
 *) ;;
 esac

+##########################################
+# SDL probe
+
+cat > $TMPC << EOF
+#include <SDL.h>
+#undef main /* We don't want SDL to override our main() */
+int main( void ) { return SDL_Init (SDL_INIT_VIDEO); }
+EOF
+
+sdl_too_old=no
+sdl=no
+if $cc -o $TMPE `sdl-config --cflags` $TMPC `sdl-config --libs`  2> /dev/null  ; then
+_sdlversion=`sdl-config --version | sed 's/[^0-9]//g'`
+if test "$_sdlversion" -lt 121 ; then
+sdl_too_old=yes
+else
+sdl=yes
+fi
+fi
+
 # find source path
 # XXX: we assume an absolute path is given when launching configure, 
 # except in './configure' case.
@@ -104,16 +125,14 @@ for opt do
  ;;
  --cpu=*) cpu=`echo $opt | cut -d '=' -f 2`
  ;;
-  --target-cpu=*) target_cpu=`echo $opt | cut -d '=' -f 2`
-  ;;
-  --target-big-endian) target_bigendian="yes"
-  ;;
-  --target-little-endian) target_bigendian="no"
+  --target-list=*) target_list=${opt#--target-list=}
  ;;
  --enable-gprof) gprof="yes"
  ;;
  --static) static="yes"
  ;;
+  --disable-sdl) sdl="no"
+  ;;
  esac
 done

@@ -147,36 +166,21 @@ fi
 else

 # if cross compiling, cannot launch a program, so make a static guess
-if test "$cpu" = "powerpc" -o "$cpu" = "mips" -o "$cpu" = "s390" -o "$cpu" = "sparc" -o "$cpu" = "sparc64"; then
+if test "$cpu" = "powerpc" -o "$cpu" = "mips" -o "$cpu" = "s390" -o "$cpu" = "sparc" -o "$cpu" = "sparc64" -o "$cpu" = "m68k"; then
    bigendian="yes"
 fi

 fi

-# check gcc version
+# check gcc options support
 cat > $TMPC <<EOF
 int main(void) {
-#if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 2)
-return 0;
-#else
-#error gcc < 3.2
-#endif
 }
 EOF

-gcc_major="2"
-if $cc -o $TMPO $TMPC 2> /dev/null ; then
-   gcc_major="3"
-fi
-
-if test "$target_bigendian" = "default" ; then
-    if test "$target_cpu" = "x86" ; then
-         target_bigendian="no"
-    elif test "$target_cpu" = "arm" ; then
-         target_bigendian="no"
-    else
-         target_bigendian="no"
-    fi
+have_gcc3_options="no"
+if $cc -fno-reorder-blocks -fno-optimize-sibling-calls -o $TMPO $TMPC 2> /dev/null ; then
+   have_gcc3_options="yes"
 fi

 if test x"$1" = x"-h" -o x"$1" = x"--help" ; then
@@ -190,7 +194,7 @@ echo "Standard options:"
 echo "  --help                   print this message"
 echo "  --prefix=PREFIX          install in PREFIX [$prefix]"
 echo "  --interp-prefix=PREFIX   where to find shared libraries, etc. [$interp_prefix]"
-echo "  --target_cpu=CPU         set target cpu (x86 or arm) [$target_cpu]"
+echo "  --target-list=LIST       set target list [$target_list]"
 echo ""
 echo "Advanced options (experts only):"
 echo "  --source-path=PATH       path of source code [$source_path]"
@@ -210,92 +214,142 @@ echo "C compiler        $cc"
 echo "make              $make"
 echo "host CPU          $cpu"
 echo "host big endian   $bigendian"
-echo "target CPU        $target_cpu"
-echo "target big endian $target_bigendian"
+echo "target list       $target_list"
 echo "gprof enabled     $gprof"
 echo "static build      $static"
+echo "SDL support       $sdl"
+if test $sdl_too_old = "yes"; then
+echo "-> Your SDL version is too old - please upgrade to have FFplay/SDL support"
+fi

-echo "Creating config.mak and config.h"
+config_mak="config-host.mak"
+config_h="config-host.h"

-echo "# Automatically generated by configure - do not modify" > config.mak
-echo "/* Automatically generated by configure - do not modify */" > $TMPH
+echo "Creating $config_mak and $config_h"

-echo "prefix=$prefix" >> config.mak
-echo "#define CONFIG_QEMU_PREFIX \"$interp_prefix\"" >> $TMPH
-echo "MAKE=$make" >> config.mak
-echo "CC=$cc" >> config.mak
-echo "GCC_MAJOR=$gcc_major" >> config.mak
-echo "HOST_CC=$host_cc" >> config.mak
-echo "AR=$ar" >> config.mak
-echo "STRIP=$strip -s -R .comment -R .note" >> config.mak
-echo "CFLAGS=$CFLAGS" >> config.mak
-echo "LDFLAGS=$LDFLAGS" >> config.mak
-if test "$cpu" = "x86" ; then
-  echo "ARCH=i386" >> config.mak
-  echo "#define HOST_I386 1" >> $TMPH
+echo "# Automatically generated by configure - do not modify" > $config_mak
+echo "/* Automatically generated by configure - do not modify */" > $config_h
+
+echo "prefix=$prefix" >> $config_mak
+echo "MAKE=$make" >> $config_mak
+echo "CC=$cc" >> $config_mak
+if test "$have_gcc3_options" = "yes" ; then
+  echo "HAVE_GCC3_OPTIONS=yes" >> $config_mak
+fi
+echo "HOST_CC=$host_cc" >> $config_mak
+echo "AR=$ar" >> $config_mak
+echo "STRIP=$strip -s -R .comment -R .note" >> $config_mak
+echo "CFLAGS=$CFLAGS" >> $config_mak
+echo "LDFLAGS=$LDFLAGS" >> $config_mak
+if test "$cpu" = "i386" ; then
+  echo "ARCH=i386" >> $config_mak
+  echo "#define HOST_I386 1" >> $config_h
 elif test "$cpu" = "armv4l" ; then
-  echo "ARCH=arm" >> config.mak
-  echo "#define HOST_ARM 1" >> $TMPH
+  echo "ARCH=arm" >> $config_mak
+  echo "#define HOST_ARM 1" >> $config_h
 elif test "$cpu" = "powerpc" ; then
-  echo "ARCH=ppc" >> config.mak
-  echo "#define HOST_PPC 1" >> $TMPH
+  echo "ARCH=ppc" >> $config_mak
+  echo "#define HOST_PPC 1" >> $config_h
 elif test "$cpu" = "mips" ; then
-  echo "ARCH=mips" >> config.mak
-  echo "#define HOST_MIPS 1" >> $TMPH
+  echo "ARCH=mips" >> $config_mak
+  echo "#define HOST_MIPS 1" >> $config_h
 elif test "$cpu" = "s390" ; then
-  echo "ARCH=s390" >> config.mak
-  echo "#define HOST_S390 1" >> $TMPH
+  echo "ARCH=s390" >> $config_mak
+  echo "#define HOST_S390 1" >> $config_h
 elif test "$cpu" = "alpha" ; then
-  echo "ARCH=alpha" >> config.mak
-  echo "#define HOST_ALPHA 1" >> $TMPH
+  echo "ARCH=alpha" >> $config_mak
+  echo "#define HOST_ALPHA 1" >> $config_h
 elif test "$cpu" = "sparc" ; then
-  echo "ARCH=sparc" >> config.mak
-  echo "#define HOST_SPARC 1" >> $TMPH
+  echo "ARCH=sparc" >> $config_mak
+  echo "#define HOST_SPARC 1" >> $config_h
 elif test "$cpu" = "sparc64" ; then
-  echo "ARCH=sparc64" >> config.mak
-  echo "#define HOST_SPARC64 1" >> $TMPH
+  echo "ARCH=sparc64" >> $config_mak
+  echo "#define HOST_SPARC64 1" >> $config_h
 elif test "$cpu" = "ia64" ; then
-  echo "ARCH=ia64" >> config.mak
-  echo "#define HOST_IA64 1" >> $TMPH
+  echo "ARCH=ia64" >> $config_mak
+  echo "#define HOST_IA64 1" >> $config_h
+elif test "$cpu" = "m68k" ; then
+  echo "ARCH=m68k" >> config.mak
+  echo "#define HOST_M68K 1" >> $TMPH
 else
  echo "Unsupported CPU"
  exit 1
 fi
 if test "$bigendian" = "yes" ; then
-  echo "WORDS_BIGENDIAN=yes" >> config.mak
-  echo "#define WORDS_BIGENDIAN 1" >> $TMPH
+  echo "WORDS_BIGENDIAN=yes" >> $config_mak
+  echo "#define WORDS_BIGENDIAN 1" >> $config_h
+fi
+echo "#define HAVE_BYTESWAP_H 1" >> $config_h
+if test "$gprof" = "yes" ; then
+  echo "TARGET_GPROF=yes" >> $config_mak
+  echo "#define HAVE_GPROF 1" >> $config_h
+fi
+if test "$static" = "yes" ; then
+  echo "CONFIG_STATIC=yes" >> $config_mak
+fi
+if test "$sdl" = "yes" ; then
+  echo "CONFIG_SDL=yes" >> $config_mak
+  echo "#define CONFIG_SDL 1" >> $config_h
+  echo "SDL_LIBS=`sdl-config --libs`" >> $config_mak
+  echo "SDL_CFLAGS=`sdl-config --cflags`" >> $config_mak
+fi
+echo -n "VERSION=" >>$config_mak
+head $source_path/VERSION >>$config_mak
+echo "" >>$config_mak
+echo -n "#define QEMU_VERSION \"" >> $config_h
+head $source_path/VERSION >> $config_h
+echo "\"" >> $config_h
+
+echo "SRC_PATH=$source_path" >> $config_mak
+echo "TARGET_DIRS=$target_list" >> $config_mak
+
+for target in $target_list; do 
+
+target_dir="$target"
+config_mak=$target_dir/config.mak
+config_h=$target_dir/config.h
+target_cpu=`echo $target | cut -d '-' -f 1`
+target_bigendian="no"
+target_softmmu="no"
+if expr $target : '.*-softmmu' > /dev/null ; then
+  target_softmmu="yes"
 fi

-if test "$target_cpu" = "x86" ; then
-  echo "TARGET_ARCH=i386" >> config.mak
-  echo "#define TARGET_ARCH \"i386\"" >> $TMPH
-  echo "#define TARGET_I386 1" >> $TMPH
+echo "Creating $config_mak, $config_h and $target_dir/Makefile"
+
+mkdir -p $target_dir
+ln -sf $source_path/Makefile.target $target_dir/Makefile
+
+echo "# Automatically generated by configure - do not modify" > $config_mak
+echo "/* Automatically generated by configure - do not modify */" > $config_h
+
+
+echo "include ../config-host.mak" >> $config_mak
+echo "#include \"../config-host.h\"" >> $config_h
+echo "#define CONFIG_QEMU_PREFIX \"$interp_prefix\"" >> $config_h
+
+if test "$target_cpu" = "i386" ; then
+  echo "TARGET_ARCH=i386" >> $config_mak
+  echo "#define TARGET_ARCH \"i386\"" >> $config_h
+  echo "#define TARGET_I386 1" >> $config_h
 elif test "$target_cpu" = "arm" ; then
-  echo "TARGET_ARCH=arm" >> config.mak
-  echo "#define TARGET_ARCH \"arm\"" >> $TMPH
-  echo "#define TARGET_ARM 1" >> $TMPH
+  echo "TARGET_ARCH=arm" >> $config_mak
+  echo "#define TARGET_ARCH \"arm\"" >> $config_h
+  echo "#define TARGET_ARM 1" >> $config_h
 else
  echo "Unsupported target CPU"
  exit 1
 fi
 if test "$target_bigendian" = "yes" ; then
-  echo "TARGET_WORDS_BIGENDIAN=yes" >> config.mak
-  echo "#define TARGET_WORDS_BIGENDIAN 1" >> $TMPH
+  echo "TARGET_WORDS_BIGENDIAN=yes" >> $config_mak
+  echo "#define TARGET_WORDS_BIGENDIAN 1" >> $config_h
+fi
+if test "$target_softmmu" = "yes" ; then
+  echo "CONFIG_SOFTMMU=yes" >> $config_mak
+  echo "#define CONFIG_SOFTMMU 1" >> $config_h
 fi

-if test "$gprof" = "yes" ; then
-  echo "TARGET_GPROF=yes" >> config.mak
-  echo "#define HAVE_GPROF 1" >> $TMPH
-fi
-if test "$static" = "yes" ; then
-  echo "CONFIG_STATIC=yes" >> config.mak
-fi
-echo -n "VERSION=" >>config.mak
-head $source_path/VERSION >>config.mak
-echo "" >>config.mak
-echo -n "#define QEMU_VERSION \"" >> $TMPH
-head $source_path/VERSION >> $TMPH
-echo "\"" >> $TMPH
+done # for target in $targets

 # build tree in object directory if source path is different from current one
 if test "$source_path_used" = "yes" ; then
@@ -308,13 +362,5 @@ if test "$source_path_used" = "yes" ; then
        ln -sf $source_path/$f $f
    done
 fi
-echo "SRC_PATH=$source_path" >> config.mak

-diff $TMPH config.h >/dev/null 2>&1
-if test $? -ne 0 ; then
-	mv -f $TMPH config.h
-else
-	echo "config.h is unchanged"
-fi
-
-rm -f $TMPO $TMPC $TMPE $TMPS $TMPH
+rm -f $TMPO $TMPC $TMPE $TMPS
--- a/cpu-all.h
+++ b/cpu-all.h
@@ -140,6 +140,7 @@ static inline void stfl(void *ptr, float v)
    stl(ptr, u.i);
 }

+
 #if defined(__arm__) && !defined(WORDS_BIGENDIAN)

 /* NOTE: arm is horrible as double 32 bit words are stored in big endian ! */
@@ -302,6 +303,35 @@ void page_unprotect_range(uint8_t *data, unsigned long data_size);

 #endif

-#endif
+#endif /* SINGLE_CPU_DEFINES */
+
+#define DEFAULT_GDBSTUB_PORT 1234
+
+void cpu_abort(CPUState *env, const char *fmt, ...);
+extern CPUState *cpu_single_env;
+
+#define CPU_INTERRUPT_EXIT 0x01 /* wants exit from main loop */
+#define CPU_INTERRUPT_HARD 0x02 /* hardware interrupt pending */
+void cpu_interrupt(CPUState *s, int mask);
+
+int cpu_breakpoint_insert(CPUState *env, uint32_t pc);
+int cpu_breakpoint_remove(CPUState *env, uint32_t pc);
+void cpu_single_step(CPUState *env, int enabled);
+
+/* memory API */
+
+typedef void CPUWriteMemoryFunc(uint32_t addr, uint32_t value);
+typedef uint32_t CPUReadMemoryFunc(uint32_t addr);
+
+void cpu_register_physical_memory(unsigned long start_addr, unsigned long size,
+                                  long phys_offset);
+int cpu_register_io_memory(int io_index,
+                           CPUReadMemoryFunc **mem_read,
+                           CPUWriteMemoryFunc **mem_write);
+
+/* gdb stub API */
+extern int gdbstub_fd;
+CPUState *cpu_gdbstub_get_env(void *opaque);
+int cpu_gdbstub(void *opaque, int (*main_loop)(void *opaque), int port);

 #endif /* CPU_ALL_H */
--- a/cpu-arm.h
+++ b/cpu-arm.h
@@ -20,12 +20,10 @@
 #ifndef CPU_ARM_H
 #define CPU_ARM_H

-#include "config.h"
-#include <setjmp.h>
+#include "cpu-defs.h"

 #define EXCP_UDEF       1   /* undefined instruction */
 #define EXCP_SWI        2   /* software interrupt */
-#define EXCP_INTERRUPT 	256 /* async interruption */

 typedef struct CPUARMState {
    uint32_t regs[16];
@@ -40,6 +38,8 @@ typedef struct CPUARMState {
    jmp_buf jmp_env;
    int exception_index;
    int interrupt_request;
+    struct TranslationBlock *current_tb;
+    int user_mode_only;

    /* user data */
    void *opaque;
@@ -47,7 +47,6 @@ typedef struct CPUARMState {

 CPUARMState *cpu_arm_init(void);
 int cpu_arm_exec(CPUARMState *s);
-void cpu_arm_interrupt(CPUARMState *s);
 void cpu_arm_close(CPUARMState *s);
 /* you can call this signal handler from your SIGBUS and SIGSEGV
   signal handlers to inform the virtual CPU of exceptions. non zero
--- a/cpu-defs.h
+++ b/cpu-defs.h
@@ -0,0 +1,39 @@
+/*
+ * common defines for all CPUs
+ * 
+ * Copyright (c) 2003 Fabrice Bellard
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+#ifndef CPU_DEFS_H
+#define CPU_DEFS_H
+
+#include "config.h"
+#include <setjmp.h>
+
+#define EXCP_INTERRUPT 	256 /* async interruption */
+#define EXCP_HLT        257 /* hlt instruction reached */
+#define EXCP_DEBUG      258 /* cpu stopped after a breakpoint or singlestep */
+
+#define MAX_BREAKPOINTS 32
+
+#define CPU_TLB_SIZE 256
+
+typedef struct CPUTLBEntry {
+    uint32_t address;
+    uint32_t addend;
+} CPUTLBEntry;
+
+#endif
--- a/cpu-exec.c
+++ b/cpu-exec.c
@@ -71,7 +71,7 @@ int cpu_exec(CPUState *env1)
 #ifdef __sparc__
    int saved_i7, tmp_T0;
 #endif
-    int code_gen_size, ret;
+    int code_gen_size, ret, interrupt_request;
    void (*gen_func)(void);
    TranslationBlock *tb, **ptb;
    uint8_t *tc_ptr, *cs_base, *pc;
@@ -139,147 +139,193 @@ int cpu_exec(CPUState *env1)
 #else
 #error unsupported target CPU
 #endif
-    env->interrupt_request = 0;
+    env->exception_index = -1;

    /* prepare setjmp context for exception handling */
-    if (setjmp(env->jmp_env) == 0) {
-        T0 = 0; /* force lookup of first TB */
-        for(;;) {
-#ifdef __sparc__
-	  /* g1 can be modified by some libc? functions */ 
-	    tmp_T0 = T0;
-#endif	    
-            if (env->interrupt_request) {
-                env->exception_index = EXCP_INTERRUPT;
-                cpu_loop_exit();
-            }
-#ifdef DEBUG_EXEC
-            if (loglevel) {
+    for(;;) {
+        if (setjmp(env->jmp_env) == 0) {
+            /* if an exception is pending, we execute it here */
+            if (env->exception_index >= 0) {
+                if (env->exception_index >= EXCP_INTERRUPT) {
+                    /* exit request from the cpu execution loop */
+                    ret = env->exception_index;
+                    break;
+                } else if (env->user_mode_only) {
+                    /* if user mode only, we simulate a fake exception
+                       which will be hanlded outside the cpu execution
+                       loop */
 #if defined(TARGET_I386)
-                /* restore flags in standard format */
-                env->regs[R_EAX] = EAX;
-                env->regs[R_EBX] = EBX;
-                env->regs[R_ECX] = ECX;
-                env->regs[R_EDX] = EDX;
-                env->regs[R_ESI] = ESI;
-                env->regs[R_EDI] = EDI;
-                env->regs[R_EBP] = EBP;
-                env->regs[R_ESP] = ESP;
-                env->eflags = env->eflags | cc_table[CC_OP].compute_all() | (DF & DF_MASK);
-                cpu_x86_dump_state(env, logfile, 0);
-                env->eflags &= ~(DF_MASK | CC_O | CC_S | CC_Z | CC_A | CC_P | CC_C);
+                    do_interrupt_user(env->exception_index, 
+                                      env->exception_is_int, 
+                                      env->error_code, 
+                                      env->exception_next_eip);
+#endif
+                    ret = env->exception_index;
+                    break;
+                } else {
+#if defined(TARGET_I386)
+                    /* simulate a real cpu exception. On i386, it can
+                       trigger new exceptions, but we do not handle
+                       double or triple faults yet. */
+                    do_interrupt(env->exception_index, 
+                                 env->exception_is_int, 
+                                 env->error_code, 
+                                 env->exception_next_eip, 0);
+#endif
+                }
+                env->exception_index = -1;
+            }
+            T0 = 0; /* force lookup of first TB */
+            for(;;) {
+#ifdef __sparc__
+                /* g1 can be modified by some libc? functions */ 
+                tmp_T0 = T0;
+#endif	    
+                interrupt_request = env->interrupt_request;
+                if (__builtin_expect(interrupt_request, 0)) {
+#if defined(TARGET_I386)
+                    /* if hardware interrupt pending, we execute it */
+                    if ((interrupt_request & CPU_INTERRUPT_HARD) &&
+                        (env->eflags & IF_MASK) && 
+                        !(env->hflags & HF_INHIBIT_IRQ_MASK)) {
+                        int intno;
+                        intno = cpu_x86_get_pic_interrupt(env);
+                        if (loglevel) {
+                            fprintf(logfile, "Servicing hardware INT=0x%02x\n", intno);
+                        }
+                        do_interrupt(intno, 0, 0, 0, 1);
+                        env->interrupt_request &= ~CPU_INTERRUPT_HARD;
+                        /* ensure that no TB jump will be modified as
+                           the program flow was changed */
+#ifdef __sparc__
+                        tmp_T0 = 0;
+#else
+                        T0 = 0;
+#endif
+                    }
+#endif
+                    if (interrupt_request & CPU_INTERRUPT_EXIT) {
+                        env->interrupt_request &= ~CPU_INTERRUPT_EXIT;
+                        env->exception_index = EXCP_INTERRUPT;
+                        cpu_loop_exit();
+                    }
+                }
+#ifdef DEBUG_EXEC
+                if (loglevel) {
+#if defined(TARGET_I386)
+                    /* restore flags in standard format */
+                    env->regs[R_EAX] = EAX;
+                    env->regs[R_EBX] = EBX;
+                    env->regs[R_ECX] = ECX;
+                    env->regs[R_EDX] = EDX;
+                    env->regs[R_ESI] = ESI;
+                    env->regs[R_EDI] = EDI;
+                    env->regs[R_EBP] = EBP;
+                    env->regs[R_ESP] = ESP;
+                    env->eflags = env->eflags | cc_table[CC_OP].compute_all() | (DF & DF_MASK);
+                    cpu_x86_dump_state(env, logfile, X86_DUMP_CCOP);
+                    env->eflags &= ~(DF_MASK | CC_O | CC_S | CC_Z | CC_A | CC_P | CC_C);
 #elif defined(TARGET_ARM)
-                cpu_arm_dump_state(env, logfile, 0);
+                    env->cpsr = compute_cpsr();
+                    cpu_arm_dump_state(env, logfile, 0);
+                    env->cpsr &= ~0xf0000000;
 #else
 #error unsupported target CPU 
 #endif
-            }
+                }
 #endif
-            /* we compute the CPU state. We assume it will not
-               change during the whole generated block. */
+                /* we record a subset of the CPU state. It will
+                   always be the same before a given translated block
+                   is executed. */
 #if defined(TARGET_I386)
-            flags = env->seg_cache[R_CS].seg_32bit << GEN_FLAG_CODE32_SHIFT;
-            flags |= env->seg_cache[R_SS].seg_32bit << GEN_FLAG_SS32_SHIFT;
-            flags |= (((unsigned long)env->seg_cache[R_DS].base | 
-                       (unsigned long)env->seg_cache[R_ES].base |
-                       (unsigned long)env->seg_cache[R_SS].base) != 0) << 
-                GEN_FLAG_ADDSEG_SHIFT;
-            if (!(env->eflags & VM_MASK)) {
-                flags |= (env->segs[R_CS] & 3) << GEN_FLAG_CPL_SHIFT;
-            } else {
-                /* NOTE: a dummy CPL is kept */
-                flags |= (1 << GEN_FLAG_VM_SHIFT);
-                flags |= (3 << GEN_FLAG_CPL_SHIFT);
-            }
-            flags |= (env->eflags & (IOPL_MASK | TF_MASK));
-            cs_base = env->seg_cache[R_CS].base;
-            pc = cs_base + env->eip;
+                flags = env->hflags;
+                flags |= (env->eflags & (IOPL_MASK | TF_MASK | VM_MASK));
+                cs_base = env->segs[R_CS].base;
+                pc = cs_base + env->eip;
 #elif defined(TARGET_ARM)
-            flags = 0;
-            cs_base = 0;
-            pc = (uint8_t *)env->regs[15];
+                flags = 0;
+                cs_base = 0;
+                pc = (uint8_t *)env->regs[15];
 #else
 #error unsupported CPU
 #endif
-            tb = tb_find(&ptb, (unsigned long)pc, (unsigned long)cs_base, 
-                         flags);
-            if (!tb) {
-                spin_lock(&tb_lock);
-                /* if no translated code available, then translate it now */
-                tb = tb_alloc((unsigned long)pc);
+                tb = tb_find(&ptb, (unsigned long)pc, (unsigned long)cs_base, 
+                             flags);
                if (!tb) {
-                    /* flush must be done */
-                    tb_flush();
-                    /* cannot fail at this point */
+                    spin_lock(&tb_lock);
+                    /* if no translated code available, then translate it now */
                    tb = tb_alloc((unsigned long)pc);
-                    /* don't forget to invalidate previous TB info */
-                    ptb = &tb_hash[tb_hash_func((unsigned long)pc)];
-                    T0 = 0;
-                }
-                tc_ptr = code_gen_ptr;
-                tb->tc_ptr = tc_ptr;
-                tb->cs_base = (unsigned long)cs_base;
-                tb->flags = flags;
-                ret = cpu_gen_code(tb, CODE_GEN_MAX_SIZE, &code_gen_size);
-#if defined(TARGET_I386)
-                /* XXX: suppress that, this is incorrect */
-                /* if invalid instruction, signal it */
-                if (ret != 0) {
-                    /* NOTE: the tb is allocated but not linked, so we
-                       can leave it */
+                    if (!tb) {
+                        /* flush must be done */
+                        tb_flush();
+                        /* cannot fail at this point */
+                        tb = tb_alloc((unsigned long)pc);
+                        /* don't forget to invalidate previous TB info */
+                        ptb = &tb_hash[tb_hash_func((unsigned long)pc)];
+                        T0 = 0;
+                    }
+                    tc_ptr = code_gen_ptr;
+                    tb->tc_ptr = tc_ptr;
+                    tb->cs_base = (unsigned long)cs_base;
+                    tb->flags = flags;
+                    cpu_gen_code(env, tb, CODE_GEN_MAX_SIZE, &code_gen_size);
+                    *ptb = tb;
+                    tb->hash_next = NULL;
+                    tb_link(tb);
+                    code_gen_ptr = (void *)(((unsigned long)code_gen_ptr + code_gen_size + CODE_GEN_ALIGN - 1) & ~(CODE_GEN_ALIGN - 1));
                    spin_unlock(&tb_lock);
-                    raise_exception(EXCP06_ILLOP);
                }
-#endif
-                *ptb = tb;
-                tb->hash_next = NULL;
-                tb_link(tb);
-                code_gen_ptr = (void *)(((unsigned long)code_gen_ptr + code_gen_size + CODE_GEN_ALIGN - 1) & ~(CODE_GEN_ALIGN - 1));
-                spin_unlock(&tb_lock);
-            }
 #ifdef DEBUG_EXEC
-	    if (loglevel) {
-		fprintf(logfile, "Trace 0x%08lx [0x%08lx] %s\n",
-			(long)tb->tc_ptr, (long)tb->pc,
-			lookup_symbol((void *)tb->pc));
-	    }
+                if (loglevel) {
+                    fprintf(logfile, "Trace 0x%08lx [0x%08lx] %s\n",
+                            (long)tb->tc_ptr, (long)tb->pc,
+                            lookup_symbol((void *)tb->pc));
+                }
 #endif
 #ifdef __sparc__
-	    T0 = tmp_T0;
+                T0 = tmp_T0;
 #endif	    
-            /* see if we can patch the calling TB. XXX: remove TF test */
-            if (T0 != 0 
-#if defined(TARGET_I386)
-                && !(env->eflags & TF_MASK)
-#endif
-                ) {
-                spin_lock(&tb_lock);
-                tb_add_jump((TranslationBlock *)(T0 & ~3), T0 & 3, tb);
-                spin_unlock(&tb_lock);
-            }
-            tc_ptr = tb->tc_ptr;
-
-            /* execute the generated code */
-            gen_func = (void *)tc_ptr;
+                /* see if we can patch the calling TB. */
+                if (T0 != 0) {
+                    spin_lock(&tb_lock);
+                    tb_add_jump((TranslationBlock *)(T0 & ~3), T0 & 3, tb);
+                    spin_unlock(&tb_lock);
+                }
+                tc_ptr = tb->tc_ptr;
+                env->current_tb = tb;
+                /* execute the generated code */
+                gen_func = (void *)tc_ptr;
 #if defined(__sparc__)
-	    __asm__ __volatile__("call	%0\n\t"
-				 "mov	%%o7,%%i0"
-				 : /* no outputs */
-				 : "r" (gen_func) 
-				 : "i0", "i1", "i2", "i3", "i4", "i5");
+                __asm__ __volatile__("call	%0\n\t"
+                                     "mov	%%o7,%%i0"
+                                     : /* no outputs */
+                                     : "r" (gen_func) 
+                                     : "i0", "i1", "i2", "i3", "i4", "i5");
 #elif defined(__arm__)
-            asm volatile ("mov pc, %0\n\t"
-                          ".global exec_loop\n\t"
-                          "exec_loop:\n\t"
-                          : /* no outputs */
-                          : "r" (gen_func)
-                          : "r1", "r2", "r3", "r8", "r9", "r10", "r12", "r14");
+                asm volatile ("mov pc, %0\n\t"
+                              ".global exec_loop\n\t"
+                              "exec_loop:\n\t"
+                              : /* no outputs */
+                              : "r" (gen_func)
+                              : "r1", "r2", "r3", "r8", "r9", "r10", "r12", "r14");
 #else
-            gen_func();
+                gen_func();
 #endif
+                env->current_tb = NULL;
+                /* reset soft MMU for next block (it can currently
+                   only be set by a memory fault) */
+#if defined(TARGET_I386) && !defined(CONFIG_SOFTMMU)
+                if (env->hflags & HF_SOFTMMU_MASK) {
+                    env->hflags &= ~HF_SOFTMMU_MASK;
+                    /* do not allow linking to another block */
+                    T0 = 0;
+                }
+#endif
+            }
+        } else {
        }
-    }
-    ret = env->exception_index;
+    } /* for(;;) */
+

 #if defined(TARGET_I386)
    /* restore flags in standard format */
@@ -311,12 +357,7 @@ int cpu_exec(CPUState *env1)
    EDI = saved_EDI;
 #endif
 #elif defined(TARGET_ARM)
-    {
-        int ZF;
-        ZF = (env->NZF == 0);
-        env->cpsr = env->cpsr | (env->NZF & 0x80000000) | (ZF << 30) | 
-            (env->CF << 29) | ((env->VF & 0x80000000) >> 3);
-    }
+    env->cpsr = compute_cpsr();
 #else
 #error unsupported target CPU
 #endif
@@ -330,12 +371,6 @@ int cpu_exec(CPUState *env1)
    return ret;
 }

-void cpu_interrupt(CPUState *s)
-{
-    s->interrupt_request = 1;
-}
-
-
 #if defined(TARGET_I386)

 void cpu_x86_load_seg(CPUX86State *s, int seg_reg, int selector)
@@ -344,16 +379,10 @@ void cpu_x86_load_seg(CPUX86State *s, int seg_reg, int selector)

    saved_env = env;
    env = s;
-    if (env->eflags & VM_MASK) {
-        SegmentCache *sc;
+    if (!(env->cr[0] & CR0_PE_MASK) || (env->eflags & VM_MASK)) {
        selector &= 0xffff;
-        sc = &env->seg_cache[seg_reg];
-        /* NOTE: in VM86 mode, limit and seg_32bit are never reloaded,
-           so we must load them here */
-        sc->base = (void *)(selector << 4);
-        sc->limit = 0xffff;
-        sc->seg_32bit = 0;
-        env->segs[seg_reg] = selector;
+        cpu_x86_load_seg_cache(env, seg_reg, selector, 
+                               (uint8_t *)(selector << 4), 0xffff, 0);
    } else {
        load_seg(seg_reg, selector, 0);
    }
@@ -398,6 +427,8 @@ void cpu_x86_frstor(CPUX86State *s, uint8_t *ptr, int data32)
 #include <signal.h>
 #include <sys/ucontext.h>

+#if defined(TARGET_I386)
+
 /* 'pc' is the host PC at which the exception was raised. 'address' is
   the effective address of the memory exception. 'is_write' is 1 if a
   write caused the exception and otherwise 0'. 'old_set' is the
@@ -407,42 +438,59 @@ static inline int handle_cpu_signal(unsigned long pc, unsigned long address,
 {
    TranslationBlock *tb;
    int ret;
-    uint32_t found_pc;
-    
+
+    if (cpu_single_env)
+        env = cpu_single_env; /* XXX: find a correct solution for multithread */
 #if defined(DEBUG_SIGNAL)
-    printf("qemu: SIGSEGV pc=0x%08lx address=%08lx wr=%d oldset=0x%08lx\n", 
+    printf("qemu: SIGSEGV pc=0x%08lx address=%08lx w=%d oldset=0x%08lx\n", 
           pc, address, is_write, *(unsigned long *)old_set);
 #endif
    /* XXX: locking issue */
    if (is_write && page_unprotect(address)) {
        return 1;
    }
+    /* see if it is an MMU fault */
+    ret = cpu_x86_handle_mmu_fault(env, address, is_write);
+    if (ret < 0)
+        return 0; /* not an MMU fault */
+    if (ret == 0)
+        return 1; /* the MMU fault was handled without causing real CPU fault */
+    /* now we have a real cpu fault */
    tb = tb_find_pc(pc);
    if (tb) {
        /* the PC is inside the translated code. It means that we have
           a virtual CPU fault */
-        ret = cpu_search_pc(tb, &found_pc, pc);
-        if (ret < 0)
-            return 0;
-#if defined(TARGET_I386)
-        env->eip = found_pc - tb->cs_base;
-        env->cr2 = address;
+        cpu_restore_state(tb, env, pc);
+    }
+    if (ret == 1) {
+#if 0
+        printf("PF exception: EIP=0x%08x CR2=0x%08x error=0x%x\n", 
+               env->eip, env->cr[2], env->error_code);
+#endif
        /* we restore the process signal mask as the sigreturn should
           do it (XXX: use sigsetjmp) */
        sigprocmask(SIG_SETMASK, old_set, NULL);
-        raise_exception_err(EXCP0E_PAGE, 4 | (is_write << 1));
+        raise_exception_err(EXCP0E_PAGE, env->error_code);
+    } else {
+        /* activate soft MMU for this block */
+        env->hflags |= HF_SOFTMMU_MASK;
+        sigprocmask(SIG_SETMASK, old_set, NULL);
+        cpu_loop_exit();
+    }
+    /* never comes here */
+    return 1;
+}
+
 #elif defined(TARGET_ARM)
-        env->regs[15] = found_pc;
-        /* XXX: do more */
+static inline int handle_cpu_signal(unsigned long pc, unsigned long address,
+                                    int is_write, sigset_t *old_set)
+{
+    /* XXX: do more */
+    return 0;
+}
 #else
 #error unsupported target CPU
 #endif
-        /* never comes here */
-        return 1;
-    } else {
-        return 0;
-    }
-}

 #if defined(__i386__)

@@ -568,8 +616,25 @@ int cpu_signal_handler(int host_signum, struct siginfo *info,
                             &uc->uc_sigmask);
 }

+#elif defined(__mc68000)
+
+int cpu_signal_handler(int host_signum, struct siginfo *info, 
+                       void *puc)
+{
+    struct ucontext *uc = puc;
+    unsigned long pc;
+    int is_write;
+    
+    pc = uc->uc_mcontext.gregs[16];
+    /* XXX: compute is_write */
+    is_write = 0;
+    return handle_cpu_signal(pc, (unsigned long)info->si_addr, 
+                             is_write,
+                             &uc->uc_sigmask);
+}
+
 #else

-#error CPU specific signal handler needed
+#error host CPU specific signal handler needed

 #endif
--- a/cpu-i386.h
+++ b/cpu-i386.h
@@ -20,8 +20,7 @@
 #ifndef CPU_I386_H
 #define CPU_I386_H

-#include "config.h"
-#include <setjmp.h>
+#include "cpu-defs.h"

 #define R_EAX 0
 #define R_ECX 1
@@ -50,7 +49,8 @@

 /* segment descriptor fields */
 #define DESC_G_MASK     (1 << 23)
-#define DESC_B_MASK     (1 << 22)
+#define DESC_B_SHIFT    22
+#define DESC_B_MASK     (1 << DESC_B_SHIFT)
 #define DESC_AVL_MASK   (1 << 20)
 #define DESC_P_MASK     (1 << 15)
 #define DESC_DPL_SHIFT  13
@@ -73,6 +73,10 @@
 #define CC_S    0x0080
 #define CC_O    0x0800

+#define TF_SHIFT   8
+#define IOPL_SHIFT 12
+#define VM_SHIFT   17
+
 #define TF_MASK 		0x00000100
 #define IF_MASK 		0x00000200
 #define DF_MASK 		0x00000400
@@ -85,6 +89,77 @@
 #define VIP_MASK                0x00100000
 #define ID_MASK                 0x00200000

+/* hidden flags - used internally by qemu to represent additionnal cpu
+   states. Only the CPL and INHIBIT_IRQ are not redundant. We avoid
+   using the IOPL_MASK, TF_MASK and VM_MASK bit position to ease oring
+   with eflags. */
+/* current cpl */
+#define HF_CPL_SHIFT         0
+/* true if soft mmu is being used */
+#define HF_SOFTMMU_SHIFT     2
+/* true if hardware interrupts must be disabled for next instruction */
+#define HF_INHIBIT_IRQ_SHIFT 3
+/* 16 or 32 segments */
+#define HF_CS32_SHIFT        4
+#define HF_SS32_SHIFT        5
+/* zero base for DS, ES and SS */
+#define HF_ADDSEG_SHIFT      6
+
+#define HF_CPL_MASK          (3 << HF_CPL_SHIFT)
+#define HF_SOFTMMU_MASK      (1 << HF_SOFTMMU_SHIFT)
+#define HF_INHIBIT_IRQ_MASK  (1 << HF_INHIBIT_IRQ_SHIFT)
+#define HF_CS32_MASK         (1 << HF_CS32_SHIFT)
+#define HF_SS32_MASK         (1 << HF_SS32_SHIFT)
+#define HF_ADDSEG_MASK       (1 << HF_ADDSEG_SHIFT)
+
+#define CR0_PE_MASK  (1 << 0)
+#define CR0_TS_MASK  (1 << 3)
+#define CR0_WP_MASK  (1 << 16)
+#define CR0_AM_MASK  (1 << 18)
+#define CR0_PG_MASK  (1 << 31)
+
+#define CR4_VME_MASK  (1 << 0)
+#define CR4_PVI_MASK  (1 << 1)
+#define CR4_TSD_MASK  (1 << 2)
+#define CR4_DE_MASK   (1 << 3)
+#define CR4_PSE_MASK  (1 << 4)
+
+#define PG_PRESENT_BIT	0
+#define PG_RW_BIT	1
+#define PG_USER_BIT	2
+#define PG_PWT_BIT	3
+#define PG_PCD_BIT	4
+#define PG_ACCESSED_BIT	5
+#define PG_DIRTY_BIT	6
+#define PG_PSE_BIT	7
+#define PG_GLOBAL_BIT	8
+
+#define PG_PRESENT_MASK  (1 << PG_PRESENT_BIT)
+#define PG_RW_MASK	 (1 << PG_RW_BIT)
+#define PG_USER_MASK	 (1 << PG_USER_BIT)
+#define PG_PWT_MASK	 (1 << PG_PWT_BIT)
+#define PG_PCD_MASK	 (1 << PG_PCD_BIT)
+#define PG_ACCESSED_MASK (1 << PG_ACCESSED_BIT)
+#define PG_DIRTY_MASK	 (1 << PG_DIRTY_BIT)
+#define PG_PSE_MASK	 (1 << PG_PSE_BIT)
+#define PG_GLOBAL_MASK	 (1 << PG_GLOBAL_BIT)
+
+#define PG_ERROR_W_BIT     1
+
+#define PG_ERROR_P_MASK    0x01
+#define PG_ERROR_W_MASK    (1 << PG_ERROR_W_BIT)
+#define PG_ERROR_U_MASK    0x04
+#define PG_ERROR_RSVD_MASK 0x08
+
+#define MSR_IA32_APICBASE               0x1b
+#define MSR_IA32_APICBASE_BSP           (1<<8)
+#define MSR_IA32_APICBASE_ENABLE        (1<<11)
+#define MSR_IA32_APICBASE_BASE          (0xfffff<<12)
+
+#define MSR_IA32_SYSENTER_CS            0x174
+#define MSR_IA32_SYSENTER_ESP           0x175
+#define MSR_IA32_SYSENTER_EIP           0x176
+
 #define EXCP00_DIVZ	0
 #define EXCP01_SSTP	1
 #define EXCP02_NMI	2
@@ -104,8 +179,6 @@
 #define EXCP11_ALGN	17
 #define EXCP12_MCHK	18

-#define EXCP_INTERRUPT 	256 /* async interruption */
-
 enum {
    CC_OP_DYNAMIC, /* must use dynamic code to get cc_op */
    CC_OP_EFLAGS,  /* all cc are explicitely computed, CC_SRC = flags */
@@ -161,19 +234,12 @@ typedef double CPU86_LDouble;
 #endif

 typedef struct SegmentCache {
+    uint32_t selector;
    uint8_t *base;
-    unsigned long limit;
-    uint8_t seg_32bit;
+    uint32_t limit;
+    uint32_t flags;
 } SegmentCache;

-typedef struct SegmentDescriptorTable {
-    uint8_t *base;
-    unsigned long limit;
-    /* this is the returned base when reading the register, just to
-    avoid that the emulated program modifies it */
-    unsigned long emu_base;
-} SegmentDescriptorTable;
-
 typedef struct CPUX86State {
    /* standard registers */
    uint32_t regs[8];
@@ -187,6 +253,7 @@ typedef struct CPUX86State {
    uint32_t cc_dst;
    uint32_t cc_op;
    int32_t df; /* D flag : 1 if D = 0, -1 if D = 1 */
+    uint32_t hflags; /* hidden flags, see HF_xxx constants */

    /* FPU state */
    unsigned int fpstt; /* top of stack index */
@@ -205,18 +272,38 @@ typedef struct CPUX86State {
    } fp_convert;
    
    /* segments */
-    uint32_t segs[6]; /* selector values */
-    SegmentCache seg_cache[6]; /* info taken from LDT/GDT */
-    SegmentDescriptorTable gdt;
-    SegmentDescriptorTable ldt;
-    SegmentDescriptorTable idt;
+    SegmentCache segs[6]; /* selector values */
+    SegmentCache ldt;
+    SegmentCache tr;
+    SegmentCache gdt; /* only base and limit are used */
+    SegmentCache idt; /* only base and limit are used */
+    
+    /* sysenter registers */
+    uint32_t sysenter_cs;
+    uint32_t sysenter_esp;
+    uint32_t sysenter_eip;
    
    /* exception/interrupt handling */
    jmp_buf jmp_env;
    int exception_index;
    int error_code;
-    uint32_t cr2;
-    int interrupt_request;
+    int exception_is_int;
+    int exception_next_eip;
+    struct TranslationBlock *current_tb; /* currently executing TB */
+    uint32_t cr[5]; /* NOTE: cr1 is unused */
+    uint32_t dr[8]; /* debug registers */
+    int interrupt_request; 
+    int user_mode_only; /* user mode only simulation */
+
+    /* soft mmu support */
+    /* 0 = kernel, 1 = user */
+    CPUTLBEntry tlb_read[2][CPU_TLB_SIZE];
+    CPUTLBEntry tlb_write[2][CPU_TLB_SIZE];
+    
+    /* ice debug support */
+    uint32_t breakpoints[MAX_BREAKPOINTS];
+    int nb_breakpoints;
+    int singlestep_enabled;

    /* user data */
    void *opaque;
@@ -233,13 +320,60 @@ int cpu_x86_inl(CPUX86State *env, int addr);

 CPUX86State *cpu_x86_init(void);
 int cpu_x86_exec(CPUX86State *s);
-void cpu_x86_interrupt(CPUX86State *s);
 void cpu_x86_close(CPUX86State *s);
+int cpu_x86_get_pic_interrupt(CPUX86State *s);

-/* needed to load some predefinied segment registers */
+/* this function must always be used to load data in the segment
+   cache: it synchronizes the hflags with the segment cache values */
+static inline void cpu_x86_load_seg_cache(CPUX86State *env, 
+                                          int seg_reg, unsigned int selector,
+                                          uint8_t *base, unsigned int limit, 
+                                          unsigned int flags)
+{
+    SegmentCache *sc;
+    unsigned int new_hflags;
+    
+    sc = &env->segs[seg_reg];
+    sc->selector = selector;
+    sc->base = base;
+    sc->limit = limit;
+    sc->flags = flags;
+
+    /* update the hidden flags */
+    new_hflags = (env->segs[R_CS].flags & DESC_B_MASK)
+        >> (DESC_B_SHIFT - HF_CS32_SHIFT);
+    new_hflags |= (env->segs[R_SS].flags & DESC_B_MASK)
+        >> (DESC_B_SHIFT - HF_SS32_SHIFT);
+    if (!(env->cr[0] & CR0_PE_MASK) || (env->eflags & VM_MASK)) {
+        /* XXX: try to avoid this test. The problem comes from the
+           fact that is real mode or vm86 mode we only modify the
+           'base' and 'selector' fields of the segment cache to go
+           faster. A solution may be to force addseg to one in
+           translate-i386.c. */
+        new_hflags |= HF_ADDSEG_MASK;
+    } else {
+        new_hflags |= (((unsigned long)env->segs[R_DS].base | 
+                        (unsigned long)env->segs[R_ES].base |
+                        (unsigned long)env->segs[R_SS].base) != 0) << 
+            HF_ADDSEG_SHIFT;
+    }
+    env->hflags = (env->hflags & 
+                   ~(HF_CS32_MASK | HF_SS32_MASK | HF_ADDSEG_MASK)) | new_hflags;
+}
+
+/* wrapper, just in case memory mappings must be changed */
+static inline void cpu_x86_set_cpl(CPUX86State *s, int cpl)
+{
+#if HF_CPL_MASK == 3
+    s->hflags = (s->hflags & ~HF_CPL_MASK) | cpl;
+#else
+#error HF_CPL_MASK is hardcoded
+#endif
+}
+
+/* the following helpers are only usable in user mode simulation as
+   they can trigger unexpected exceptions */
 void cpu_x86_load_seg(CPUX86State *s, int seg_reg, int selector);
-
-/* simulate fsave/frstor */
 void cpu_x86_fsave(CPUX86State *s, uint8_t *ptr, int data32);
 void cpu_x86_frstor(CPUX86State *s, uint8_t *ptr, int data32);

@@ -250,6 +384,12 @@ struct siginfo;
 int cpu_x86_signal_handler(int host_signum, struct siginfo *info, 
                           void *puc);

+/* MMU defines */
+void cpu_x86_init_mmu(CPUX86State *env);
+extern int phys_ram_size;
+extern int phys_ram_fd;
+extern uint8_t *phys_ram_base;
+
 /* used to debug */
 #define X86_DUMP_FPU  0x0001 /* dump FPU state too */
 #define X86_DUMP_CCOP 0x0002 /* dump qemu flag cache */
--- a/disas.c
+++ b/disas.c
@@ -42,7 +42,7 @@ perror_memory (status, memaddr, info)
    /* Actually, address between memaddr and memaddr + len was
       out of bounds.  */
    (*info->fprintf_func) (info->stream,
-			   "Address 0x%x is out of bounds.\n", memaddr);
+			   "Address 0x%llx is out of bounds.\n", memaddr);
 }

 /* This could be in a separate file, to save miniscule amounts of space
@@ -57,7 +57,7 @@ generic_print_address (addr, info)
     bfd_vma addr;
     struct disassemble_info *info;
 {
-  (*info->fprintf_func) (info->stream, "0x%x", addr);
+  (*info->fprintf_func) (info->stream, "0x%llx", addr);
 }

 /* Just return the given address.  */
--- a/dyngen-exec.h
+++ b/dyngen-exec.h
@@ -109,6 +109,13 @@ extern int printf(const char *, ...);
 #define AREG5 "$13"
 #define AREG6 "$14"
 #endif
+#ifdef __mc68000
+#define AREG0 "%a5"
+#define AREG1 "%a4"
+#define AREG2 "%d7"
+#define AREG3 "%d6"
+#define AREG4 "%d5"
+#endif
 #ifdef __ia64__
 #define AREG0 "r27"
 #define AREG1 "r24"
@@ -125,6 +132,8 @@ extern int printf(const char *, ...);

 #define xglue(x, y) x ## y
 #define glue(x, y) xglue(x, y)
+#define stringify(s)	tostring(s)
+#define tostring(s)	#s

 #ifdef __alpha__
 /* the symbols are considered non exported so a br immediate is generated */
@@ -152,4 +161,30 @@ extern int __op_param1, __op_param2, __op_param3;
 #define PARAM3 ((long)(&__op_param3))
 #endif

-extern int __op_jmp0, __op_jmp1;
+extern int __op_jmp0, __op_jmp1, __op_jmp2, __op_jmp3;
+
+#ifdef __i386__
+#define EXIT_TB() asm volatile ("ret")
+#endif
+#ifdef __powerpc__
+#define EXIT_TB() asm volatile ("blr")
+#endif
+#ifdef __s390__
+#define EXIT_TB() asm volatile ("br %r14")
+#endif
+#ifdef __alpha__
+#define EXIT_TB() asm volatile ("ret")
+#endif
+#ifdef __ia64__
+#define EXIT_TB() asm volatile ("br.ret.sptk.many b0;;")
+#endif
+#ifdef __sparc__
+#define EXIT_TB() asm volatile ("jmpl %i0 + 8, %g0\n" \
+                                "nop")
+#endif
+#ifdef __arm__
+#define EXIT_TB() asm volatile ("b exec_loop")
+#endif
+#ifdef __mc68000
+#define EXIT_TB() asm volatile ("rts")
+#endif
--- a/dyngen.c
+++ b/dyngen.c
@@ -25,7 +25,7 @@
 #include <unistd.h>
 #include <fcntl.h>

-#include "config.h"
+#include "config-host.h"

 /* elf format definitions. We use these macros to test the CPU to
   allow cross compilation (this tool must be ran on the build
@@ -86,6 +86,13 @@
 #define elf_check_arch(x) ((x) == EM_ARM)
 #define ELF_USES_RELOC

+#elif defined(HOST_M68K)
+
+#define ELF_CLASS	ELFCLASS32
+#define ELF_ARCH	EM_68K
+#define elf_check_arch(x) ((x) == EM_68K)
+#define ELF_USES_RELOCA
+
 #else
 #error unsupported CPU - please update the code
 #endif
@@ -108,7 +115,7 @@ typedef uint64_t host_ulong;
 #define SHT_RELOC SHT_REL
 #endif

-#include "thunk.h"
+#include "bswap.h"

 enum {
    OUT_GEN_OP,
@@ -575,6 +582,21 @@ void gen_code(const char *name, host_ulong offset, host_ulong size,
                                      relocs, nb_relocs);
        break;
 #endif
+    case EM_68K:
+	{
+	    uint8_t *p;
+	    p = (void *)(p_end - 2);
+	    if (p == p_start)
+		error("empty code for %s", name);
+	    // remove NOP's, probably added for alignment
+	    while ((get16((uint16_t *)p) == 0x4e71) &&
+		   (p>p_start)) 
+	      p -= 2;
+	    if (get16((uint16_t *)p) != 0x4e75)
+		error("rts expected at the end of %s", name);
+	    copy_size = p - p_start;
+	}
+        break;
    default:
 	error("unknown ELF architecture");
    }
@@ -647,14 +669,13 @@ void gen_code(const char *name, host_ulong offset, host_ulong size,
        {
            ElfW(Sym) *sym;
            const char *sym_name, *p;
-            target_ulong val;
+            unsigned long val;
            int n;

            for(i = 0, sym = symtab; i < nb_syms; i++, sym++) {
                sym_name = strtab + sym->st_name;
                if (strstart(sym_name, "__op_label", &p)) {
                    uint8_t *ptr;
-                    int addend;
                    unsigned long offset;
                    
                    /* test if the variable refers to a label inside
@@ -663,7 +684,7 @@ void gen_code(const char *name, host_ulong offset, host_ulong size,
                    if (!ptr)
                        error("__op_labelN in invalid section");
                    offset = sym->st_value;
-                    addend = 0;
+                    val = *(unsigned long *)(ptr + offset);
 #ifdef ELF_USES_RELOCA
                    {
                        int reloc_shndx, nb_relocs1, j;
@@ -676,7 +697,7 @@ void gen_code(const char *name, host_ulong offset, host_ulong size,
                            rel = (ELF_RELOC *)sdata[reloc_shndx];
                            for(j = 0; j < nb_relocs1; j++) {
                                if (rel->r_offset == offset) {
-                                    addend = rel->r_addend;
+				    val = rel->r_addend;
                                    break;
                                }
 				rel++;
@@ -684,12 +705,10 @@ void gen_code(const char *name, host_ulong offset, host_ulong size,
                        }
                    }
 #endif                    
-                    val = *(target_ulong *)(ptr + offset);
-                    val += addend;

                    if (val >= start_offset && val < start_offset + copy_size) {
                        n = strtol(p, NULL, 10);
-                        fprintf(outfile, "    label_offsets[%d] = %d + (gen_code_ptr - gen_code_buf);\n", n, val - start_offset);
+                        fprintf(outfile, "    label_offsets[%d] = %ld + (gen_code_ptr - gen_code_buf);\n", n, val - start_offset);
                    }
                }
            }
@@ -1065,6 +1084,41 @@ void gen_code(const char *name, host_ulong offset, host_ulong size,
                }
                }
            }
+#elif defined(HOST_M68K)
+            {
+                char name[256];
+                int type;
+                int addend;
+		Elf32_Sym *sym;
+                for(i = 0, rel = relocs;i < nb_relocs; i++, rel++) {
+                if (rel->r_offset >= start_offset &&
+		    rel->r_offset < start_offset + copy_size) {
+		    sym = &(symtab[ELFW(R_SYM)(rel->r_info)]);
+                    sym_name = strtab + symtab[ELFW(R_SYM)(rel->r_info)].st_name;
+                    if (strstart(sym_name, "__op_param", &p)) {
+                        snprintf(name, sizeof(name), "param%s", p);
+                    } else {
+                        snprintf(name, sizeof(name), "(long)(&%s)", sym_name);
+                    }
+                    type = ELF32_R_TYPE(rel->r_info);
+                    addend = get32((uint32_t *)(text + rel->r_offset)) + rel->r_addend;
+                    switch(type) {
+                    case R_68K_32:
+		        fprintf(outfile, "    /* R_68K_32 RELOC, offset %x */\n", rel->r_offset) ;
+                        fprintf(outfile, "    *(uint32_t *)(gen_code_ptr + %d) = %s + %#x;\n", 
+                                rel->r_offset - start_offset, name, addend );
+                        break;
+                    case R_68K_PC32:
+		        fprintf(outfile, "    /* R_68K_PC32 RELOC, offset %x */\n", rel->r_offset);
+                        fprintf(outfile, "    *(uint32_t *)(gen_code_ptr + %d) = %s - (long)(gen_code_ptr + %#x) + %#x;\n", 
+                                rel->r_offset - start_offset, name, rel->r_offset - start_offset, /*sym->st_value+*/ addend);
+                        break;
+                    default:
+                        error("unsupported m68k relocation (%d)", type);
+                    }
+                }
+                }
+            }
 #else
 #error unsupported CPU
 #endif
@@ -1203,6 +1257,10 @@ int load_elf(const char *filename, FILE *outfile, int out_type)

    if (out_type == OUT_INDEX_OP) {
        fprintf(outfile, "DEF(end, 0, 0)\n");
+        fprintf(outfile, "DEF(nop, 0, 0)\n");
+        fprintf(outfile, "DEF(nop1, 1, 0)\n");
+        fprintf(outfile, "DEF(nop2, 2, 0)\n");
+        fprintf(outfile, "DEF(nop3, 3, 0)\n");
        for(i = 0, sym = symtab; i < nb_syms; i++, sym++) {
            const char *name, *p;
            name = strtab + sym->st_name;
@@ -1272,6 +1330,17 @@ fprintf(outfile,
        }

 fprintf(outfile,
+"        case INDEX_op_nop:\n"
+"            break;\n"
+"        case INDEX_op_nop1:\n"
+"            opparam_ptr++;\n"
+"            break;\n"
+"        case INDEX_op_nop2:\n"
+"            opparam_ptr += 2;\n"
+"            break;\n"
+"        case INDEX_op_nop3:\n"
+"            opparam_ptr += 3;\n"
+"            break;\n"
 "        default:\n"
 "            goto the_end;\n"
 "        }\n");
@@ -1293,38 +1362,10 @@ fprintf(outfile,
 " the_end:\n"
 );

-/* generate epilogue */ 
-    switch(ELF_ARCH) {
-    case EM_386:
-        fprintf(outfile, "*gen_code_ptr++ = 0xc3; /* ret */\n");
-        break;
-    case EM_PPC:
-        fprintf(outfile, "*((uint32_t *)gen_code_ptr)++ = 0x4e800020; /* blr */\n");
-        break;
-    case EM_S390:
-        fprintf(outfile, "*((uint16_t *)gen_code_ptr)++ = 0x07fe; /* br %%r14 */\n");
-        break;
-    case EM_ALPHA:
-        fprintf(outfile, "*((uint32_t *)gen_code_ptr)++ = 0x6bfa8001; /* ret */\n");
-        break;
-    case EM_IA_64:
-        fprintf(outfile, "*((uint32_t *)gen_code_ptr)++ = 0x00840008; /* br.ret.sptk.many b0;; */\n");
-        break;
-    case EM_SPARC:
-    case EM_SPARC32PLUS:
-	fprintf(outfile, "*((uint32_t *)gen_code_ptr)++ = 0x81c62008; /* jmpl %%i0 + 8, %%g0 */\n");
-	fprintf(outfile, "*((uint32_t *)gen_code_ptr)++ = 0x01000000; /* nop */\n");
-        break;
-    case EM_SPARCV9:
-	fprintf(outfile, "*((uint32_t *)gen_code_ptr)++ = 0x81c7e008; /* ret */\n");
-	fprintf(outfile, "*((uint32_t *)gen_code_ptr)++ = 0x81e80000; /* restore */\n");
-        break;
-    case EM_ARM:
-	fprintf(outfile, "gen_code_ptr = arm_flush_ldr(gen_code_ptr, arm_ldr_table, arm_ldr_ptr, arm_data_table, arm_data_ptr, 0);\n");
-        break;
-    default:
-	error("unknown ELF architecture");
-    }
+/* generate some code patching */ 
+#ifdef HOST_ARM
+fprintf(outfile, "gen_code_ptr = arm_flush_ldr(gen_code_ptr, arm_ldr_table, arm_ldr_ptr, arm_data_table, arm_data_ptr, 0);\n");
+#endif
    /* flush instruction cache */
    fprintf(outfile, "flush_icache_range((unsigned long)gen_code_buf, (unsigned long)gen_code_ptr);\n");

--- a/dyngen.h
+++ b/dyngen.h
@@ -19,7 +19,7 @@
 */

 int __op_param1, __op_param2, __op_param3;
-int __op_jmp0, __op_jmp1;
+int __op_jmp0, __op_jmp1, __op_jmp2, __op_jmp3;

 #ifdef __i386__
 static inline void flush_icache_range(unsigned long start, unsigned long stop)
@@ -94,6 +94,14 @@ static inline void flush_icache_range(unsigned long start, unsigned long stop)
 }
 #endif

+#ifdef __mc68000
+#include <asm/cachectl.h>
+static inline void flush_icache_range(unsigned long start, unsigned long stop)
+{
+    cacheflush(start,FLUSH_SCOPE_LINE,FLUSH_CACHE_BOTH,stop-start+16);
+}
+#endif
+
 #ifdef __alpha__

 register int gp asm("$29");
@@ -152,11 +160,7 @@ static uint8_t *arm_flush_ldr(uint8_t *gen_code_ptr,
 
    data_size = (uint8_t *)data_end - (uint8_t *)data_start;

-    if (!gen_jmp) {
-        /* b exec_loop */
-        arm_reloc_pc24((uint32_t *)gen_code_ptr, 0xeafffffe, (long)(&exec_loop)); 
-        gen_code_ptr += 4;
-    } else {
+    if (gen_jmp) {
        /* generate branch to skip the data */
        if (data_size == 0)
            return gen_code_ptr;
--- a/exec-arm.h
+++ b/exec-arm.h
@@ -30,3 +30,11 @@ register uint32_t T2 asm(AREG3);
 void cpu_lock(void);
 void cpu_unlock(void);
 void cpu_loop_exit(void);
+
+static inline int compute_cpsr(void)
+{
+    int ZF;
+    ZF = (env->NZF == 0);
+    return env->cpsr | (env->NZF & 0x80000000) | (ZF << 30) | 
+        (env->CF << 29) | ((env->VF & 0x80000000) >> 3);
+}
--- a/exec-i386.h
+++ b/exec-i386.h
@@ -123,8 +123,28 @@ typedef struct CCTable {
 extern CCTable cc_table[];

 void load_seg(int seg_reg, int selector, unsigned cur_eip);
+void helper_ljmp_protected_T0_T1(void);
+void helper_lcall_real_T0_T1(int shift, int next_eip);
+void helper_lcall_protected_T0_T1(int shift, int next_eip);
+void helper_iret_real(int shift);
+void helper_iret_protected(int shift);
+void helper_lret_protected(int shift, int addend);
+void helper_lldt_T0(void);
+void helper_ltr_T0(void);
+void helper_movl_crN_T0(int reg);
+void helper_movl_drN_T0(int reg);
+void helper_invlpg(unsigned int addr);
+void cpu_x86_update_cr0(CPUX86State *env);
+void cpu_x86_update_cr3(CPUX86State *env);
+void cpu_x86_flush_tlb(CPUX86State *env, uint32_t addr);
+int cpu_x86_handle_mmu_fault(CPUX86State *env, uint32_t addr, int is_write);
+void tlb_fill(unsigned long addr, int is_write, void *retaddr);
 void __hidden cpu_lock(void);
 void __hidden cpu_unlock(void);
+void do_interrupt(int intno, int is_int, int error_code, 
+                  unsigned int next_eip, int is_hw);
+void do_interrupt_user(int intno, int is_int, int error_code, 
+                       unsigned int next_eip);
 void raise_interrupt(int intno, int is_int, int error_code, 
                     unsigned int next_eip);
 void raise_exception_err(int exception_index, int error_code);
@@ -144,6 +164,8 @@ void helper_idivl_EAX_T0(uint32_t eip);
 void helper_cmpxchg8b(void);
 void helper_cpuid(void);
 void helper_rdtsc(void);
+void helper_rdmsr(void);
+void helper_wrmsr(void);
 void helper_lsl(void);
 void helper_lar(void);

@@ -324,3 +346,71 @@ void helper_frstor(uint8_t *ptr, int data32);
 const uint8_t parity_table[256];
 const uint8_t rclw_table[32];
 const uint8_t rclb_table[32];
+
+static inline uint32_t compute_eflags(void)
+{
+    return env->eflags | cc_table[CC_OP].compute_all() | (DF & DF_MASK);
+}
+
+#define FL_UPDATE_MASK32 (TF_MASK | AC_MASK | ID_MASK)
+
+#define FL_UPDATE_CPL0_MASK (TF_MASK | IF_MASK | IOPL_MASK | NT_MASK | \
+                             RF_MASK | AC_MASK | ID_MASK)
+
+/* NOTE: CC_OP must be modified manually to CC_OP_EFLAGS */
+static inline void load_eflags(int eflags, int update_mask)
+{
+    CC_SRC = eflags & (CC_O | CC_S | CC_Z | CC_A | CC_P | CC_C);
+    DF = 1 - (2 * ((eflags >> 10) & 1));
+    env->eflags = (env->eflags & ~update_mask) | 
+        (eflags & update_mask);
+}
+
+/* memory access macros */
+
+#define ldul ldl
+#define lduq ldq
+#define ldul_user ldl_user
+#define ldul_kernel ldl_kernel
+
+#define ldub_raw ldub
+#define ldsb_raw ldsb
+#define lduw_raw lduw
+#define ldsw_raw ldsw
+#define ldl_raw ldl
+#define ldq_raw ldq
+
+#define stb_raw stb
+#define stw_raw stw
+#define stl_raw stl
+#define stq_raw stq
+
+#define MEMUSER 0
+#define DATA_SIZE 1
+#include "softmmu_header.h"
+
+#define DATA_SIZE 2
+#include "softmmu_header.h"
+
+#define DATA_SIZE 4
+#include "softmmu_header.h"
+
+#define DATA_SIZE 8
+#include "softmmu_header.h"
+
+#undef MEMUSER
+#define MEMUSER 1
+#define DATA_SIZE 1
+#include "softmmu_header.h"
+
+#define DATA_SIZE 2
+#include "softmmu_header.h"
+
+#define DATA_SIZE 4
+#include "softmmu_header.h"
+
+#define DATA_SIZE 8
+#include "softmmu_header.h"
+
+#undef MEMUSER
+
--- a/exec.c
+++ b/exec.c
@@ -26,11 +26,17 @@
 #include <inttypes.h>
 #include <sys/mman.h>

+#include "config.h"
+#ifdef TARGET_I386
 #include "cpu-i386.h"
+#endif
+#ifdef TARGET_ARM
+#include "cpu-arm.h"
+#endif
 #include "exec.h"

 //#define DEBUG_TB_INVALIDATE
-#define DEBUG_FLUSH
+//#define DEBUG_FLUSH

 /* make various TB consistency checks */
 //#define DEBUG_TB_CHECK 
@@ -62,6 +68,7 @@ typedef struct PageDesc {
 #define L2_SIZE (1 << L2_BITS)

 static void tb_invalidate_page(unsigned long address);
+static void io_mem_init(void);

 unsigned long real_host_page_size;
 unsigned long host_page_bits;
@@ -70,6 +77,12 @@ unsigned long host_page_mask;

 static PageDesc *l1_map[L1_SIZE];

+/* io memory support */
+static unsigned long *l1_physmap[L1_SIZE];
+CPUWriteMemoryFunc *io_mem_write[IO_MEM_NB_ENTRIES][4];
+CPUReadMemoryFunc *io_mem_read[IO_MEM_NB_ENTRIES][4];
+static int io_mem_nb;
+
 static void page_init(void)
 {
    /* NOTE: we can always suppose that host_page_size >=
@@ -195,6 +208,7 @@ void cpu_exec_init(void)
    if (!code_gen_ptr) {
        code_gen_ptr = code_gen_buffer;
        page_init();
+        io_mem_init();
    }
 }

@@ -563,3 +577,308 @@ TranslationBlock *tb_find_pc(unsigned long tc_ptr)
    } 
    return &tbs[m_max];
 }
+
+static void tb_reset_jump_recursive(TranslationBlock *tb);
+
+static inline void tb_reset_jump_recursive2(TranslationBlock *tb, int n)
+{
+    TranslationBlock *tb1, *tb_next, **ptb;
+    unsigned int n1;
+
+    tb1 = tb->jmp_next[n];
+    if (tb1 != NULL) {
+        /* find head of list */
+        for(;;) {
+            n1 = (long)tb1 & 3;
+            tb1 = (TranslationBlock *)((long)tb1 & ~3);
+            if (n1 == 2)
+                break;
+            tb1 = tb1->jmp_next[n1];
+        }
+        /* we are now sure now that tb jumps to tb1 */
+        tb_next = tb1;
+
+        /* remove tb from the jmp_first list */
+        ptb = &tb_next->jmp_first;
+        for(;;) {
+            tb1 = *ptb;
+            n1 = (long)tb1 & 3;
+            tb1 = (TranslationBlock *)((long)tb1 & ~3);
+            if (n1 == n && tb1 == tb)
+                break;
+            ptb = &tb1->jmp_next[n1];
+        }
+        *ptb = tb->jmp_next[n];
+        tb->jmp_next[n] = NULL;
+        
+        /* suppress the jump to next tb in generated code */
+        tb_reset_jump(tb, n);
+
+        /* suppress jumps in the tb on which we could have jump */
+        tb_reset_jump_recursive(tb_next);
+    }
+}
+
+static void tb_reset_jump_recursive(TranslationBlock *tb)
+{
+    tb_reset_jump_recursive2(tb, 0);
+    tb_reset_jump_recursive2(tb, 1);
+}
+
+/* add a breakpoint. EXCP_DEBUG is returned by the CPU loop if a
+   breakpoint is reached */
+int cpu_breakpoint_insert(CPUState *env, uint32_t pc)
+{
+#if defined(TARGET_I386)
+    int i;
+
+    for(i = 0; i < env->nb_breakpoints; i++) {
+        if (env->breakpoints[i] == pc)
+            return 0;
+    }
+
+    if (env->nb_breakpoints >= MAX_BREAKPOINTS)
+        return -1;
+    env->breakpoints[env->nb_breakpoints++] = pc;
+    tb_invalidate_page(pc);
+    return 0;
+#else
+    return -1;
+#endif
+}
+
+/* remove a breakpoint */
+int cpu_breakpoint_remove(CPUState *env, uint32_t pc)
+{
+#if defined(TARGET_I386)
+    int i;
+    for(i = 0; i < env->nb_breakpoints; i++) {
+        if (env->breakpoints[i] == pc)
+            goto found;
+    }
+    return -1;
+ found:
+    memmove(&env->breakpoints[i], &env->breakpoints[i + 1],
+            (env->nb_breakpoints - (i + 1)) * sizeof(env->breakpoints[0]));
+    env->nb_breakpoints--;
+    tb_invalidate_page(pc);
+    return 0;
+#else
+    return -1;
+#endif
+}
+
+/* enable or disable single step mode. EXCP_DEBUG is returned by the
+   CPU loop after each instruction */
+void cpu_single_step(CPUState *env, int enabled)
+{
+#if defined(TARGET_I386)
+    if (env->singlestep_enabled != enabled) {
+        env->singlestep_enabled = enabled;
+        /* must flush all the translated code to avoid inconsistancies */
+        tb_flush();
+    }
+#endif
+}
+
+
+/* mask must never be zero */
+void cpu_interrupt(CPUState *env, int mask)
+{
+    TranslationBlock *tb;
+    
+    env->interrupt_request |= mask;
+    /* if the cpu is currently executing code, we must unlink it and
+       all the potentially executing TB */
+    tb = env->current_tb;
+    if (tb) {
+        tb_reset_jump_recursive(tb);
+    }
+}
+
+
+void cpu_abort(CPUState *env, const char *fmt, ...)
+{
+    va_list ap;
+
+    va_start(ap, fmt);
+    fprintf(stderr, "qemu: fatal: ");
+    vfprintf(stderr, fmt, ap);
+    fprintf(stderr, "\n");
+#ifdef TARGET_I386
+    cpu_x86_dump_state(env, stderr, X86_DUMP_FPU | X86_DUMP_CCOP);
+#endif
+    va_end(ap);
+    abort();
+}
+
+#ifdef TARGET_I386
+/* unmap all maped pages and flush all associated code */
+void page_unmap(void)
+{
+    PageDesc *p, *pmap;
+    unsigned long addr;
+    int i, j, ret, j1;
+
+    for(i = 0; i < L1_SIZE; i++) {
+        pmap = l1_map[i];
+        if (pmap) {
+            p = pmap;
+            for(j = 0;j < L2_SIZE;) {
+                if (p->flags & PAGE_VALID) {
+                    addr = (i << (32 - L1_BITS)) | (j << TARGET_PAGE_BITS);
+                    /* we try to find a range to make less syscalls */
+                    j1 = j;
+                    p++;
+                    j++;
+                    while (j < L2_SIZE && (p->flags & PAGE_VALID)) {
+                        p++;
+                        j++;
+                    }
+                    ret = munmap((void *)addr, (j - j1) << TARGET_PAGE_BITS);
+                    if (ret != 0) {
+                        fprintf(stderr, "Could not unmap page 0x%08lx\n", addr);
+                        exit(1);
+                    }
+                } else {
+                    p++;
+                    j++;
+                }
+            }
+            free(pmap);
+            l1_map[i] = NULL;
+        }
+    }
+    tb_flush();
+}
+#endif
+
+void tlb_flush(CPUState *env)
+{
+#if defined(TARGET_I386)
+    int i;
+    for(i = 0; i < CPU_TLB_SIZE; i++) {
+        env->tlb_read[0][i].address = -1;
+        env->tlb_write[0][i].address = -1;
+        env->tlb_read[1][i].address = -1;
+        env->tlb_write[1][i].address = -1;
+    }
+#endif
+}
+
+void tlb_flush_page(CPUState *env, uint32_t addr)
+{
+#if defined(TARGET_I386)
+    int i;
+
+    i = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1);
+    env->tlb_read[0][i].address = -1;
+    env->tlb_write[0][i].address = -1;
+    env->tlb_read[1][i].address = -1;
+    env->tlb_write[1][i].address = -1;
+#endif
+}
+
+static inline unsigned long *physpage_find_alloc(unsigned int page)
+{
+    unsigned long **lp, *p;
+    unsigned int index, i;
+
+    index = page >> TARGET_PAGE_BITS;
+    lp = &l1_physmap[index >> L2_BITS];
+    p = *lp;
+    if (!p) {
+        /* allocate if not found */
+        p = malloc(sizeof(unsigned long) * L2_SIZE);
+        for(i = 0; i < L2_SIZE; i++)
+            p[i] = IO_MEM_UNASSIGNED;
+        *lp = p;
+    }
+    return p + (index & (L2_SIZE - 1));
+}
+
+/* return NULL if no page defined (unused memory) */
+unsigned long physpage_find(unsigned long page)
+{
+    unsigned long *p;
+    unsigned int index;
+    index = page >> TARGET_PAGE_BITS;
+    p = l1_physmap[index >> L2_BITS];
+    if (!p)
+        return IO_MEM_UNASSIGNED;
+    return p[index & (L2_SIZE - 1)];
+}
+
+/* register physical memory. 'size' must be a multiple of the target
+   page size. If (phys_offset & ~TARGET_PAGE_MASK) != 0, then it is an
+   io memory page */
+void cpu_register_physical_memory(unsigned long start_addr, unsigned long size,
+                                  long phys_offset)
+{
+    unsigned long addr, end_addr;
+    unsigned long *p;
+
+    end_addr = start_addr + size;
+    for(addr = start_addr; addr < end_addr; addr += TARGET_PAGE_SIZE) {
+        p = physpage_find_alloc(addr);
+        *p = phys_offset;
+        if ((phys_offset & ~TARGET_PAGE_MASK) == 0)
+            phys_offset += TARGET_PAGE_SIZE;
+    }
+}
+
+static uint32_t unassigned_mem_readb(uint32_t addr)
+{
+    return 0;
+}
+
+static void unassigned_mem_writeb(uint32_t addr, uint32_t val)
+{
+}
+
+static CPUReadMemoryFunc *unassigned_mem_read[3] = {
+    unassigned_mem_readb,
+    unassigned_mem_readb,
+    unassigned_mem_readb,
+};
+
+static CPUWriteMemoryFunc *unassigned_mem_write[3] = {
+    unassigned_mem_writeb,
+    unassigned_mem_writeb,
+    unassigned_mem_writeb,
+};
+
+
+static void io_mem_init(void)
+{
+    io_mem_nb = 1;
+    cpu_register_io_memory(0, unassigned_mem_read, unassigned_mem_write);
+}
+
+/* mem_read and mem_write are arrays of functions containing the
+   function to access byte (index 0), word (index 1) and dword (index
+   2). All functions must be supplied. If io_index is non zero, the
+   corresponding io zone is modified. If it is zero, a new io zone is
+   allocated. The return value can be used with
+   cpu_register_physical_memory(). (-1) is returned if error. */
+int cpu_register_io_memory(int io_index,
+                           CPUReadMemoryFunc **mem_read,
+                           CPUWriteMemoryFunc **mem_write)
+{
+    int i;
+
+    if (io_index <= 0) {
+        if (io_index >= IO_MEM_NB_ENTRIES)
+            return -1;
+        io_index = io_mem_nb++;
+    } else {
+        if (io_index >= IO_MEM_NB_ENTRIES)
+            return -1;
+    }
+    
+    for(i = 0;i < 3; i++) {
+        io_mem_read[io_index][i] = mem_read[i];
+        io_mem_write[io_index][i] = mem_write[i];
+    }
+    return io_index << IO_MEM_SHIFT;
+}
--- a/exec.h
+++ b/exec.h
@@ -21,6 +21,23 @@
 /* allow to see translation results - the slowdown should be negligible, so we leave it */
 #define DEBUG_DISAS

+#ifndef glue
+#define xglue(x, y) x ## y
+#define glue(x, y) xglue(x, y)
+#define stringify(s)	tostring(s)
+#define tostring(s)	#s
+#endif
+
+#if GCC_MAJOR < 3
+#define __builtin_expect(x, n) (x)
+#endif
+
+#ifdef __i386__
+#define REGPARM(n) __attribute((regparm(n)))
+#else
+#define REGPARM(n)
+#endif
+
 /* is_jmp field values */
 #define DISAS_NEXT    0 /* next instruction can be analyzed */
 #define DISAS_JUMP    1 /* only pc was modified dynamically */
@@ -39,32 +56,30 @@ struct TranslationBlock;
 extern uint16_t gen_opc_buf[OPC_BUF_SIZE];
 extern uint32_t gen_opparam_buf[OPPARAM_BUF_SIZE];
 extern uint32_t gen_opc_pc[OPC_BUF_SIZE];
+extern uint8_t gen_opc_cc_op[OPC_BUF_SIZE];
 extern uint8_t gen_opc_instr_start[OPC_BUF_SIZE];

 #if defined(TARGET_I386)

-#define GEN_FLAG_CODE32_SHIFT 0
-#define GEN_FLAG_ADDSEG_SHIFT 1
-#define GEN_FLAG_SS32_SHIFT   2
-#define GEN_FLAG_VM_SHIFT     3
-#define GEN_FLAG_ST_SHIFT     4
-#define GEN_FLAG_TF_SHIFT     8 /* same position as eflags */
-#define GEN_FLAG_CPL_SHIFT    9
-#define GEN_FLAG_IOPL_SHIFT   12 /* same position as eflags */
+void optimize_flags_init(void);

 #endif

 extern FILE *logfile;
 extern int loglevel;

-int gen_intermediate_code(struct TranslationBlock *tb, int search_pc);
+int gen_intermediate_code(CPUState *env, struct TranslationBlock *tb);
+int gen_intermediate_code_pc(CPUState *env, struct TranslationBlock *tb);
 void dump_ops(const uint16_t *opc_buf, const uint32_t *opparam_buf);
-int cpu_gen_code(struct TranslationBlock *tb,
+int cpu_gen_code(CPUState *env, struct TranslationBlock *tb,
                 int max_code_size, int *gen_code_size_ptr);
-int cpu_search_pc(struct TranslationBlock *tb, 
-                  uint32_t *found_pc, unsigned long searched_pc);
+int cpu_restore_state(struct TranslationBlock *tb, 
+                      CPUState *env, unsigned long searched_pc);
 void cpu_exec_init(void);
 int page_unprotect(unsigned long address);
+void page_unmap(void);
+void tlb_flush_page(CPUState *env, uint32_t addr);
+void tlb_flush(CPUState *env);

 #define CODE_GEN_MAX_SIZE        65536
 #define CODE_GEN_ALIGN           16 /* must be >= of the size of a icache line */
@@ -93,7 +108,7 @@ typedef struct TranslationBlock {
       the code of this one. */
    uint16_t tb_next_offset[2]; /* offset of original jump target */
 #ifdef USE_DIRECT_JUMP
-    uint16_t tb_jmp_offset[2]; /* offset of jump instruction */
+    uint16_t tb_jmp_offset[4]; /* offset of jump instruction */
 #else
    uint32_t tb_next[2]; /* address of jump generated code */
 #endif
@@ -145,18 +160,14 @@ static inline TranslationBlock *tb_find(TranslationBlock ***pptb,

 #if defined(__powerpc__)

-static inline void tb_set_jmp_target(TranslationBlock *tb, 
-                                     int n, unsigned long addr)
+static inline void tb_set_jmp_target1(unsigned long jmp_addr, unsigned long addr)
 {
    uint32_t val, *ptr;
-    unsigned long offset;
-
-    offset = (unsigned long)(tb->tc_ptr + tb->tb_jmp_offset[n]);

    /* patch the branch destination */
-    ptr = (uint32_t *)offset;
+    ptr = (uint32_t *)jmp_addr;
    val = *ptr;
-    val = (val & ~0x03fffffc) | ((addr - offset) & 0x03fffffc);
+    val = (val & ~0x03fffffc) | ((addr - jmp_addr) & 0x03fffffc);
    *ptr = val;
    /* flush icache */
    asm volatile ("dcbst 0,%0" : : "r"(ptr) : "memory");
@@ -166,6 +177,18 @@ static inline void tb_set_jmp_target(TranslationBlock *tb,
    asm volatile ("isync" : : : "memory");
 }

+static inline void tb_set_jmp_target(TranslationBlock *tb, 
+                                     int n, unsigned long addr)
+{
+    unsigned long offset;
+
+    offset = tb->tb_jmp_offset[n];
+    tb_set_jmp_target1((unsigned long)(tb->tc_ptr + offset), addr);
+    offset = tb->tb_jmp_offset[n + 2];
+    if (offset != 0xffff)
+        tb_set_jmp_target1((unsigned long)(tb->tc_ptr + offset), addr);
+}
+
 #else

 /* set the jump target */
@@ -200,30 +223,59 @@ TranslationBlock *tb_find_pc(unsigned long pc_ptr);
 #if defined(__powerpc__)

 /* on PowerPC we patch the jump instruction directly */
-#define JUMP_TB(tbparam, n, eip)\
+#define JUMP_TB(opname, tbparam, n, eip)\
 do {\
-    static void __attribute__((unused)) *__op_label ## n = &&label ## n;\
-    asm volatile ("b %0" : : "i" (&__op_jmp ## n));\
-label ## n:\
+    asm volatile (".section \".data\"\n"\
+		  "__op_label" #n "." stringify(opname) ":\n"\
+		  ".long 1f\n"\
+		  ".previous\n"\
+                  "b __op_jmp" #n "\n"\
+		  "1:\n");\
    T0 = (long)(tbparam) + (n);\
    EIP = eip;\
+    EXIT_TB();\
+} while (0)
+
+#define JUMP_TB2(opname, tbparam, n)\
+do {\
+    asm volatile ("b __op_jmp%0\n" : : "i" (n + 2));\
 } while (0)

 #else

 /* jump to next block operations (more portable code, does not need
   cache flushing, but slower because of indirect jump) */
-#define JUMP_TB(tbparam, n, eip)\
+#define JUMP_TB(opname, tbparam, n, eip)\
 do {\
    static void __attribute__((unused)) *__op_label ## n = &&label ## n;\
+    static void __attribute__((unused)) *dummy ## n = &&dummy_label ## n;\
    goto *(void *)(((TranslationBlock *)tbparam)->tb_next[n]);\
 label ## n:\
    T0 = (long)(tbparam) + (n);\
    EIP = eip;\
+dummy_label ## n:\
+    EXIT_TB();\
+} while (0)
+
+/* second jump to same destination 'n' */
+#define JUMP_TB2(opname, tbparam, n)\
+do {\
+    goto *(void *)(((TranslationBlock *)tbparam)->tb_next[n]);\
 } while (0)

 #endif

+/* physical memory access */
+#define IO_MEM_NB_ENTRIES  256
+#define TLB_INVALID_MASK   (1 << 3)
+#define IO_MEM_SHIFT       4
+#define IO_MEM_UNASSIGNED  (1 << IO_MEM_SHIFT)
+
+unsigned long physpage_find(unsigned long page);
+
+extern CPUWriteMemoryFunc *io_mem_write[IO_MEM_NB_ENTRIES][4];
+extern CPUReadMemoryFunc *io_mem_read[IO_MEM_NB_ENTRIES][4];
+
 #ifdef __powerpc__
 static inline int testandset (int *p)
 {
@@ -315,10 +367,23 @@ static inline int testandset (int *spinlock)
 }
 #endif

+#ifdef __mc68000
+static inline int testandset (int *p)
+{
+    char ret;
+    __asm__ __volatile__("tas %1; sne %0"
+                         : "=r" (ret)
+                         : "m" (p)
+                         : "cc","memory");
+    return ret == 0;
+}
+#endif
+
 typedef int spinlock_t;

 #define SPIN_LOCK_UNLOCKED 0

+#if 1
 static inline void spin_lock(spinlock_t *lock)
 {
    while (testandset(lock));
@@ -333,6 +398,20 @@ static inline int spin_trylock(spinlock_t *lock)
 {
    return !testandset(lock);
 }
+#else
+static inline void spin_lock(spinlock_t *lock)
+{
+}
+
+static inline void spin_unlock(spinlock_t *lock)
+{
+}
+
+static inline int spin_trylock(spinlock_t *lock)
+{
+    return 1;
+}
+#endif

 extern spinlock_t tb_lock;

--- a/gdbstub.c
+++ b/gdbstub.c
@@ -0,0 +1,453 @@
+/*
+ * gdb server stub
+ * 
+ * Copyright (c) 2003 Fabrice Bellard
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+#include <unistd.h>
+#include <errno.h>
+#include <sys/socket.h>
+#include <netinet/in.h>
+#include <netinet/tcp.h>
+#include <signal.h>
+
+#include "config.h"
+#ifdef TARGET_I386
+#include "cpu-i386.h"
+#endif
+#ifdef TARGET_ARM
+#include "cpu-arm.h"
+#endif
+#include "thunk.h"
+#include "exec.h"
+
+//#define DEBUG_GDB
+
+int gdbstub_fd = -1;
+
+/* return 0 if OK */
+static int gdbstub_open(int port)
+{
+    struct sockaddr_in sockaddr;
+    socklen_t len;
+    int fd, val, ret;
+
+    fd = socket(PF_INET, SOCK_STREAM, 0);
+    if (fd < 0) {
+        perror("socket");
+        return -1;
+    }
+
+    /* allow fast reuse */
+    val = 1;
+    setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, &val, sizeof(val));
+
+    sockaddr.sin_family = AF_INET;
+    sockaddr.sin_port = htons(port);
+    sockaddr.sin_addr.s_addr = 0;
+    ret = bind(fd, (struct sockaddr *)&sockaddr, sizeof(sockaddr));
+    if (ret < 0) {
+        perror("bind");
+        return -1;
+    }
+    ret = listen(fd, 0);
+    if (ret < 0) {
+        perror("listen");
+        return -1;
+    }
+    
+    /* now wait for one connection */
+    for(;;) {
+        len = sizeof(sockaddr);
+        gdbstub_fd = accept(fd, (struct sockaddr *)&sockaddr, &len);
+        if (gdbstub_fd < 0 && errno != EINTR) {
+            perror("accept");
+            return -1;
+        } else if (gdbstub_fd >= 0) {
+            break;
+        }
+    }
+    
+    /* set short latency */
+    val = 1;
+    setsockopt(gdbstub_fd, SOL_TCP, TCP_NODELAY, &val, sizeof(val));
+    return 0;
+}
+
+static int get_char(void)
+{
+    uint8_t ch;
+    int ret;
+
+    for(;;) {
+        ret = read(gdbstub_fd, &ch, 1);
+        if (ret < 0) {
+            if (errno != EINTR && errno != EAGAIN)
+                return -1;
+        } else if (ret == 0) {
+            return -1;
+        } else {
+            break;
+        }
+    }
+    return ch;
+}
+
+static void put_buffer(const uint8_t *buf, int len)
+{
+    int ret;
+
+    while (len > 0) {
+        ret = write(gdbstub_fd, buf, len);
+        if (ret < 0) {
+            if (errno != EINTR && errno != EAGAIN)
+                return;
+        } else {
+            buf += ret;
+            len -= ret;
+        }
+    }
+}
+
+static inline int fromhex(int v)
+{
+    if (v >= '0' && v <= '9')
+        return v - '0';
+    else if (v >= 'A' && v <= 'F')
+        return v - 'A' + 10;
+    else if (v >= 'a' && v <= 'f')
+        return v - 'a' + 10;
+    else
+        return 0;
+}
+
+static inline int tohex(int v)
+{
+    if (v < 10)
+        return v + '0';
+    else
+        return v - 10 + 'a';
+}
+
+static void memtohex(char *buf, const uint8_t *mem, int len)
+{
+    int i, c;
+    char *q;
+    q = buf;
+    for(i = 0; i < len; i++) {
+        c = mem[i];
+        *q++ = tohex(c >> 4);
+        *q++ = tohex(c & 0xf);
+    }
+    *q = '\0';
+}
+
+static void hextomem(uint8_t *mem, const char *buf, int len)
+{
+    int i;
+
+    for(i = 0; i < len; i++) {
+        mem[i] = (fromhex(buf[0]) << 4) | fromhex(buf[1]);
+        buf += 2;
+    }
+}
+
+/* return -1 if error or EOF */
+static int get_packet(char *buf, int buf_size)
+{
+    int ch, len, csum, csum1;
+    char reply[1];
+    
+    for(;;) {
+        for(;;) {
+            ch = get_char();
+            if (ch < 0)
+                return -1;
+            if (ch == '$')
+                break;
+        }
+        len = 0;
+        csum = 0;
+        for(;;) {
+            ch = get_char();
+            if (ch < 0)
+                return -1;
+            if (ch == '#')
+                break;
+            if (len > buf_size - 1)
+                return -1;
+            buf[len++] = ch;
+            csum += ch;
+        }
+        buf[len] = '\0';
+        ch = get_char();
+        if (ch < 0)
+            return -1;
+        csum1 = fromhex(ch) << 4;
+        ch = get_char();
+        if (ch < 0)
+            return -1;
+        csum1 |= fromhex(ch);
+        if ((csum & 0xff) != csum1) {
+            reply[0] = '-';
+            put_buffer(reply, 1);
+        } else {
+            reply[0] = '+';
+            put_buffer(reply, 1);
+            break;
+        }
+    }
+#ifdef DEBUG_GDB
+    printf("command='%s'\n", buf);
+#endif
+    return len;
+}
+
+/* return -1 if error, 0 if OK */
+static int put_packet(char *buf)
+{
+    char buf1[3];
+    int len, csum, ch, i;
+
+#ifdef DEBUG_GDB
+    printf("reply='%s'\n", buf);
+#endif
+
+    for(;;) {
+        buf1[0] = '$';
+        put_buffer(buf1, 1);
+        len = strlen(buf);
+        put_buffer(buf, len);
+        csum = 0;
+        for(i = 0; i < len; i++) {
+            csum += buf[i];
+        }
+        buf1[0] = '#';
+        buf1[1] = tohex((csum >> 4) & 0xf);
+        buf1[2] = tohex((csum) & 0xf);
+
+        put_buffer(buf1, 3);
+
+        ch = get_char();
+        if (ch < 0)
+            return -1;
+        if (ch == '+')
+            break;
+    }
+    return 0;
+}
+
+static int memory_rw(uint8_t *buf, uint32_t addr, int len, int is_write)
+{
+    int l, flags;
+    uint32_t page;
+
+    while (len > 0) {
+        page = addr & TARGET_PAGE_MASK;
+        l = (page + TARGET_PAGE_SIZE) - addr;
+        if (l > len)
+            l = len;
+        flags = page_get_flags(page);
+        if (!(flags & PAGE_VALID))
+            return -1;
+        if (is_write) {
+            if (!(flags & PAGE_WRITE))
+                return -1;
+            memcpy((uint8_t *)addr, buf, l);
+        } else {
+            if (!(flags & PAGE_READ))
+                return -1;
+            memcpy(buf, (uint8_t *)addr, l);
+        }
+        len -= l;
+        buf += l;
+        addr += l;
+    }
+    return 0;
+}
+
+/* port = 0 means default port */
+int cpu_gdbstub(void *opaque, int (*main_loop)(void *opaque), int port)
+{
+    CPUState *env;
+    const char *p;
+    int ret, ch, nb_regs, i, type;
+    char buf[4096];
+    uint8_t mem_buf[2000];
+    uint32_t *registers;
+    uint32_t addr, len;
+    
+    printf("Waiting gdb connection on port %d\n", port);
+    if (gdbstub_open(port) < 0)
+        return -1;
+    printf("Connected\n");
+    for(;;) {
+        ret = get_packet(buf, sizeof(buf));
+        if (ret < 0)
+            break;
+        p = buf;
+        ch = *p++;
+        switch(ch) {
+        case '?':
+            snprintf(buf, sizeof(buf), "S%02x", SIGTRAP);
+            put_packet(buf);
+            break;
+        case 'c':
+            if (*p != '\0') {
+                addr = strtoul(p, (char **)&p, 16);
+                env = cpu_gdbstub_get_env(opaque);
+#if defined(TARGET_I386)
+                env->eip = addr;
+#endif
+            }
+            ret = main_loop(opaque);
+            if (ret == EXCP_DEBUG)
+                ret = SIGTRAP;
+            else
+                ret = 0;
+            snprintf(buf, sizeof(buf), "S%02x", ret);
+            put_packet(buf);
+            break;
+        case 's':
+            env = cpu_gdbstub_get_env(opaque);
+            if (*p != '\0') {
+                addr = strtoul(p, (char **)&p, 16);
+#if defined(TARGET_I386)
+                env->eip = addr;
+#endif
+            }
+            cpu_single_step(env, 1);
+            ret = main_loop(opaque);
+            cpu_single_step(env, 0);
+            if (ret == EXCP_DEBUG)
+                ret = SIGTRAP;
+            else
+                ret = 0;
+            snprintf(buf, sizeof(buf), "S%02x", ret);
+            put_packet(buf);
+            break;
+        case 'g':
+            env = cpu_gdbstub_get_env(opaque);
+            registers = (void *)mem_buf;
+#if defined(TARGET_I386)
+            for(i = 0; i < 8; i++) {
+                registers[i] = tswapl(env->regs[i]);
+            }
+            registers[8] = env->eip;
+            registers[9] = env->eflags;
+            registers[10] = env->segs[R_CS].selector;
+            registers[11] = env->segs[R_SS].selector;
+            registers[12] = env->segs[R_DS].selector;
+            registers[13] = env->segs[R_ES].selector;
+            registers[14] = env->segs[R_FS].selector;
+            registers[15] = env->segs[R_GS].selector;
+            nb_regs = 16;
+#endif
+            memtohex(buf, (const uint8_t *)registers, 
+                     sizeof(registers[0]) * nb_regs);
+            put_packet(buf);
+            break;
+        case 'G':
+            env = cpu_gdbstub_get_env(opaque);
+            registers = (void *)mem_buf;
+#if defined(TARGET_I386)
+            hextomem((uint8_t *)registers, p, 16 * 4);
+            for(i = 0; i < 8; i++) {
+                env->regs[i] = tswapl(registers[i]);
+            }
+            env->eip = registers[8];
+            env->eflags = registers[9];
+#define LOAD_SEG(index, sreg)\
+            if (tswapl(registers[index]) != env->segs[sreg].selector)\
+                cpu_x86_load_seg(env, sreg, tswapl(registers[index]));
+            LOAD_SEG(10, R_CS);
+            LOAD_SEG(11, R_SS);
+            LOAD_SEG(12, R_DS);
+            LOAD_SEG(13, R_ES);
+            LOAD_SEG(14, R_FS);
+            LOAD_SEG(15, R_GS);
+#endif
+            put_packet("OK");
+            break;
+        case 'm':
+            addr = strtoul(p, (char **)&p, 16);
+            if (*p == ',')
+                p++;
+            len = strtoul(p, NULL, 16);
+            if (memory_rw(mem_buf, addr, len, 0) != 0)
+                memset(mem_buf, 0, len);
+            memtohex(buf, mem_buf, len);
+            put_packet(buf);
+            break;
+        case 'M':
+            addr = strtoul(p, (char **)&p, 16);
+            if (*p == ',')
+                p++;
+            len = strtoul(p, (char **)&p, 16);
+            if (*p == ',')
+                p++;
+            hextomem(mem_buf, p, len);
+            if (memory_rw(mem_buf, addr, len, 1) != 0)
+                put_packet("ENN");
+            else
+                put_packet("OK");
+            break;
+        case 'Z':
+            type = strtoul(p, (char **)&p, 16);
+            if (*p == ',')
+                p++;
+            addr = strtoul(p, (char **)&p, 16);
+            if (*p == ',')
+                p++;
+            len = strtoul(p, (char **)&p, 16);
+            if (type == 0 || type == 1) {
+                env = cpu_gdbstub_get_env(opaque);
+                if (cpu_breakpoint_insert(env, addr) < 0)
+                    goto breakpoint_error;
+                put_packet("OK");
+            } else {
+            breakpoint_error:
+                put_packet("ENN");
+            }
+            break;
+        case 'z':
+            type = strtoul(p, (char **)&p, 16);
+            if (*p == ',')
+                p++;
+            addr = strtoul(p, (char **)&p, 16);
+            if (*p == ',')
+                p++;
+            len = strtoul(p, (char **)&p, 16);
+            if (type == 0 || type == 1) {
+                env = cpu_gdbstub_get_env(opaque);
+                cpu_breakpoint_remove(env, addr);
+                put_packet("OK");
+            } else {
+                goto breakpoint_error;
+            }
+            break;
+        default:
+            /* put empty packet */
+            buf[0] = '\0';
+            put_packet(buf);
+            break;
+        }
+    }
+    return 0;
+}
--- a/helper-i386.c
+++ b/helper-i386.c
--- a/helper2-i386.c
+++ b/helper2-i386.c
@@ -0,0 +1,390 @@
+/*
+ *  i386 helpers (without register variable usage)
+ * 
+ *  Copyright (c) 2003 Fabrice Bellard
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+#include <stdarg.h>
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+#include <inttypes.h>
+#include <signal.h>
+#include <assert.h>
+#include <sys/mman.h>
+
+#include "cpu-i386.h"
+#include "exec.h"
+
+//#define DEBUG_MMU
+
+CPUX86State *cpu_x86_init(void)
+{
+    CPUX86State *env;
+    int i;
+    static int inited;
+
+    cpu_exec_init();
+
+    env = malloc(sizeof(CPUX86State));
+    if (!env)
+        return NULL;
+    memset(env, 0, sizeof(CPUX86State));
+    /* basic FPU init */
+    for(i = 0;i < 8; i++)
+        env->fptags[i] = 1;
+    env->fpuc = 0x37f;
+    /* flags setup : we activate the IRQs by default as in user mode */
+    env->eflags = 0x2 | IF_MASK;
+
+    tlb_flush(env);
+#ifdef CONFIG_SOFTMMU
+    env->hflags |= HF_SOFTMMU_MASK;
+#endif
+    /* init various static tables */
+    if (!inited) {
+        inited = 1;
+        optimize_flags_init();
+    }
+    return env;
+}
+
+void cpu_x86_close(CPUX86State *env)
+{
+    free(env);
+}
+
+/***********************************************************/
+/* x86 debug */
+
+static const char *cc_op_str[] = {
+    "DYNAMIC",
+    "EFLAGS",
+    "MUL",
+    "ADDB",
+    "ADDW",
+    "ADDL",
+    "ADCB",
+    "ADCW",
+    "ADCL",
+    "SUBB",
+    "SUBW",
+    "SUBL",
+    "SBBB",
+    "SBBW",
+    "SBBL",
+    "LOGICB",
+    "LOGICW",
+    "LOGICL",
+    "INCB",
+    "INCW",
+    "INCL",
+    "DECB",
+    "DECW",
+    "DECL",
+    "SHLB",
+    "SHLW",
+    "SHLL",
+    "SARB",
+    "SARW",
+    "SARL",
+};
+
+void cpu_x86_dump_state(CPUX86State *env, FILE *f, int flags)
+{
+    int eflags;
+    char cc_op_name[32];
+
+    eflags = env->eflags;
+    fprintf(f, "EAX=%08x EBX=%08x ECX=%08x EDX=%08x\n"
+            "ESI=%08x EDI=%08x EBP=%08x ESP=%08x\n"
+            "EIP=%08x EFL=%08x [%c%c%c%c%c%c%c]\n",
+            env->regs[R_EAX], env->regs[R_EBX], env->regs[R_ECX], env->regs[R_EDX], 
+            env->regs[R_ESI], env->regs[R_EDI], env->regs[R_EBP], env->regs[R_ESP], 
+            env->eip, eflags,
+            eflags & DF_MASK ? 'D' : '-',
+            eflags & CC_O ? 'O' : '-',
+            eflags & CC_S ? 'S' : '-',
+            eflags & CC_Z ? 'Z' : '-',
+            eflags & CC_A ? 'A' : '-',
+            eflags & CC_P ? 'P' : '-',
+            eflags & CC_C ? 'C' : '-');
+    fprintf(f, "CS=%04x SS=%04x DS=%04x ES=%04x FS=%04x GS=%04x\n",
+            env->segs[R_CS].selector,
+            env->segs[R_SS].selector,
+            env->segs[R_DS].selector,
+            env->segs[R_ES].selector,
+            env->segs[R_FS].selector,
+            env->segs[R_GS].selector);
+    if (flags & X86_DUMP_CCOP) {
+        if ((unsigned)env->cc_op < CC_OP_NB)
+            strcpy(cc_op_name, cc_op_str[env->cc_op]);
+        else
+            snprintf(cc_op_name, sizeof(cc_op_name), "[%d]", env->cc_op);
+        fprintf(f, "CCS=%08x CCD=%08x CCO=%-8s\n",
+                env->cc_src, env->cc_dst, cc_op_name);
+    }
+    if (flags & X86_DUMP_FPU) {
+        fprintf(f, "ST0=%f ST1=%f ST2=%f ST3=%f\n", 
+                (double)env->fpregs[0], 
+                (double)env->fpregs[1], 
+                (double)env->fpregs[2], 
+                (double)env->fpregs[3]);
+        fprintf(f, "ST4=%f ST5=%f ST6=%f ST7=%f\n", 
+                (double)env->fpregs[4], 
+                (double)env->fpregs[5], 
+                (double)env->fpregs[7], 
+                (double)env->fpregs[8]);
+    }
+}
+
+/***********************************************************/
+/* x86 mmu */
+/* XXX: add PGE support */
+
+/* called when cr3 or PG bit are modified */
+static int last_pg_state = -1;
+static int last_pe_state = 0;
+int phys_ram_size;
+int phys_ram_fd;
+uint8_t *phys_ram_base;
+
+void cpu_x86_update_cr0(CPUX86State *env)
+{
+    int pg_state, pe_state;
+
+#ifdef DEBUG_MMU
+    printf("CR0 update: CR0=0x%08x\n", env->cr[0]);
+#endif
+    pg_state = env->cr[0] & CR0_PG_MASK;
+    if (pg_state != last_pg_state) {
+        page_unmap();
+        tlb_flush(env);
+        last_pg_state = pg_state;
+    }
+    pe_state = env->cr[0] & CR0_PE_MASK;
+    if (last_pe_state != pe_state) {
+        tb_flush();
+        last_pe_state = pe_state;
+    }
+}
+
+void cpu_x86_update_cr3(CPUX86State *env)
+{
+    if (env->cr[0] & CR0_PG_MASK) {
+#if defined(DEBUG_MMU)
+        printf("CR3 update: CR3=%08x\n", env->cr[3]);
+#endif
+        page_unmap();
+        tlb_flush(env);
+    }
+}
+
+void cpu_x86_init_mmu(CPUX86State *env)
+{
+    last_pg_state = -1;
+    cpu_x86_update_cr0(env);
+}
+
+/* XXX: also flush 4MB pages */
+void cpu_x86_flush_tlb(CPUX86State *env, uint32_t addr)
+{
+    int flags;
+    unsigned long virt_addr;
+
+    tlb_flush_page(env, addr);
+
+    flags = page_get_flags(addr);
+    if (flags & PAGE_VALID) {
+        virt_addr = addr & ~0xfff;
+        munmap((void *)virt_addr, 4096);
+        page_set_flags(virt_addr, virt_addr + 4096, 0);
+    }
+}
+
+/* return value:
+   -1 = cannot handle fault 
+   0  = nothing more to do 
+   1  = generate PF fault
+   2  = soft MMU activation required for this block
+*/
+int cpu_x86_handle_mmu_fault(CPUX86State *env, uint32_t addr, int is_write)
+{
+    uint8_t *pde_ptr, *pte_ptr;
+    uint32_t pde, pte, virt_addr;
+    int cpl, error_code, is_dirty, is_user, prot, page_size, ret;
+    unsigned long pd;
+    
+    cpl = env->hflags & HF_CPL_MASK;
+    is_user = (cpl == 3);
+    
+#ifdef DEBUG_MMU
+    printf("MMU fault: addr=0x%08x w=%d u=%d eip=%08x\n", 
+           addr, is_write, is_user, env->eip);
+#endif
+
+    if (env->user_mode_only) {
+        /* user mode only emulation */
+        error_code = 0;
+        goto do_fault;
+    }
+
+    if (!(env->cr[0] & CR0_PG_MASK)) {
+        pte = addr;
+        virt_addr = addr & ~0xfff;
+        prot = PROT_READ | PROT_WRITE;
+        page_size = 4096;
+        goto do_mapping;
+    }
+
+    /* page directory entry */
+    pde_ptr = phys_ram_base + ((env->cr[3] & ~0xfff) + ((addr >> 20) & ~3));
+    pde = ldl(pde_ptr);
+    if (!(pde & PG_PRESENT_MASK)) {
+        error_code = 0;
+        goto do_fault;
+    }
+    if (is_user) {
+        if (!(pde & PG_USER_MASK))
+            goto do_fault_protect;
+        if (is_write && !(pde & PG_RW_MASK))
+            goto do_fault_protect;
+    } else {
+        if ((env->cr[0] & CR0_WP_MASK) && (pde & PG_USER_MASK) &&
+            is_write && !(pde & PG_RW_MASK)) 
+            goto do_fault_protect;
+    }
+    /* if PSE bit is set, then we use a 4MB page */
+    if ((pde & PG_PSE_MASK) && (env->cr[4] & CR4_PSE_MASK)) {
+        is_dirty = is_write && !(pde & PG_DIRTY_MASK);
+        if (!(pde & PG_ACCESSED_MASK)) {
+            pde |= PG_ACCESSED_MASK;
+            if (is_dirty)
+                pde |= PG_DIRTY_MASK;
+            stl(pde_ptr, pde);
+        }
+        
+        pte = pde & ~0x003ff000; /* align to 4MB */
+        page_size = 4096 * 1024;
+        virt_addr = addr & ~0x003fffff;
+    } else {
+        if (!(pde & PG_ACCESSED_MASK)) {
+            pde |= PG_ACCESSED_MASK;
+            stl(pde_ptr, pde);
+        }
+
+        /* page directory entry */
+        pte_ptr = phys_ram_base + ((pde & ~0xfff) + ((addr >> 10) & 0xffc));
+        pte = ldl(pte_ptr);
+        if (!(pte & PG_PRESENT_MASK)) {
+            error_code = 0;
+            goto do_fault;
+        }
+        if (is_user) {
+            if (!(pte & PG_USER_MASK))
+                goto do_fault_protect;
+            if (is_write && !(pte & PG_RW_MASK))
+                goto do_fault_protect;
+        } else {
+            if ((env->cr[0] & CR0_WP_MASK) && (pte & PG_USER_MASK) &&
+                is_write && !(pte & PG_RW_MASK)) 
+                goto do_fault_protect;
+        }
+        is_dirty = is_write && !(pte & PG_DIRTY_MASK);
+        if (!(pte & PG_ACCESSED_MASK) || is_dirty) {
+            pte |= PG_ACCESSED_MASK;
+            if (is_dirty)
+                pte |= PG_DIRTY_MASK;
+            stl(pte_ptr, pte);
+        }
+        page_size = 4096;
+        virt_addr = addr & ~0xfff;
+    }
+    /* the page can be put in the TLB */
+    prot = PROT_READ;
+    if (is_user) {
+        if (pte & PG_RW_MASK)
+            prot |= PROT_WRITE;
+    } else {
+        if (!(env->cr[0] & CR0_WP_MASK) || !(pte & PG_USER_MASK) ||
+            (pte & PG_RW_MASK))
+            prot |= PROT_WRITE;
+    }
+    
+ do_mapping:
+    if (env->hflags & HF_SOFTMMU_MASK) {
+        unsigned long paddr, vaddr, address, addend, page_offset;
+        int index;
+
+        /* software MMU case. Even if 4MB pages, we map only one 4KB
+           page in the cache to avoid filling it too fast */
+        page_offset = (addr & ~0xfff) & (page_size - 1);
+        paddr = (pte & ~0xfff) + page_offset;
+        vaddr = virt_addr + page_offset;
+        index = (addr >> 12) & (CPU_TLB_SIZE - 1);
+        pd = physpage_find(paddr);
+        if (pd & 0xfff) {
+            /* IO memory case */
+            address = vaddr | pd;
+            addend = paddr;
+        } else {
+            /* standard memory */
+            address = vaddr;
+            addend = (unsigned long)phys_ram_base + pd;
+        }
+        addend -= vaddr;
+        env->tlb_read[is_user][index].address = address;
+        env->tlb_read[is_user][index].addend = addend;
+        if (prot & PROT_WRITE) {
+            env->tlb_write[is_user][index].address = address;
+            env->tlb_write[is_user][index].addend = addend;
+        }
+    }
+    ret = 0;
+    /* XXX: incorrect for 4MB pages */
+    pd = physpage_find(pte & ~0xfff);
+    if ((pd & 0xfff) != 0) {
+        /* IO access: no mapping is done as it will be handled by the
+           soft MMU */
+        if (!(env->hflags & HF_SOFTMMU_MASK))
+            ret = 2;
+    } else {
+        void *map_addr;
+        map_addr = mmap((void *)virt_addr, page_size, prot, 
+                        MAP_SHARED | MAP_FIXED, phys_ram_fd, pd);
+        if (map_addr == MAP_FAILED) {
+            fprintf(stderr, 
+                    "mmap failed when mapped physical address 0x%08x to virtual address 0x%08x\n",
+                    pte & ~0xfff, virt_addr);
+            exit(1);
+        }
+#ifdef DEBUG_MMU
+        printf("mmaping 0x%08x to virt 0x%08x pse=%d\n", 
+               pte & ~0xfff, virt_addr, (page_size != 4096));
+#endif
+        page_set_flags(virt_addr, virt_addr + page_size, 
+                       PAGE_VALID | PAGE_EXEC | prot);
+    }
+    return ret;
+ do_fault_protect:
+    error_code = PG_ERROR_P_MASK;
+ do_fault:
+    env->cr[2] = addr;
+    env->error_code = (is_write << PG_ERROR_W_BIT) | error_code;
+    if (is_user)
+        env->error_code |= PG_ERROR_U_MASK;
+    return 1;
+}
--- a/hw/vga.c
+++ b/hw/vga.c
--- a/hw/vga_template.h
+++ b/hw/vga_template.h
@@ -0,0 +1,420 @@
+/*
+ * QEMU VGA Emulator templates
+ * 
+ * Copyright (c) 2003 Fabrice Bellard
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#if DEPTH == 8
+#define BPP 1
+#define PIXEL_TYPE uint8_t 
+#elif DEPTH == 15 || DEPTH == 16
+#define BPP 2
+#define PIXEL_TYPE uint16_t 
+#elif DEPTH == 32
+#define BPP 4
+#define PIXEL_TYPE uint32_t 
+#else
+#error unsupport depth
+#endif
+
+#if DEPTH != 15
+
+static inline void glue(vga_draw_glyph_line_, DEPTH)(uint8_t *d, 
+                                                     uint32_t font_data,
+                                                     uint32_t xorcol, 
+                                                     uint32_t bgcol)
+{
+#if BPP == 1
+        ((uint32_t *)d)[0] = (dmask16[(font_data >> 4)] & xorcol) ^ bgcol;
+        ((uint32_t *)d)[3] = (dmask16[(font_data >> 0) & 0xf] & xorcol) ^ bgcol;
+#elif BPP == 2
+        ((uint32_t *)d)[0] = (dmask4[(font_data >> 6)] & xorcol) ^ bgcol;
+        ((uint32_t *)d)[1] = (dmask4[(font_data >> 4) & 3] & xorcol) ^ bgcol;
+        ((uint32_t *)d)[2] = (dmask4[(font_data >> 2) & 3] & xorcol) ^ bgcol;
+        ((uint32_t *)d)[3] = (dmask4[(font_data >> 0) & 3] & xorcol) ^ bgcol;
+#else
+        ((uint32_t *)d)[0] = (-((font_data >> 7)) & xorcol) ^ bgcol;
+        ((uint32_t *)d)[1] = (-((font_data >> 6) & 1) & xorcol) ^ bgcol;
+        ((uint32_t *)d)[2] = (-((font_data >> 5) & 1) & xorcol) ^ bgcol;
+        ((uint32_t *)d)[3] = (-((font_data >> 4) & 1) & xorcol) ^ bgcol;
+        ((uint32_t *)d)[4] = (-((font_data >> 3) & 1) & xorcol) ^ bgcol;
+        ((uint32_t *)d)[5] = (-((font_data >> 2) & 1) & xorcol) ^ bgcol;
+        ((uint32_t *)d)[6] = (-((font_data >> 1) & 1) & xorcol) ^ bgcol;
+        ((uint32_t *)d)[7] = (-((font_data >> 0) & 1) & xorcol) ^ bgcol;
+#endif
+}
+
+static void glue(vga_draw_glyph8_, DEPTH)(uint8_t *d, int linesize,
+                                          const uint8_t *font_ptr, int h,
+                                          uint32_t fgcol, uint32_t bgcol)
+{
+    uint32_t font_data, xorcol;
+    
+    xorcol = bgcol ^ fgcol;
+    do {
+        font_data = font_ptr[0];
+        glue(vga_draw_glyph_line_, DEPTH)(d, font_data, xorcol, bgcol);
+        font_ptr += 4;
+        d += linesize;
+    } while (--h);
+}
+
+static void glue(vga_draw_glyph16_, DEPTH)(uint8_t *d, int linesize,
+                                          const uint8_t *font_ptr, int h,
+                                          uint32_t fgcol, uint32_t bgcol)
+{
+    uint32_t font_data, xorcol;
+    
+    xorcol = bgcol ^ fgcol;
+    do {
+        font_data = font_ptr[0];
+        glue(vga_draw_glyph_line_, DEPTH)(d, 
+                                          expand4to8[font_data >> 4], 
+                                          xorcol, bgcol);
+        glue(vga_draw_glyph_line_, DEPTH)(d + 8 * BPP, 
+                                          expand4to8[font_data & 0x0f], 
+                                          xorcol, bgcol);
+        font_ptr += 4;
+        d += linesize;
+    } while (--h);
+}
+
+static void glue(vga_draw_glyph9_, DEPTH)(uint8_t *d, int linesize,
+                                          const uint8_t *font_ptr, int h, 
+                                          uint32_t fgcol, uint32_t bgcol, int dup9)
+{
+    uint32_t font_data, xorcol, v;
+    
+    xorcol = bgcol ^ fgcol;
+    do {
+        font_data = font_ptr[0];
+        /* XXX: unaligned accesses are done */
+#if BPP == 1
+        ((uint32_t *)d)[0] = (dmask16[(font_data >> 4)] & xorcol) ^ bgcol;
+        v = (dmask16[(font_data >> 0) & 0xf] & xorcol) ^ bgcol;
+        ((uint32_t *)d)[3] = v;
+        if (dup9)
+            ((uint8_t *)d)[8] = v >> (24 * (1 - BIG));
+        else
+            ((uint8_t *)d)[8] = bgcol;
+        
+#elif BPP == 2
+        ((uint32_t *)d)[0] = (dmask4[(font_data >> 6)] & xorcol) ^ bgcol;
+        ((uint32_t *)d)[1] = (dmask4[(font_data >> 4) & 3] & xorcol) ^ bgcol;
+        ((uint32_t *)d)[2] = (dmask4[(font_data >> 2) & 3] & xorcol) ^ bgcol;
+        v = (dmask4[(font_data >> 0) & 3] & xorcol) ^ bgcol;
+        ((uint32_t *)d)[3] = v;
+        if (dup9)
+            ((uint16_t *)d)[8] = v >> (16 * (1 - BIG));
+        else
+            ((uint16_t *)d)[8] = bgcol;
+#else
+        ((uint32_t *)d)[0] = ((-(font_data >> 7)) & xorcol) ^ bgcol;
+        ((uint32_t *)d)[1] = ((-(font_data >> 6) & 1) & xorcol) ^ bgcol;
+        ((uint32_t *)d)[2] = ((-(font_data >> 5) & 1) & xorcol) ^ bgcol;
+        ((uint32_t *)d)[3] = ((-(font_data >> 4) & 1) & xorcol) ^ bgcol;
+        ((uint32_t *)d)[4] = ((-(font_data >> 3) & 1) & xorcol) ^ bgcol;
+        ((uint32_t *)d)[5] = ((-(font_data >> 2) & 1) & xorcol) ^ bgcol;
+        ((uint32_t *)d)[6] = ((-(font_data >> 1) & 1) & xorcol) ^ bgcol;
+        v = ((-(font_data >> 0) & 1) & xorcol) ^ bgcol;
+        ((uint32_t *)d)[7] = v;
+        if (dup9)
+            ((uint32_t *)d)[8] = v;
+        else
+            ((uint32_t *)d)[8] = bgcol;
+#endif
+        font_ptr += 4;
+        d += linesize;
+    } while (--h);
+}
+
+/* 
+ * 4 color mode
+ */
+static void glue(vga_draw_line2_, DEPTH)(VGAState *s1, uint8_t *d, 
+                                         const uint8_t *s, int width)
+{
+    uint32_t plane_mask, *palette, data, v;
+    int x;
+
+    palette = s1->last_palette;
+    plane_mask = mask16[s1->ar[0x12] & 0xf];
+    width >>= 3;
+    for(x = 0; x < width; x++) {
+        data = ((uint32_t *)s)[0];
+        data &= plane_mask;
+        v = expand2[GET_PLANE(data, 0)];
+        v |= expand2[GET_PLANE(data, 2)] << 2;
+        ((PIXEL_TYPE *)d)[0] = palette[v >> 12];
+        ((PIXEL_TYPE *)d)[1] = palette[(v >> 8) & 0xf];
+        ((PIXEL_TYPE *)d)[2] = palette[(v >> 4) & 0xf];
+        ((PIXEL_TYPE *)d)[3] = palette[(v >> 0) & 0xf];
+
+        v = expand2[GET_PLANE(data, 1)];
+        v |= expand2[GET_PLANE(data, 3)] << 2;
+        ((PIXEL_TYPE *)d)[4] = palette[v >> 12];
+        ((PIXEL_TYPE *)d)[5] = palette[(v >> 8) & 0xf];
+        ((PIXEL_TYPE *)d)[6] = palette[(v >> 4) & 0xf];
+        ((PIXEL_TYPE *)d)[7] = palette[(v >> 0) & 0xf];
+        d += BPP * 8;
+        s += 4;
+    }
+}
+
+#if BPP == 1
+#define PUT_PIXEL2(d, n, v) ((uint16_t *)d)[(n)] = (v)
+#elif BPP == 2
+#define PUT_PIXEL2(d, n, v) ((uint32_t *)d)[(n)] = (v)
+#else
+#define PUT_PIXEL2(d, n, v) \
+((uint32_t *)d)[2*(n)] = ((uint32_t *)d)[2*(n)+1] = (v)
+#endif
+
+/* 
+ * 4 color mode, dup2 horizontal
+ */
+static void glue(vga_draw_line2d2_, DEPTH)(VGAState *s1, uint8_t *d, 
+                                           const uint8_t *s, int width)
+{
+    uint32_t plane_mask, *palette, data, v;
+    int x;
+
+    palette = s1->last_palette;
+    plane_mask = mask16[s1->ar[0x12] & 0xf];
+    width >>= 3;
+    for(x = 0; x < width; x++) {
+        data = ((uint32_t *)s)[0];
+        data &= plane_mask;
+        v = expand2[GET_PLANE(data, 0)];
+        v |= expand2[GET_PLANE(data, 2)] << 2;
+        PUT_PIXEL2(d, 0, palette[v >> 12]);
+        PUT_PIXEL2(d, 1, palette[(v >> 8) & 0xf]);
+        PUT_PIXEL2(d, 2, palette[(v >> 4) & 0xf]);
+        PUT_PIXEL2(d, 3, palette[(v >> 0) & 0xf]);
+
+        v = expand2[GET_PLANE(data, 1)];
+        v |= expand2[GET_PLANE(data, 3)] << 2;
+        PUT_PIXEL2(d, 4, palette[v >> 12]);
+        PUT_PIXEL2(d, 5, palette[(v >> 8) & 0xf]);
+        PUT_PIXEL2(d, 6, palette[(v >> 4) & 0xf]);
+        PUT_PIXEL2(d, 7, palette[(v >> 0) & 0xf]);
+        d += BPP * 16;
+        s += 4;
+    }
+}
+
+/* 
+ * 16 color mode
+ */
+static void glue(vga_draw_line4_, DEPTH)(VGAState *s1, uint8_t *d, 
+                                         const uint8_t *s, int width)
+{
+    uint32_t plane_mask, data, v, *palette;
+    int x;
+
+    palette = s1->last_palette;
+    plane_mask = mask16[s1->ar[0x12] & 0xf];
+    width >>= 3;
+    for(x = 0; x < width; x++) {
+        data = ((uint32_t *)s)[0];
+        data &= plane_mask;
+        v = expand4[GET_PLANE(data, 0)];
+        v |= expand4[GET_PLANE(data, 1)] << 1;
+        v |= expand4[GET_PLANE(data, 2)] << 2;
+        v |= expand4[GET_PLANE(data, 3)] << 3;
+        ((PIXEL_TYPE *)d)[0] = palette[v >> 28];
+        ((PIXEL_TYPE *)d)[1] = palette[(v >> 24) & 0xf];
+        ((PIXEL_TYPE *)d)[2] = palette[(v >> 20) & 0xf];
+        ((PIXEL_TYPE *)d)[3] = palette[(v >> 16) & 0xf];
+        ((PIXEL_TYPE *)d)[4] = palette[(v >> 12) & 0xf];
+        ((PIXEL_TYPE *)d)[5] = palette[(v >> 8) & 0xf];
+        ((PIXEL_TYPE *)d)[6] = palette[(v >> 4) & 0xf];
+        ((PIXEL_TYPE *)d)[7] = palette[(v >> 0) & 0xf];
+        d += BPP * 8;
+        s += 4;
+    }
+}
+
+/* 
+ * 16 color mode, dup2 horizontal
+ */
+static void glue(vga_draw_line4d2_, DEPTH)(VGAState *s1, uint8_t *d, 
+                                           const uint8_t *s, int width)
+{
+    uint32_t plane_mask, data, v, *palette;
+    int x;
+
+    palette = s1->last_palette;
+    plane_mask = mask16[s1->ar[0x12] & 0xf];
+    width >>= 3;
+    for(x = 0; x < width; x++) {
+        data = ((uint32_t *)s)[0];
+        data &= plane_mask;
+        v = expand4[GET_PLANE(data, 0)];
+        v |= expand4[GET_PLANE(data, 1)] << 1;
+        v |= expand4[GET_PLANE(data, 2)] << 2;
+        v |= expand4[GET_PLANE(data, 3)] << 3;
+        PUT_PIXEL2(d, 0, palette[v >> 28]);
+        PUT_PIXEL2(d, 1, palette[(v >> 24) & 0xf]);
+        PUT_PIXEL2(d, 2, palette[(v >> 20) & 0xf]);
+        PUT_PIXEL2(d, 3, palette[(v >> 16) & 0xf]);
+        PUT_PIXEL2(d, 4, palette[(v >> 12) & 0xf]);
+        PUT_PIXEL2(d, 5, palette[(v >> 8) & 0xf]);
+        PUT_PIXEL2(d, 6, palette[(v >> 4) & 0xf]);
+        PUT_PIXEL2(d, 7, palette[(v >> 0) & 0xf]);
+        d += BPP * 16;
+        s += 4;
+    }
+}
+
+/* 
+ * 256 color mode, double pixels
+ *
+ * XXX: add plane_mask support (never used in standard VGA modes)
+ */
+static void glue(vga_draw_line8d2_, DEPTH)(VGAState *s1, uint8_t *d, 
+                                           const uint8_t *s, int width)
+{
+    uint32_t *palette;
+    int x;
+
+    palette = s1->last_palette;
+    width >>= 3;
+    for(x = 0; x < width; x++) {
+        PUT_PIXEL2(d, 0, palette[s[0]]);
+        PUT_PIXEL2(d, 1, palette[s[1]]);
+        PUT_PIXEL2(d, 2, palette[s[2]]);
+        PUT_PIXEL2(d, 3, palette[s[3]]);
+        d += BPP * 8;
+        s += 4;
+    }
+}
+
+/* 
+ * standard 256 color mode
+ *
+ * XXX: add plane_mask support (never used in standard VGA modes)
+ */
+static void glue(vga_draw_line8_, DEPTH)(VGAState *s1, uint8_t *d, 
+                                         const uint8_t *s, int width)
+{
+    uint32_t *palette;
+    int x;
+
+    palette = s1->last_palette;
+    width >>= 3;
+    for(x = 0; x < width; x++) {
+        ((PIXEL_TYPE *)d)[0] = palette[s[0]];
+        ((PIXEL_TYPE *)d)[1] = palette[s[1]];
+        ((PIXEL_TYPE *)d)[2] = palette[s[2]];
+        ((PIXEL_TYPE *)d)[3] = palette[s[3]];
+        ((PIXEL_TYPE *)d)[4] = palette[s[4]];
+        ((PIXEL_TYPE *)d)[5] = palette[s[5]];
+        ((PIXEL_TYPE *)d)[6] = palette[s[6]];
+        ((PIXEL_TYPE *)d)[7] = palette[s[7]];
+        d += BPP * 8;
+        s += 8;
+    }
+}
+
+#endif /* DEPTH != 15 */
+
+
+/* XXX: optimize */
+
+/* 
+ * 15 bit color
+ */
+static void glue(vga_draw_line15_, DEPTH)(VGAState *s1, uint8_t *d, 
+                                          const uint8_t *s, int width)
+{
+#if DEPTH == 15 && !defined(WORDS_BIGENDIAN)
+    memcpy(d, s, width * 2);
+#else
+    int w;
+    uint32_t v, r, g, b;
+
+    w = width;
+    do {
+        v = lduw((void *)s);
+        r = (v >> 7) & 0xf8;
+        g = (v >> 2) & 0xf8;
+        b = (v << 3) & 0xf8;
+        ((PIXEL_TYPE *)d)[0] = glue(rgb_to_pixel, DEPTH)(r, g, b);
+        s += 2;
+        d += BPP;
+    } while (--w != 0);
+#endif    
+}
+
+/* 
+ * 16 bit color
+ */
+static void glue(vga_draw_line16_, DEPTH)(VGAState *s1, uint8_t *d, 
+                                          const uint8_t *s, int width)
+{
+#if DEPTH == 16 && !defined(WORDS_BIGENDIAN)
+    memcpy(d, s, width * 2);
+#else
+    int w;
+    uint32_t v, r, g, b;
+
+    w = width;
+    do {
+        v = lduw((void *)s);
+        r = (v >> 8) & 0xf8;
+        g = (v >> 3) & 0xfc;
+        b = (v << 3) & 0xf8;
+        ((PIXEL_TYPE *)d)[0] = glue(rgb_to_pixel, DEPTH)(r, g, b);
+        s += 2;
+        d += BPP;
+    } while (--w != 0);
+#endif    
+}
+
+/* 
+ * 32 bit color
+ */
+static void glue(vga_draw_line32_, DEPTH)(VGAState *s1, uint8_t *d, 
+                                          const uint8_t *s, int width)
+{
+#if DEPTH == 32 && !defined(WORDS_BIGENDIAN)
+    memcpy(d, s, width * 4);
+#else
+    int w;
+    uint32_t r, g, b;
+
+    w = width;
+    do {
+        b = s[0];
+        g = s[1];
+        r = s[2];
+        ((PIXEL_TYPE *)d)[0] = glue(rgb_to_pixel, DEPTH)(r, g, b);
+        s += 4;
+        d += BPP;
+    } while (--w != 0);
+#endif
+}
+
+#undef PUT_PIXEL2
+#undef DEPTH
+#undef BPP
+#undef PIXEL_TYPE
--- a/i386-vl.ld
+++ b/i386-vl.ld
@@ -0,0 +1,140 @@
+/* ld script to make i386 Linux kernel
+ * Written by Martin Mares <mj@atrey.karlin.mff.cuni.cz>;
+ */
+OUTPUT_FORMAT("elf32-i386", "elf32-i386", "elf32-i386")
+OUTPUT_ARCH(i386)
+SEARCH_DIR(/lib); SEARCH_DIR(/usr/lib); SEARCH_DIR(/usr/local/lib); SEARCH_DIR(/usr/alpha-unknown-linux-gnu/lib);
+ENTRY(_start)
+SECTIONS
+{
+  /* Read-only sections, merged into text segment: */
+  . = 0xa8000000 + SIZEOF_HEADERS;
+  .interp     : { *(.interp) 	}
+  .hash          : { *(.hash)		}
+  .dynsym        : { *(.dynsym)		}
+  .dynstr        : { *(.dynstr)		}
+  .gnu.version   : { *(.gnu.version)	}
+  .gnu.version_d   : { *(.gnu.version_d)	}
+  .gnu.version_r   : { *(.gnu.version_r)	}
+  .rel.text      :
+    { *(.rel.text) *(.rel.gnu.linkonce.t*) }
+  .rela.text     :
+    { *(.rela.text) *(.rela.gnu.linkonce.t*) }
+  .rel.data      :
+    { *(.rel.data) *(.rel.gnu.linkonce.d*) }
+  .rela.data     :
+    { *(.rela.data) *(.rela.gnu.linkonce.d*) }
+  .rel.rodata    :
+    { *(.rel.rodata) *(.rel.gnu.linkonce.r*) }
+  .rela.rodata   :
+    { *(.rela.rodata) *(.rela.gnu.linkonce.r*) }
+  .rel.got       : { *(.rel.got)		}
+  .rela.got      : { *(.rela.got)		}
+  .rel.ctors     : { *(.rel.ctors)	}
+  .rela.ctors    : { *(.rela.ctors)	}
+  .rel.dtors     : { *(.rel.dtors)	}
+  .rela.dtors    : { *(.rela.dtors)	}
+  .rel.init      : { *(.rel.init)	}
+  .rela.init     : { *(.rela.init)	}
+  .rel.fini      : { *(.rel.fini)	}
+  .rela.fini     : { *(.rela.fini)	}
+  .rel.bss       : { *(.rel.bss)		}
+  .rela.bss      : { *(.rela.bss)		}
+  .rel.plt       : { *(.rel.plt)		}
+  .rela.plt      : { *(.rela.plt)		}
+  .init          : { *(.init)	} =0x47ff041f
+  .text      :
+  {
+    *(.text)
+    /* .gnu.warning sections are handled specially by elf32.em.  */
+    *(.gnu.warning)
+    *(.gnu.linkonce.t*)
+  } =0x47ff041f
+  _etext = .;
+  PROVIDE (etext = .);
+  .fini      : { *(.fini)    } =0x47ff041f
+  .rodata    : { *(.rodata) *(.gnu.linkonce.r*) }
+  .rodata1   : { *(.rodata1) }
+  .reginfo : { *(.reginfo) }
+  __preinit_array_start = .;
+  .preinit_array : { *(.preinit_array) }
+  __preinit_array_end = .;
+  __init_array_start = .;
+  .init_array : { *(.init_array) }
+  __init_array_end = .;
+  __fini_array_start = .;
+  .fini_array : { *(.fini_array) }
+  __fini_array_end = .;
+
+  /* Adjust the address for the data segment.  We want to adjust up to
+     the same address within the page on the next page up.  */
+  . = ALIGN(0x100000) + (. & (0x100000 - 1));
+  .data    :
+  {
+    *(.data)
+    *(.gnu.linkonce.d*)
+    CONSTRUCTORS
+  }
+  .data1   : { *(.data1) }
+  .ctors         :
+  {
+    *(.ctors)
+  }
+  .dtors         :
+  {
+    *(.dtors)
+  }
+  .plt      : { *(.plt)	}
+  .got           : { *(.got.plt) *(.got) }
+  .dynamic       : { *(.dynamic) }
+  /* We want the small data sections together, so single-instruction offsets
+     can access them all, and initialized data all before uninitialized, so
+     we can shorten the on-disk segment size.  */
+  .sdata     : { *(.sdata) }
+  _edata  =  .;
+  PROVIDE (edata = .);
+  __bss_start = .;
+  .sbss      : { *(.sbss) *(.scommon) }
+  .bss       :
+  {
+   *(.dynbss)
+   *(.bss)
+   *(COMMON)
+  }
+  _end = . ;
+  PROVIDE (end = .);
+  /* Stabs debugging sections.  */
+  .stab 0 : { *(.stab) }
+  .stabstr 0 : { *(.stabstr) }
+  .stab.excl 0 : { *(.stab.excl) }
+  .stab.exclstr 0 : { *(.stab.exclstr) }
+  .stab.index 0 : { *(.stab.index) }
+  .stab.indexstr 0 : { *(.stab.indexstr) }
+  .comment 0 : { *(.comment) }
+  /* DWARF debug sections.
+     Symbols in the DWARF debugging sections are relative to the beginning
+     of the section so we begin them at 0.  */
+  /* DWARF 1 */
+  .debug          0 : { *(.debug) }
+  .line           0 : { *(.line) }
+  /* GNU DWARF 1 extensions */
+  .debug_srcinfo  0 : { *(.debug_srcinfo) }
+  .debug_sfnames  0 : { *(.debug_sfnames) }
+  /* DWARF 1.1 and DWARF 2 */
+  .debug_aranges  0 : { *(.debug_aranges) }
+  .debug_pubnames 0 : { *(.debug_pubnames) }
+  /* DWARF 2 */
+  .debug_info     0 : { *(.debug_info) }
+  .debug_abbrev   0 : { *(.debug_abbrev) }
+  .debug_line     0 : { *(.debug_line) }
+  .debug_frame    0 : { *(.debug_frame) }
+  .debug_str      0 : { *(.debug_str) }
+  .debug_loc      0 : { *(.debug_loc) }
+  .debug_macinfo  0 : { *(.debug_macinfo) }
+  /* SGI/MIPS DWARF 2 extensions */
+  .debug_weaknames 0 : { *(.debug_weaknames) }
+  .debug_funcnames 0 : { *(.debug_funcnames) }
+  .debug_typenames 0 : { *(.debug_typenames) }
+  .debug_varnames  0 : { *(.debug_varnames) }
+  /* These must appear regardless of  .  */
+}
--- a/linux-user/elfload.c
+++ b/linux-user/elfload.c
@@ -74,7 +74,8 @@ static inline void init_thread(struct target_pt_regs *regs, struct image_info *i
    regs->ARM_sp = infop->start_stack;
    regs->ARM_r2 = tswapl(stack[2]); /* envp */
    regs->ARM_r1 = tswapl(stack[1]); /* argv */
-    regs->ARM_r0 = tswapl(stack[0]); /* argc */
+    /* XXX: it seems that r0 is zeroed after ! */
+    //    regs->ARM_r0 = tswapl(stack[0]); /* argc */
 }

 #define USE_ELF_CORE_DUMP
@@ -490,7 +491,7 @@ static unsigned int * create_elf_tables(char *p, int argc, int envc,
         * Force 16 byte alignment here for generality.
         */
        sp = (unsigned int *) (~15UL & (unsigned long) p);
-        sp -= exec ? DLINFO_ITEMS*2 : 2;
+        sp -= DLINFO_ITEMS*2;
        dlinfo = sp;
        sp -= envc+1;
        envp = sp;
@@ -505,21 +506,20 @@ static unsigned int * create_elf_tables(char *p, int argc, int envc,
          put_user (tswapl(id), dlinfo++); \
          put_user (tswapl(val), dlinfo++)

-        if (exec) { /* Put this here for an ELF program interpreter */
-          NEW_AUX_ENT (AT_PHDR, (target_ulong)(load_addr + exec->e_phoff));
-          NEW_AUX_ENT (AT_PHENT, (target_ulong)(sizeof (struct elf_phdr)));
-          NEW_AUX_ENT (AT_PHNUM, (target_ulong)(exec->e_phnum));
-          NEW_AUX_ENT (AT_PAGESZ, (target_ulong)(TARGET_PAGE_SIZE));
-          NEW_AUX_ENT (AT_BASE, (target_ulong)(interp_load_addr));
-          NEW_AUX_ENT (AT_FLAGS, (target_ulong)0);
-          NEW_AUX_ENT (AT_ENTRY, load_bias + exec->e_entry);
-          NEW_AUX_ENT (AT_UID, (target_ulong) getuid());
-          NEW_AUX_ENT (AT_EUID, (target_ulong) geteuid());
-          NEW_AUX_ENT (AT_GID, (target_ulong) getgid());
-          NEW_AUX_ENT (AT_EGID, (target_ulong) getegid());
-        }
+        NEW_AUX_ENT (AT_PHDR, (target_ulong)(load_addr + exec->e_phoff));
+        NEW_AUX_ENT (AT_PHENT, (target_ulong)(sizeof (struct elf_phdr)));
+        NEW_AUX_ENT (AT_PHNUM, (target_ulong)(exec->e_phnum));
+        NEW_AUX_ENT (AT_PAGESZ, (target_ulong)(TARGET_PAGE_SIZE));
+        NEW_AUX_ENT (AT_BASE, (target_ulong)(interp_load_addr));
+        NEW_AUX_ENT (AT_FLAGS, (target_ulong)0);
+        NEW_AUX_ENT (AT_ENTRY, load_bias + exec->e_entry);
+        NEW_AUX_ENT (AT_UID, (target_ulong) getuid());
+        NEW_AUX_ENT (AT_EUID, (target_ulong) geteuid());
+        NEW_AUX_ENT (AT_GID, (target_ulong) getgid());
+        NEW_AUX_ENT (AT_EGID, (target_ulong) getegid());
        NEW_AUX_ENT (AT_NULL, 0);
 #undef NEW_AUX_ENT
+
        put_user(tswapl(argc),--sp);
        info->arg_start = (unsigned int)((unsigned long)p & 0xffffffff);
        while (argc-->0) {
@@ -1087,7 +1087,7 @@ static int load_elf_binary(struct linux_binprm * bprm, struct target_pt_regs * r
      create_elf_tables((char *)bprm->p,
 		    bprm->argc,
 		    bprm->envc,
-		    (interpreter_type == INTERPRETER_ELF ? &elf_ex : NULL),
+                    &elf_ex,
                    load_addr, load_bias,
 		    interp_load_addr,
 		    (interpreter_type == INTERPRETER_AOUT ? 0 : 1),
--- a/linux-user/main.c
+++ b/linux-user/main.c
@@ -26,8 +26,6 @@

 #include "qemu.h"

-#include "cpu-i386.h"
-
 #define DEBUG_LOGFILE "/tmp/qemu.log"

 FILE *logfile = NULL;
@@ -100,6 +98,11 @@ int cpu_x86_inl(CPUX86State *env, int addr)
    return 0;
 }

+int cpu_x86_get_pic_interrupt(CPUX86State *env)
+{
+    return -1;
+}
+
 static void write_dt(void *ptr, unsigned long addr, unsigned long limit, 
                     int flags)
 {
@@ -176,7 +179,7 @@ void cpu_loop(CPUX86State *env)
                info.si_code = TARGET_SEGV_MAPERR;
            else
                info.si_code = TARGET_SEGV_ACCERR;
-            info._sifields._sigfault._addr = env->cr2;
+            info._sifields._sigfault._addr = env->cr[2];
            queue_signal(info.si_signo, &info);
            break;
        case EXCP00_DIVZ:
@@ -231,7 +234,7 @@ void cpu_loop(CPUX86State *env)
            /* just indicate that signals should be handled asap */
            break;
        default:
-            pc = env->seg_cache[R_CS].base + env->eip;
+            pc = env->segs[R_CS].base + env->eip;
            fprintf(stderr, "qemu: 0x%08lx: unhandled CPU exception 0x%x - aborting\n", 
                    (long)pc, trapnr);
            abort();
@@ -243,8 +246,6 @@ void cpu_loop(CPUX86State *env)

 #ifdef TARGET_ARM

-#define ARM_SYSCALL_BASE	0x900000
-
 void cpu_loop(CPUARMState *env)
 {
    int trapnr;
@@ -282,6 +283,9 @@ void cpu_loop(CPUARMState *env)
                }
            }
            break;
+        case EXCP_INTERRUPT:
+            /* just indicate that signals should be handled asap */
+            break;
        default:
        error:
            fprintf(stderr, "qemu: unhandled CPU exception 0x%x - aborting\n", 
@@ -317,6 +321,9 @@ void usage(void)

 /* XXX: currently only used for async signals (see signal.c) */
 CPUState *global_env;
+/* used only if single thread */
+CPUState *cpu_single_env = NULL;
+
 /* used to free thread contexts */
 TaskState *first_task_state;

@@ -395,7 +402,7 @@ int main(int argc, char **argv)
    /* NOTE: we need to init the CPU at this stage to get the
       host_page_size */
    env = cpu_init();
-
+    
    if (elf_exec(filename, argv+optind, environ, regs, info) != 0) {
 	printf("Error loading %s\n", filename);
 	_exit(1);
@@ -423,8 +430,13 @@ int main(int argc, char **argv)
    memset(ts, 0, sizeof(TaskState));
    env->opaque = ts;
    ts->used = 1;
+    env->user_mode_only = 1;
    
 #if defined(TARGET_I386)
+    cpu_x86_set_cpl(env, 3);
+
+    env->cr[0] = CR0_PG_MASK | CR0_WP_MASK | CR0_PE_MASK;
+
    /* linux register setup */
    env->regs[R_EAX] = regs->eax;
    env->regs[R_EBX] = regs->ebx;
@@ -476,6 +488,7 @@ int main(int argc, char **argv)
    cpu_x86_load_seg(env, R_SS, __USER_DS);
    cpu_x86_load_seg(env, R_FS, __USER_DS);
    cpu_x86_load_seg(env, R_GS, __USER_DS);
+
 #elif defined(TARGET_ARM)
    {
        int i;
--- a/linux-user/qemu.h
+++ b/linux-user/qemu.h
@@ -6,8 +6,15 @@
 #include <signal.h>
 #include "syscall_defs.h"

-#include "cpu-" TARGET_ARCH ".h"
-#include "syscall-" TARGET_ARCH ".h"
+#if defined(TARGET_I386)
+#include "cpu-i386.h"
+#include "syscall-i386.h"
+#elif defined(TARGET_ARM)
+#include "cpu-arm.h"
+#include "syscall-arm.h"
+#else
+#error unsupported target CPU
+#endif

 /* This struct is used to hold certain information about the image.
 * Basically, it replicates in user space what would be certain
--- a/linux-user/signal.c
+++ b/linux-user/signal.c
@@ -60,44 +60,122 @@ static int signal_pending; /* non zero if a signal may be pending */
 static void host_signal_handler(int host_signum, siginfo_t *info, 
                                void *puc);

-/* XXX: do it properly */
+static uint8_t host_to_target_signal_table[65] = {
+    [SIGHUP] = TARGET_SIGHUP,
+    [SIGINT] = TARGET_SIGINT,
+    [SIGQUIT] = TARGET_SIGQUIT,
+    [SIGILL] = TARGET_SIGILL,
+    [SIGTRAP] = TARGET_SIGTRAP,
+    [SIGABRT] = TARGET_SIGABRT,
+    [SIGIOT] = TARGET_SIGIOT,
+    [SIGBUS] = TARGET_SIGBUS,
+    [SIGFPE] = TARGET_SIGFPE,
+    [SIGKILL] = TARGET_SIGKILL,
+    [SIGUSR1] = TARGET_SIGUSR1,
+    [SIGSEGV] = TARGET_SIGSEGV,
+    [SIGUSR2] = TARGET_SIGUSR2,
+    [SIGPIPE] = TARGET_SIGPIPE,
+    [SIGALRM] = TARGET_SIGALRM,
+    [SIGTERM] = TARGET_SIGTERM,
+#ifdef SIGSTKFLT
+    [SIGSTKFLT] = TARGET_SIGSTKFLT,
+#endif
+    [SIGCHLD] = TARGET_SIGCHLD,
+    [SIGCONT] = TARGET_SIGCONT,
+    [SIGSTOP] = TARGET_SIGSTOP,
+    [SIGTSTP] = TARGET_SIGTSTP,
+    [SIGTTIN] = TARGET_SIGTTIN,
+    [SIGTTOU] = TARGET_SIGTTOU,
+    [SIGURG] = TARGET_SIGURG,
+    [SIGXCPU] = TARGET_SIGXCPU,
+    [SIGXFSZ] = TARGET_SIGXFSZ,
+    [SIGVTALRM] = TARGET_SIGVTALRM,
+    [SIGPROF] = TARGET_SIGPROF,
+    [SIGWINCH] = TARGET_SIGWINCH,
+    [SIGIO] = TARGET_SIGIO,
+    [SIGPWR] = TARGET_SIGPWR,
+    [SIGSYS] = TARGET_SIGSYS,
+    /* next signals stay the same */
+};
+static uint8_t target_to_host_signal_table[65];
+
 static inline int host_to_target_signal(int sig)
 {
-    return sig;
+    return host_to_target_signal_table[sig];
 }

 static inline int target_to_host_signal(int sig)
 {
-    return sig;
+    return target_to_host_signal_table[sig];
 }

-void host_to_target_sigset(target_sigset_t *d, sigset_t *s)
+void host_to_target_sigset(target_sigset_t *d, const sigset_t *s)
 {
    int i;
-    for(i = 0;i < TARGET_NSIG_WORDS; i++) {
+    unsigned long sigmask;
+    uint32_t target_sigmask;
+    
+    sigmask = ((unsigned long *)s)[0];
+    target_sigmask = 0;
+    for(i = 0; i < 32; i++) {
+        if (sigmask & (1 << i)) 
+            target_sigmask |= 1 << (host_to_target_signal(i + 1) - 1);
+    }
+#if TARGET_LONG_BITS == 32 && HOST_LONG_BITS == 32
+    d->sig[0] = tswapl(target_sigmask);
+    for(i = 1;i < TARGET_NSIG_WORDS; i++) {
        d->sig[i] = tswapl(((unsigned long *)s)[i]);
    }
+#elif TARGET_LONG_BITS == 32 && HOST_LONG_BITS == 64 && TARGET_NSIG_WORDS == 2
+    d->sig[0] = tswapl(target_sigmask);
+    d->sig[1] = tswapl(sigmask >> 32);
+#else
+#error host_to_target_sigset
+#endif
 }

-void target_to_host_sigset(sigset_t *d, target_sigset_t *s)
+void target_to_host_sigset(sigset_t *d, const target_sigset_t *s)
 {
    int i;
-    for(i = 0;i < TARGET_NSIG_WORDS; i++) {
+    unsigned long sigmask;
+    target_ulong target_sigmask;
+
+    target_sigmask = tswapl(s->sig[0]);
+    sigmask = 0;
+    for(i = 0; i < 32; i++) {
+        if (target_sigmask & (1 << i)) 
+            sigmask |= 1 << (target_to_host_signal(i + 1) - 1);
+    }
+#if TARGET_LONG_BITS == 32 && HOST_LONG_BITS == 32
+    ((unsigned long *)d)[0] = sigmask;
+    for(i = 1;i < TARGET_NSIG_WORDS; i++) {
        ((unsigned long *)d)[i] = tswapl(s->sig[i]);
    }
+#elif TARGET_LONG_BITS == 32 && HOST_LONG_BITS == 64 && TARGET_NSIG_WORDS == 2
+    ((unsigned long *)d)[0] = sigmask | ((unsigned long)tswapl(s->sig[1]) << 32);
+#else
+#error target_to_host_sigset
+#endif /* TARGET_LONG_BITS */
 }

 void host_to_target_old_sigset(target_ulong *old_sigset, 
                               const sigset_t *sigset)
 {
-    *old_sigset = tswap32(*(unsigned long *)sigset & 0xffffffff);
+    target_sigset_t d;
+    host_to_target_sigset(&d, sigset);
+    *old_sigset = d.sig[0];
 }

 void target_to_host_old_sigset(sigset_t *sigset, 
                               const target_ulong *old_sigset)
 {
-    sigemptyset(sigset);
-    *(unsigned long *)sigset = tswapl(*old_sigset);
+    target_sigset_t d;
+    int i;
+
+    d.sig[0] = *old_sigset;
+    for(i = 1;i < TARGET_NSIG_WORDS; i++)
+        d.sig[i] = 0;
+    target_to_host_sigset(sigset, &d);
 }

 /* siginfo conversion */
@@ -167,8 +245,18 @@ void target_to_host_siginfo(siginfo_t *info, const target_siginfo_t *tinfo)
 void signal_init(void)
 {
    struct sigaction act;
-    int i;
+    int i, j;

+    /* generate signal conversion tables */
+    for(i = 1; i <= 64; i++) {
+        if (host_to_target_signal_table[i] == 0)
+            host_to_target_signal_table[i] = i;
+    }
+    for(i = 1; i <= 64; i++) {
+        j = host_to_target_signal_table[i];
+        target_to_host_signal_table[j] = i;
+    }
+        
    /* set all host signal handlers. ALL signals are blocked during
       the handlers to serialize them. */
    sigfillset(&act.sa_mask);
@@ -333,7 +421,7 @@ static void host_signal_handler(int host_signum, siginfo_t *info,
    host_to_target_siginfo_noswap(&tinfo, info);
    if (queue_signal(sig, &tinfo) == 1) {
        /* interrupt the virtual CPU as soon as possible */
-        cpu_interrupt(global_env);
+        cpu_interrupt(global_env, CPU_INTERRUPT_EXIT);
    }
 }

@@ -364,6 +452,80 @@ int do_sigaction(int sig, const struct target_sigaction *act,
    return 0;
 }

+#define __put_user(x,ptr)\
+({\
+    int size = sizeof(*ptr);\
+    switch(size) {\
+    case 1:\
+        stb(ptr, (typeof(*ptr))(x));\
+        break;\
+    case 2:\
+        stw(ptr, (typeof(*ptr))(x));\
+        break;\
+    case 4:\
+        stl(ptr, (typeof(*ptr))(x));\
+        break;\
+    case 8:\
+        stq(ptr, (typeof(*ptr))(x));\
+        break;\
+    default:\
+        abort();\
+    }\
+    0;\
+})
+
+#define __get_user(x, ptr) \
+({\
+    int size = sizeof(*ptr);\
+    switch(size) {\
+    case 1:\
+        x = (typeof(*ptr))ldub(ptr);\
+        break;\
+    case 2:\
+        x = (typeof(*ptr))lduw(ptr);\
+        break;\
+    case 4:\
+        x = (typeof(*ptr))ldl(ptr);\
+        break;\
+    case 8:\
+        x = (typeof(*ptr))ldq(ptr);\
+        break;\
+    default:\
+        abort();\
+    }\
+    0;\
+})
+
+
+#define __copy_to_user(dst, src, size)\
+({\
+    memcpy(dst, src, size);\
+    0;\
+})
+
+#define __copy_from_user(dst, src, size)\
+({\
+    memcpy(dst, src, size);\
+    0;\
+})
+
+#define __clear_user(dst, size)\
+({\
+    memset(dst, 0, size);\
+    0;\
+})
+
+#ifndef offsetof
+#define offsetof(type, field) ((size_t) &((type *)0)->field)
+#endif
+
+static inline int copy_siginfo_to_user(target_siginfo_t *tinfo, 
+                                       const target_siginfo_t *info)
+{
+    tswap_siginfo(tinfo, info);
+    return 0;
+}
+
 #ifdef TARGET_I386

 /* from the Linux kernel */
@@ -472,44 +634,6 @@ struct rt_sigframe
 * Set up a signal frame.
 */

-#define __put_user(x,ptr)\
-({\
-    int size = sizeof(*ptr);\
-    switch(size) {\
-    case 1:\
-        stb(ptr, (typeof(*ptr))(x));\
-        break;\
-    case 2:\
-        stw(ptr, (typeof(*ptr))(x));\
-        break;\
-    case 4:\
-        stl(ptr, (typeof(*ptr))(x));\
-        break;\
-    case 8:\
-        stq(ptr, (typeof(*ptr))(x));\
-        break;\
-    default:\
-        abort();\
-    }\
-    0;\
-})
-
-#define get_user(val, ptr) (typeof(*ptr))(*(ptr))
-
-
-#define __copy_to_user(dst, src, size)\
-({\
-    memcpy(dst, src, size);\
-    0;\
-})
-
-static inline int copy_siginfo_to_user(target_siginfo_t *tinfo, 
-                                       const target_siginfo_t *info)
-{
-    tswap_siginfo(tinfo, info);
-    return 0;
-}
-
 /* XXX: save x87 state */
 static int
 setup_sigcontext(struct target_sigcontext *sc, struct target_fpstate *fpstate,
@@ -517,10 +641,10 @@ setup_sigcontext(struct target_sigcontext *sc, struct target_fpstate *fpstate,
 {
 	int err = 0;

-	err |= __put_user(env->segs[R_GS], (unsigned int *)&sc->gs);
-	err |= __put_user(env->segs[R_FS], (unsigned int *)&sc->fs);
-	err |= __put_user(env->segs[R_ES], (unsigned int *)&sc->es);
-	err |= __put_user(env->segs[R_DS], (unsigned int *)&sc->ds);
+	err |= __put_user(env->segs[R_GS].selector, (unsigned int *)&sc->gs);
+	err |= __put_user(env->segs[R_FS].selector, (unsigned int *)&sc->fs);
+	err |= __put_user(env->segs[R_ES].selector, (unsigned int *)&sc->es);
+	err |= __put_user(env->segs[R_DS].selector, (unsigned int *)&sc->ds);
 	err |= __put_user(env->regs[R_EDI], &sc->edi);
 	err |= __put_user(env->regs[R_ESI], &sc->esi);
 	err |= __put_user(env->regs[R_EBP], &sc->ebp);
@@ -532,10 +656,10 @@ setup_sigcontext(struct target_sigcontext *sc, struct target_fpstate *fpstate,
 	err |= __put_user(env->exception_index, &sc->trapno);
 	err |= __put_user(env->error_code, &sc->err);
 	err |= __put_user(env->eip, &sc->eip);
-	err |= __put_user(env->segs[R_CS], (unsigned int *)&sc->cs);
+	err |= __put_user(env->segs[R_CS].selector, (unsigned int *)&sc->cs);
 	err |= __put_user(env->eflags, &sc->eflags);
 	err |= __put_user(env->regs[R_ESP], &sc->esp_at_signal);
-	err |= __put_user(env->segs[R_SS], (unsigned int *)&sc->ss);
+	err |= __put_user(env->segs[R_SS].selector, (unsigned int *)&sc->ss);

        cpu_x86_fsave(env, (void *)fpstate, 1);
        fpstate->status = fpstate->sw;
@@ -544,7 +668,7 @@ setup_sigcontext(struct target_sigcontext *sc, struct target_fpstate *fpstate,

 	/* non-iBCS2 extensions.. */
 	err |= __put_user(mask, &sc->oldmask);
-	err |= __put_user(env->cr2, &sc->cr2);
+	err |= __put_user(env->cr[2], &sc->cr2);
 	return err;
 }

@@ -567,13 +691,14 @@ get_sigframe(struct emulated_sigaction *ka, CPUX86State *env, size_t frame_size)
 	}

 	/* This is the legacy signal stack switching. */
-	else if ((regs->xss & 0xffff) != __USER_DS &&
-		 !(ka->sa.sa_flags & SA_RESTORER) &&
-		 ka->sa.sa_restorer) {
-		esp = (unsigned long) ka->sa.sa_restorer;
-	}
+	else 
 #endif
-	return (void *)((esp - frame_size) & -8ul);
+        if ((env->segs[R_SS].selector & 0xffff) != __USER_DS &&
+            !(ka->sa.sa_flags & TARGET_SA_RESTORER) &&
+            ka->sa.sa_restorer) {
+            esp = (unsigned long) ka->sa.sa_restorer;
+	}
+        return (void *)((esp - frame_size) & -8ul);
 }

 static void setup_frame(int sig, struct emulated_sigaction *ka,
@@ -824,6 +949,377 @@ badframe:
 	return 0;
 }

+#elif defined(TARGET_ARM)
+
+struct target_sigcontext {
+	target_ulong trap_no;
+	target_ulong error_code;
+	target_ulong oldmask;
+	target_ulong arm_r0;
+	target_ulong arm_r1;
+	target_ulong arm_r2;
+	target_ulong arm_r3;
+	target_ulong arm_r4;
+	target_ulong arm_r5;
+	target_ulong arm_r6;
+	target_ulong arm_r7;
+	target_ulong arm_r8;
+	target_ulong arm_r9;
+	target_ulong arm_r10;
+	target_ulong arm_fp;
+	target_ulong arm_ip;
+	target_ulong arm_sp;
+	target_ulong arm_lr;
+	target_ulong arm_pc;
+	target_ulong arm_cpsr;
+	target_ulong fault_address;
+};
+
+typedef struct target_sigaltstack {
+	target_ulong ss_sp;
+	int ss_flags;
+	target_ulong ss_size;
+} target_stack_t;
+
+struct target_ucontext {
+    target_ulong uc_flags;
+    target_ulong uc_link;
+    target_stack_t uc_stack;
+    struct target_sigcontext uc_mcontext;
+    target_sigset_t  uc_sigmask;	/* mask last for extensibility */
+};
+
+struct sigframe
+{
+    struct target_sigcontext sc;
+    target_ulong extramask[TARGET_NSIG_WORDS-1];
+    target_ulong retcode;
+};
+
+struct rt_sigframe
+{
+    struct target_siginfo *pinfo;
+    void *puc;
+    struct target_siginfo info;
+    struct target_ucontext uc;
+    target_ulong retcode;
+};
+
+#define TARGET_CONFIG_CPU_32 1
+
+/*
+ * For ARM syscalls, we encode the syscall number into the instruction.
+ */
+#define SWI_SYS_SIGRETURN	(0xef000000|(TARGET_NR_sigreturn + ARM_SYSCALL_BASE))
+#define SWI_SYS_RT_SIGRETURN	(0xef000000|(TARGET_NR_rt_sigreturn + ARM_SYSCALL_BASE))
+
+/*
+ * For Thumb syscalls, we pass the syscall number via r7.  We therefore
+ * need two 16-bit instructions.
+ */
+#define SWI_THUMB_SIGRETURN	(0xdf00 << 16 | 0x2700 | (TARGET_NR_sigreturn))
+#define SWI_THUMB_RT_SIGRETURN	(0xdf00 << 16 | 0x2700 | (TARGET_NR_rt_sigreturn))
+
+static const target_ulong retcodes[4] = {
+	SWI_SYS_SIGRETURN,	SWI_THUMB_SIGRETURN,
+	SWI_SYS_RT_SIGRETURN,	SWI_THUMB_RT_SIGRETURN
+};
+
+
+#define __put_user_error(x,p,e) __put_user(x, p)
+#define __get_user_error(x,p,e) __get_user(x, p)
+
+static inline int valid_user_regs(CPUState *regs)
+{
+    return 1;
+}
+
+static int
+setup_sigcontext(struct target_sigcontext *sc, /*struct _fpstate *fpstate,*/
+		 CPUState *env, unsigned long mask)
+{
+	int err = 0;
+
+	__put_user_error(env->regs[0], &sc->arm_r0, err);
+	__put_user_error(env->regs[1], &sc->arm_r1, err);
+	__put_user_error(env->regs[2], &sc->arm_r2, err);
+	__put_user_error(env->regs[3], &sc->arm_r3, err);
+	__put_user_error(env->regs[4], &sc->arm_r4, err);
+	__put_user_error(env->regs[5], &sc->arm_r5, err);
+	__put_user_error(env->regs[6], &sc->arm_r6, err);
+	__put_user_error(env->regs[7], &sc->arm_r7, err);
+	__put_user_error(env->regs[8], &sc->arm_r8, err);
+	__put_user_error(env->regs[9], &sc->arm_r9, err);
+	__put_user_error(env->regs[10], &sc->arm_r10, err);
+	__put_user_error(env->regs[11], &sc->arm_fp, err);
+	__put_user_error(env->regs[12], &sc->arm_ip, err);
+	__put_user_error(env->regs[13], &sc->arm_sp, err);
+	__put_user_error(env->regs[14], &sc->arm_lr, err);
+	__put_user_error(env->regs[15], &sc->arm_pc, err);
+#ifdef TARGET_CONFIG_CPU_32
+	__put_user_error(env->cpsr, &sc->arm_cpsr, err);
+#endif
+
+	__put_user_error(/* current->thread.trap_no */ 0, &sc->trap_no, err);
+	__put_user_error(/* current->thread.error_code */ 0, &sc->error_code, err);
+	__put_user_error(/* current->thread.address */ 0, &sc->fault_address, err);
+	__put_user_error(mask, &sc->oldmask, err);
+
+	return err;
+}
+
+static inline void *
+get_sigframe(struct emulated_sigaction *ka, CPUState *regs, int framesize)
+{
+	unsigned long sp = regs->regs[13];
+
+#if 0
+	/*
+	 * This is the X/Open sanctioned signal stack switching.
+	 */
+	if ((ka->sa.sa_flags & SA_ONSTACK) && !sas_ss_flags(sp))
+		sp = current->sas_ss_sp + current->sas_ss_size;
+#endif
+	/*
+	 * ATPCS B01 mandates 8-byte alignment
+	 */
+	return (void *)((sp - framesize) & ~7);
+}
+
+static int
+setup_return(CPUState *env, struct emulated_sigaction *ka,
+	     target_ulong *rc, void *frame, int usig)
+{
+	target_ulong handler = (target_ulong)ka->sa._sa_handler;
+	target_ulong retcode;
+	int thumb = 0;
+#if defined(TARGET_CONFIG_CPU_32)
+	target_ulong cpsr = env->cpsr;
+
+#if 0
+	/*
+	 * Maybe we need to deliver a 32-bit signal to a 26-bit task.
+	 */
+	if (ka->sa.sa_flags & SA_THIRTYTWO)
+		cpsr = (cpsr & ~MODE_MASK) | USR_MODE;
+
+#ifdef CONFIG_ARM_THUMB
+	if (elf_hwcap & HWCAP_THUMB) {
+		/*
+		 * The LSB of the handler determines if we're going to
+		 * be using THUMB or ARM mode for this signal handler.
+		 */
+		thumb = handler & 1;
+
+		if (thumb)
+			cpsr |= T_BIT;
+		else
+			cpsr &= ~T_BIT;
+	}
+#endif
+#endif
+#endif /* TARGET_CONFIG_CPU_32 */
+
+	if (ka->sa.sa_flags & TARGET_SA_RESTORER) {
+		retcode = (target_ulong)ka->sa.sa_restorer;
+	} else {
+		unsigned int idx = thumb;
+
+		if (ka->sa.sa_flags & TARGET_SA_SIGINFO)
+			idx += 2;
+
+		if (__put_user(retcodes[idx], rc))
+			return 1;
+#if 0
+		flush_icache_range((target_ulong)rc,
+				   (target_ulong)(rc + 1));
+#endif
+		retcode = ((target_ulong)rc) + thumb;
+	}
+
+	env->regs[0] = usig;
+	env->regs[13] = (target_ulong)frame;
+	env->regs[14] = retcode;
+	env->regs[15] = handler & (thumb ? ~1 : ~3);
+
+#ifdef TARGET_CONFIG_CPU_32
+	env->cpsr = cpsr;
+#endif
+
+	return 0;
+}
+
+static void setup_frame(int usig, struct emulated_sigaction *ka,
+			target_sigset_t *set, CPUState *regs)
+{
+	struct sigframe *frame = get_sigframe(ka, regs, sizeof(*frame));
+	int err = 0;
+
+	err |= setup_sigcontext(&frame->sc, /*&frame->fpstate,*/ regs, set->sig[0]);
+
+	if (TARGET_NSIG_WORDS > 1) {
+		err |= __copy_to_user(frame->extramask, &set->sig[1],
+				      sizeof(frame->extramask));
+	}
+
+	if (err == 0)
+            err = setup_return(regs, ka, &frame->retcode, frame, usig);
+        //	return err;
+}
+
+static void setup_rt_frame(int usig, struct emulated_sigaction *ka, 
+                           target_siginfo_t *info,
+			   target_sigset_t *set, CPUState *env)
+{
+	struct rt_sigframe *frame = get_sigframe(ka, env, sizeof(*frame));
+	int err = 0;
+
+#if 0
+	if (!access_ok(VERIFY_WRITE, frame, sizeof (*frame)))
+            return 1;
+#endif
+	__put_user_error(&frame->info, (target_ulong *)&frame->pinfo, err);
+	__put_user_error(&frame->uc, (target_ulong *)&frame->puc, err);
+	err |= copy_siginfo_to_user(&frame->info, info);
+
+	/* Clear all the bits of the ucontext we don't use.  */
+	err |= __clear_user(&frame->uc, offsetof(struct ucontext, uc_mcontext));
+
+	err |= setup_sigcontext(&frame->uc.uc_mcontext, /*&frame->fpstate,*/
+				env, set->sig[0]);
+	err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set));
+
+	if (err == 0)
+		err = setup_return(env, ka, &frame->retcode, frame, usig);
+
+	if (err == 0) {
+		/*
+		 * For realtime signals we must also set the second and third
+		 * arguments for the signal handler.
+		 *   -- Peter Maydell <pmaydell@chiark.greenend.org.uk> 2000-12-06
+		 */
+            env->regs[1] = (target_ulong)frame->pinfo;
+            env->regs[2] = (target_ulong)frame->puc;
+	}
+
+        //	return err;
+}
+
+static int
+restore_sigcontext(CPUState *env, struct target_sigcontext *sc)
+{
+	int err = 0;
+
+	__get_user_error(env->regs[0], &sc->arm_r0, err);
+	__get_user_error(env->regs[1], &sc->arm_r1, err);
+	__get_user_error(env->regs[2], &sc->arm_r2, err);
+	__get_user_error(env->regs[3], &sc->arm_r3, err);
+	__get_user_error(env->regs[4], &sc->arm_r4, err);
+	__get_user_error(env->regs[5], &sc->arm_r5, err);
+	__get_user_error(env->regs[6], &sc->arm_r6, err);
+	__get_user_error(env->regs[7], &sc->arm_r7, err);
+	__get_user_error(env->regs[8], &sc->arm_r8, err);
+	__get_user_error(env->regs[9], &sc->arm_r9, err);
+	__get_user_error(env->regs[10], &sc->arm_r10, err);
+	__get_user_error(env->regs[11], &sc->arm_fp, err);
+	__get_user_error(env->regs[12], &sc->arm_ip, err);
+	__get_user_error(env->regs[13], &sc->arm_sp, err);
+	__get_user_error(env->regs[14], &sc->arm_lr, err);
+	__get_user_error(env->regs[15], &sc->arm_pc, err);
+#ifdef TARGET_CONFIG_CPU_32
+	__get_user_error(env->cpsr, &sc->arm_cpsr, err);
+#endif
+
+	err |= !valid_user_regs(env);
+
+	return err;
+}
+
+long do_sigreturn(CPUState *env)
+{
+	struct sigframe *frame;
+	target_sigset_t set;
+        sigset_t host_set;
+
+	/*
+	 * Since we stacked the signal on a 64-bit boundary,
+	 * then 'sp' should be word aligned here.  If it's
+	 * not, then the user is trying to mess with us.
+	 */
+	if (env->regs[13] & 7)
+		goto badframe;
+
+	frame = (struct sigframe *)env->regs[13];
+
+#if 0
+	if (verify_area(VERIFY_READ, frame, sizeof (*frame)))
+		goto badframe;
+#endif
+	if (__get_user(set.sig[0], &frame->sc.oldmask)
+	    || (TARGET_NSIG_WORDS > 1
+	        && __copy_from_user(&set.sig[1], &frame->extramask,
+				    sizeof(frame->extramask))))
+		goto badframe;
+
+        target_to_host_sigset(&host_set, &set);
+        sigprocmask(SIG_SETMASK, &host_set, NULL);
+
+	if (restore_sigcontext(env, &frame->sc))
+		goto badframe;
+
+#if 0
+	/* Send SIGTRAP if we're single-stepping */
+	if (ptrace_cancel_bpt(current))
+		send_sig(SIGTRAP, current, 1);
+#endif
+	return env->regs[0];
+
+badframe:
+        force_sig(SIGSEGV /* , current */);
+	return 0;
+}
+
+long do_rt_sigreturn(CPUState *env)
+{
+	struct rt_sigframe *frame;
+	target_sigset_t set;
+        sigset_t host_set;
+
+	/*
+	 * Since we stacked the signal on a 64-bit boundary,
+	 * then 'sp' should be word aligned here.  If it's
+	 * not, then the user is trying to mess with us.
+	 */
+	if (env->regs[13] & 7)
+		goto badframe;
+
+	frame = (struct rt_sigframe *)env->regs[13];
+
+#if 0
+	if (verify_area(VERIFY_READ, frame, sizeof (*frame)))
+		goto badframe;
+#endif
+	if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set)))
+		goto badframe;
+
+        target_to_host_sigset(&host_set, &set);
+        sigprocmask(SIG_SETMASK, &host_set, NULL);
+
+	if (restore_sigcontext(env, &frame->uc.uc_mcontext))
+		goto badframe;
+
+#if 0
+	/* Send SIGTRAP if we're single-stepping */
+	if (ptrace_cancel_bpt(current))
+		send_sig(SIGTRAP, current, 1);
+#endif
+	return env->regs[0];
+
+badframe:
+        force_sig(SIGSEGV /* , current */);
+	return 0;
+}
+
 #else

 static void setup_frame(int sig, struct emulated_sigaction *ka,
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -68,6 +68,129 @@
 #define	VFAT_IOCTL_READDIR_BOTH		_IOR('r', 1, struct dirent [2])
 #define	VFAT_IOCTL_READDIR_SHORT	_IOR('r', 2, struct dirent [2])

+
+#if defined(__powerpc__)
+#undef __syscall_nr
+#undef __sc_loadargs_0
+#undef __sc_loadargs_1
+#undef __sc_loadargs_2
+#undef __sc_loadargs_3
+#undef __sc_loadargs_4
+#undef __sc_loadargs_5
+#undef __sc_asm_input_0
+#undef __sc_asm_input_1
+#undef __sc_asm_input_2
+#undef __sc_asm_input_3
+#undef __sc_asm_input_4
+#undef __sc_asm_input_5
+#undef _syscall0
+#undef _syscall1
+#undef _syscall2
+#undef _syscall3
+#undef _syscall4
+#undef _syscall5
+
+/* need to redefine syscalls as Linux kernel defines are incorrect for
+   the clobber list */
+/* On powerpc a system call basically clobbers the same registers like a
+ * function call, with the exception of LR (which is needed for the
+ * "sc; bnslr" sequence) and CR (where only CR0.SO is clobbered to signal
+ * an error return status).
+ */
+
+#define __syscall_nr(nr, type, name, args...)				\
+	unsigned long __sc_ret, __sc_err;				\
+	{								\
+		register unsigned long __sc_0  __asm__ ("r0");		\
+		register unsigned long __sc_3  __asm__ ("r3");		\
+		register unsigned long __sc_4  __asm__ ("r4");		\
+		register unsigned long __sc_5  __asm__ ("r5");		\
+		register unsigned long __sc_6  __asm__ ("r6");		\
+		register unsigned long __sc_7  __asm__ ("r7");		\
+									\
+		__sc_loadargs_##nr(name, args);				\
+		__asm__ __volatile__					\
+			("sc           \n\t"				\
+			 "mfcr %0      "				\
+			: "=&r" (__sc_0),				\
+			  "=&r" (__sc_3),  "=&r" (__sc_4),		\
+			  "=&r" (__sc_5),  "=&r" (__sc_6),		\
+			  "=&r" (__sc_7)				\
+			: __sc_asm_input_##nr				\
+			: "cr0", "ctr", "memory",			\
+			  "r8", "r9", "r10","r11", "r12");		\
+		__sc_ret = __sc_3;					\
+		__sc_err = __sc_0;					\
+	}								\
+	if (__sc_err & 0x10000000)					\
+	{								\
+		errno = __sc_ret;					\
+		__sc_ret = -1;						\
+	}								\
+	return (type) __sc_ret
+
+#define __sc_loadargs_0(name, dummy...)					\
+	__sc_0 = __NR_##name
+#define __sc_loadargs_1(name, arg1)					\
+	__sc_loadargs_0(name);						\
+	__sc_3 = (unsigned long) (arg1)
+#define __sc_loadargs_2(name, arg1, arg2)				\
+	__sc_loadargs_1(name, arg1);					\
+	__sc_4 = (unsigned long) (arg2)
+#define __sc_loadargs_3(name, arg1, arg2, arg3)				\
+	__sc_loadargs_2(name, arg1, arg2);				\
+	__sc_5 = (unsigned long) (arg3)
+#define __sc_loadargs_4(name, arg1, arg2, arg3, arg4)			\
+	__sc_loadargs_3(name, arg1, arg2, arg3);			\
+	__sc_6 = (unsigned long) (arg4)
+#define __sc_loadargs_5(name, arg1, arg2, arg3, arg4, arg5)		\
+	__sc_loadargs_4(name, arg1, arg2, arg3, arg4);			\
+	__sc_7 = (unsigned long) (arg5)
+
+#define __sc_asm_input_0 "0" (__sc_0)
+#define __sc_asm_input_1 __sc_asm_input_0, "1" (__sc_3)
+#define __sc_asm_input_2 __sc_asm_input_1, "2" (__sc_4)
+#define __sc_asm_input_3 __sc_asm_input_2, "3" (__sc_5)
+#define __sc_asm_input_4 __sc_asm_input_3, "4" (__sc_6)
+#define __sc_asm_input_5 __sc_asm_input_4, "5" (__sc_7)
+
+#define _syscall0(type,name)						\
+type name(void)								\
+{									\
+	__syscall_nr(0, type, name);					\
+}
+
+#define _syscall1(type,name,type1,arg1)					\
+type name(type1 arg1)							\
+{									\
+	__syscall_nr(1, type, name, arg1);				\
+}
+
+#define _syscall2(type,name,type1,arg1,type2,arg2)			\
+type name(type1 arg1, type2 arg2)					\
+{									\
+	__syscall_nr(2, type, name, arg1, arg2);			\
+}
+
+#define _syscall3(type,name,type1,arg1,type2,arg2,type3,arg3)		\
+type name(type1 arg1, type2 arg2, type3 arg3)				\
+{									\
+	__syscall_nr(3, type, name, arg1, arg2, arg3);			\
+}
+
+#define _syscall4(type,name,type1,arg1,type2,arg2,type3,arg3,type4,arg4) \
+type name(type1 arg1, type2 arg2, type3 arg3, type4 arg4)		\
+{									\
+	__syscall_nr(4, type, name, arg1, arg2, arg3, arg4);		\
+}
+
+#define _syscall5(type,name,type1,arg1,type2,arg2,type3,arg3,type4,arg4,type5,arg5) \
+type name(type1 arg1, type2 arg2, type3 arg3, type4 arg4, type5 arg5)	\
+{									\
+	__syscall_nr(5, type, name, arg1, arg2, arg3, arg4, arg5);	\
+}
+#endif
+
 #define __NR_sys_uname __NR_uname
 #define __NR_sys_getcwd1 __NR_getcwd
 #define __NR_sys_statfs __NR_statfs
@@ -210,6 +333,21 @@ static inline void host_to_target_fds(target_long *target_fds,
 #endif
 }

+#if defined(__alpha__)
+#define HOST_HZ 1024
+#else
+#define HOST_HZ 100
+#endif
+
+static inline long host_to_target_clock_t(long ticks)
+{
+#if HOST_HZ == TARGET_HZ
+    return ticks;
+#else
+    return ((int64_t)ticks * TARGET_HZ) / HOST_HZ;
+#endif
+}
+
 static inline void host_to_target_rusage(struct target_rusage *target_rusage, 
                                         const struct rusage *rusage)
 {
@@ -1423,11 +1561,13 @@ long do_syscall(void *cpu_env, int num, long arg1, long arg2, long arg3,
            struct tms tms;
            ret = get_errno(times(&tms));
            if (tmsp) {
-                tmsp->tms_utime = tswapl(tms.tms_utime);
-                tmsp->tms_stime = tswapl(tms.tms_stime);
-                tmsp->tms_cutime = tswapl(tms.tms_cutime);
-                tmsp->tms_cstime = tswapl(tms.tms_cstime);
+                tmsp->tms_utime = tswapl(host_to_target_clock_t(tms.tms_utime));
+                tmsp->tms_stime = tswapl(host_to_target_clock_t(tms.tms_stime));
+                tmsp->tms_cutime = tswapl(host_to_target_clock_t(tms.tms_cutime));
+                tmsp->tms_cstime = tswapl(host_to_target_clock_t(tms.tms_cstime));
            }
+            if (!is_error(ret))
+                ret = host_to_target_clock_t(ret);
        }
        break;
    case TARGET_NR_prof:
@@ -1763,7 +1903,17 @@ long do_syscall(void *cpu_env, int num, long arg1, long arg2, long arg3,
        }
        break;
    case TARGET_NR_select:
-        goto unimplemented;
+        {
+            struct target_sel_arg_struct *sel = (void *)arg1;
+            sel->n = tswapl(sel->n);
+            sel->inp = tswapl(sel->inp);
+            sel->outp = tswapl(sel->outp);
+            sel->exp = tswapl(sel->exp);
+            sel->tvp = tswapl(sel->tvp);
+            ret = do_select(sel->n, (void *)sel->inp, (void *)sel->outp,
+                            (void *)sel->exp, (void *)sel->tvp);
+        }
+        break;
    case TARGET_NR_symlink:
        ret = get_errno(symlink((const char *)arg1, (const char *)arg2));
        break;
@@ -1781,8 +1931,8 @@ long do_syscall(void *cpu_env, int num, long arg1, long arg2, long arg3,
        goto unimplemented;
    case TARGET_NR_readdir:
        goto unimplemented;
-#ifdef TARGET_I386
    case TARGET_NR_mmap:
+#if defined(TARGET_I386) || defined(TARGET_ARM)
        {
            uint32_t v1, v2, v3, v4, v5, v6, *vptr;
            vptr = (uint32_t *)arg1;
@@ -1796,13 +1946,14 @@ long do_syscall(void *cpu_env, int num, long arg1, long arg2, long arg3,
                                        target_to_host_bitmask(v4, mmap_flags_tbl),
                                        v5, v6));
        }
-        break;
-#endif
-#ifdef TARGET_I386
-    case TARGET_NR_mmap2:
 #else
-    case TARGET_NR_mmap:
+        ret = get_errno(target_mmap(arg1, arg2, arg3, 
+                                    target_to_host_bitmask(arg4, mmap_flags_tbl), 
+                                    arg5,
+                                    arg6));
 #endif
+        break;
+    case TARGET_NR_mmap2:
        ret = get_errno(target_mmap(arg1, arg2, arg3, 
                                    target_to_host_bitmask(arg4, mmap_flags_tbl), 
                                    arg5,
--- a/linux-user/syscall_defs.h
+++ b/linux-user/syscall_defs.h
@@ -383,6 +383,8 @@ struct target_itimerval {

 typedef target_long target_clock_t;

+#define TARGET_HZ 100
+
 struct target_tms {
    target_clock_t tms_utime;
    target_clock_t tms_stime;
@@ -390,6 +392,12 @@ struct target_tms {
    target_clock_t tms_cstime;
 };

+struct target_sel_arg_struct {
+    target_long n;
+    target_long inp, outp, exp;
+    target_long tvp;
+};
+
 struct target_iovec {
    target_long iov_base;   /* Starting address */
    target_long iov_len;   /* Number of bytes */
@@ -533,8 +541,8 @@ static inline void target_siginitset(target_sigset_t *d, target_ulong set)
        d->sig[i] = 0;
 }

-void host_to_target_sigset(target_sigset_t *d, sigset_t *s);
-void target_to_host_sigset(sigset_t *d, target_sigset_t *s);
+void host_to_target_sigset(target_sigset_t *d, const sigset_t *s);
+void target_to_host_sigset(sigset_t *d, const target_sigset_t *s);
 void host_to_target_old_sigset(target_ulong *old_sigset, 
                               const sigset_t *sigset);
 void target_to_host_old_sigset(sigset_t *sigset, 
@@ -584,6 +592,8 @@ int do_sigaction(int sig, const struct target_sigaction *act,
 #define TARGET_SIGPROF		27
 #define TARGET_SIGWINCH	        28
 #define TARGET_SIGIO		29
+#define TARGET_SIGPWR		30
+#define TARGET_SIGSYS		31
 #define TARGET_SIGRTMIN         32

 #define TARGET_SIG_BLOCK          0    /* for blocking signals */
--- a/linux-user/vm86.c
+++ b/linux-user/vm86.c
@@ -73,17 +73,17 @@ void save_v86_state(CPUX86State *env)
    ts->target_v86->regs.ebp = tswap32(env->regs[R_EBP]);
    ts->target_v86->regs.esp = tswap32(env->regs[R_ESP]);
    ts->target_v86->regs.eip = tswap32(env->eip);
-    ts->target_v86->regs.cs = tswap16(env->segs[R_CS]);
-    ts->target_v86->regs.ss = tswap16(env->segs[R_SS]);
-    ts->target_v86->regs.ds = tswap16(env->segs[R_DS]);
-    ts->target_v86->regs.es = tswap16(env->segs[R_ES]);
-    ts->target_v86->regs.fs = tswap16(env->segs[R_FS]);
-    ts->target_v86->regs.gs = tswap16(env->segs[R_GS]);
+    ts->target_v86->regs.cs = tswap16(env->segs[R_CS].selector);
+    ts->target_v86->regs.ss = tswap16(env->segs[R_SS].selector);
+    ts->target_v86->regs.ds = tswap16(env->segs[R_DS].selector);
+    ts->target_v86->regs.es = tswap16(env->segs[R_ES].selector);
+    ts->target_v86->regs.fs = tswap16(env->segs[R_FS].selector);
+    ts->target_v86->regs.gs = tswap16(env->segs[R_GS].selector);
    set_flags(env->eflags, ts->v86flags, VIF_MASK | ts->v86mask);
    ts->target_v86->regs.eflags = tswap32(env->eflags);
 #ifdef DEBUG_VM86
    fprintf(logfile, "save_v86_state: eflags=%08x cs:ip=%04x:%04x\n", 
-            env->eflags, env->segs[R_CS], env->eip);
+            env->eflags, env->segs[R_CS].selector, env->eip);
 #endif

    /* restore 32 bit registers */
@@ -180,6 +180,7 @@ static inline unsigned int get_vflags(CPUX86State *env)
    flags = env->eflags & RETURN_MASK;
    if (ts->v86flags & VIF_MASK)
        flags |= IF_MASK;
+    flags |= IOPL_MASK;
    return flags | (ts->v86flags & ts->v86mask);
 }

@@ -194,7 +195,7 @@ static void do_int(CPUX86State *env, int intno)
    uint8_t *ssp;
    unsigned int sp;

-    if (env->segs[R_CS] == TARGET_BIOSSEG)
+    if (env->segs[R_CS].selector == TARGET_BIOSSEG)
        goto cannot_handle;
    if (is_revectored(intno, &ts->vm86plus.int_revectored))
        goto cannot_handle;
@@ -210,10 +211,10 @@ static void do_int(CPUX86State *env, int intno)
            intno, segoffs >> 16, segoffs & 0xffff);
 #endif
    /* save old state */
-    ssp = (uint8_t *)(env->segs[R_SS] << 4);
+    ssp = (uint8_t *)(env->segs[R_SS].selector << 4);
    sp = env->regs[R_ESP] & 0xffff;
    vm_putw(ssp, sp - 2, get_vflags(env));
-    vm_putw(ssp, sp - 4, env->segs[R_CS]);
+    vm_putw(ssp, sp - 4, env->segs[R_CS].selector);
    vm_putw(ssp, sp - 6, env->eip);
    ADD16(env->regs[R_ESP], -6);
    /* goto interrupt handler */
@@ -257,16 +258,16 @@ void handle_vm86_fault(CPUX86State *env)
    unsigned int ip, sp, newflags, newip, newcs, opcode, intno;
    int data32, pref_done;

-    csp = (uint8_t *)(env->segs[R_CS] << 4);
+    csp = (uint8_t *)(env->segs[R_CS].selector << 4);
    ip = env->eip & 0xffff;
    pc = csp + ip;
    
-    ssp = (uint8_t *)(env->segs[R_SS] << 4);
+    ssp = (uint8_t *)(env->segs[R_SS].selector << 4);
    sp = env->regs[R_ESP] & 0xffff;

 #if defined(DEBUG_VM86)
    fprintf(logfile, "VM86 exception %04x:%08x %02x %02x\n",
-            env->segs[R_CS], env->eip, pc[0], pc[1]);
+            env->segs[R_CS].selector, env->eip, pc[0], pc[1]);
 #endif

    data32 = 0;
@@ -413,12 +414,12 @@ int do_vm86(CPUX86State *env, long subfunction,
    ts->vm86_saved_regs.esp = env->regs[R_ESP];
    ts->vm86_saved_regs.eflags = env->eflags;
    ts->vm86_saved_regs.eip  = env->eip;
-    ts->vm86_saved_regs.cs = env->segs[R_CS];
-    ts->vm86_saved_regs.ss = env->segs[R_SS];
-    ts->vm86_saved_regs.ds = env->segs[R_DS];
-    ts->vm86_saved_regs.es = env->segs[R_ES];
-    ts->vm86_saved_regs.fs = env->segs[R_FS];
-    ts->vm86_saved_regs.gs = env->segs[R_GS];
+    ts->vm86_saved_regs.cs = env->segs[R_CS].selector;
+    ts->vm86_saved_regs.ss = env->segs[R_SS].selector;
+    ts->vm86_saved_regs.ds = env->segs[R_DS].selector;
+    ts->vm86_saved_regs.es = env->segs[R_ES].selector;
+    ts->vm86_saved_regs.fs = env->segs[R_FS].selector;
+    ts->vm86_saved_regs.gs = env->segs[R_GS].selector;

    /* build vm86 CPU state */
    ts->v86flags = tswap32(target_v86->regs.eflags);
@@ -466,7 +467,8 @@ int do_vm86(CPUX86State *env, long subfunction,
           target_v86->vm86plus.vm86dbg_intxxtab, 32);
    
 #ifdef DEBUG_VM86
-    fprintf(logfile, "do_vm86: cs:ip=%04x:%04x\n", env->segs[R_CS], env->eip);
+    fprintf(logfile, "do_vm86: cs:ip=%04x:%04x\n", 
+            env->segs[R_CS].selector, env->eip);
 #endif
    /* now the virtual CPU is ready for vm86 execution ! */
 out:
--- a/m68k.ld
+++ b/m68k.ld
@@ -0,0 +1,177 @@
+/* Script for -z combreloc: combine and sort reloc sections */
+OUTPUT_FORMAT("elf32-m68k", "elf32-m68k",
+	      "elf32-m68k")
+OUTPUT_ARCH(m68k)
+ENTRY(_start)
+SEARCH_DIR("/usr/local/m68k-linux/lib");
+/* Do we need any of these for elf?
+   __DYNAMIC = 0;    */
+SECTIONS
+{
+  /* Read-only sections, merged into text segment: */
+  . = 0x60000000 + SIZEOF_HEADERS;
+  .interp         : { *(.interp) }
+  .hash           : { *(.hash) }
+  .dynsym         : { *(.dynsym) }
+  .dynstr         : { *(.dynstr) }
+  .gnu.version    : { *(.gnu.version) }
+  .gnu.version_d  : { *(.gnu.version_d) }
+  .gnu.version_r  : { *(.gnu.version_r) }
+  .rel.dyn        :
+    {
+      *(.rel.init)
+      *(.rel.text .rel.text.* .rel.gnu.linkonce.t.*)
+      *(.rel.fini)
+      *(.rel.rodata .rel.rodata.* .rel.gnu.linkonce.r.*)
+      *(.rel.data .rel.data.* .rel.gnu.linkonce.d.*)
+      *(.rel.tdata .rel.tdata.* .rel.gnu.linkonce.td.*)
+      *(.rel.tbss .rel.tbss.* .rel.gnu.linkonce.tb.*)
+      *(.rel.ctors)
+      *(.rel.dtors)
+      *(.rel.got)
+      *(.rel.bss .rel.bss.* .rel.gnu.linkonce.b.*)
+    }
+  .rela.dyn       :
+    {
+      *(.rela.init)
+      *(.rela.text .rela.text.* .rela.gnu.linkonce.t.*)
+      *(.rela.fini)
+      *(.rela.rodata .rela.rodata.* .rela.gnu.linkonce.r.*)
+      *(.rela.data .rela.data.* .rela.gnu.linkonce.d.*)
+      *(.rela.tdata .rela.tdata.* .rela.gnu.linkonce.td.*)
+      *(.rela.tbss .rela.tbss.* .rela.gnu.linkonce.tb.*)
+      *(.rela.ctors)
+      *(.rela.dtors)
+      *(.rela.got)
+      *(.rela.bss .rela.bss.* .rela.gnu.linkonce.b.*)
+    }
+  .rel.plt        : { *(.rel.plt) }
+  .rela.plt       : { *(.rela.plt) }
+  .init           :
+  {
+    KEEP (*(.init))
+  } =0x4e754e75
+  .plt            : { *(.plt) }
+  .text           :
+  {
+    *(.text .stub .text.* .gnu.linkonce.t.*)
+    /* .gnu.warning sections are handled specially by elf32.em.  */
+    *(.gnu.warning)
+  } =0x4e754e75
+  .fini           :
+  {
+    KEEP (*(.fini))
+  } =0x4e754e75
+  PROVIDE (__etext = .);
+  PROVIDE (_etext = .);
+  PROVIDE (etext = .);
+  .rodata         : { *(.rodata .rodata.* .gnu.linkonce.r.*) }
+  .rodata1        : { *(.rodata1) }
+  .eh_frame_hdr : { *(.eh_frame_hdr) }
+  /* Adjust the address for the data segment.  We want to adjust up to
+     the same address within the page on the next page up.  */
+  . = ALIGN(0x2000) + (. & (0x2000 - 1));
+  /* Ensure the __preinit_array_start label is properly aligned.  We
+     could instead move the label definition inside the section, but
+     the linker would then create the section even if it turns out to
+     be empty, which isn't pretty.  */
+  . = ALIGN(32 / 8);
+  PROVIDE (__preinit_array_start = .);
+  .preinit_array     : { *(.preinit_array) }
+  PROVIDE (__preinit_array_end = .);
+  PROVIDE (__init_array_start = .);
+  .init_array     : { *(.init_array) }
+  PROVIDE (__init_array_end = .);
+  PROVIDE (__fini_array_start = .);
+  .fini_array     : { *(.fini_array) }
+  PROVIDE (__fini_array_end = .);
+  .data           :
+  {
+    *(.data .data.* .gnu.linkonce.d.*)
+    SORT(CONSTRUCTORS)
+  }
+  .data1          : { *(.data1) }
+  .tdata	  : { *(.tdata .tdata.* .gnu.linkonce.td.*) }
+  .tbss		  : { *(.tbss .tbss.* .gnu.linkonce.tb.*) *(.tcommon) }
+  .eh_frame       : { KEEP (*(.eh_frame)) }
+  .gcc_except_table   : { *(.gcc_except_table) }
+  .dynamic        : { *(.dynamic) }
+  .ctors          :
+  {
+    /* gcc uses crtbegin.o to find the start of
+       the constructors, so we make sure it is
+       first.  Because this is a wildcard, it
+       doesn't matter if the user does not
+       actually link against crtbegin.o; the
+       linker won't look for a file to match a
+       wildcard.  The wildcard also means that it
+       doesn't matter which directory crtbegin.o
+       is in.  */
+    KEEP (*crtbegin.o(.ctors))
+    /* We don't want to include the .ctor section from
+       from the crtend.o file until after the sorted ctors.
+       The .ctor section from the crtend file contains the
+       end of ctors marker and it must be last */
+    KEEP (*(EXCLUDE_FILE (*crtend.o ) .ctors))
+    KEEP (*(SORT(.ctors.*)))
+    KEEP (*(.ctors))
+  }
+  .dtors          :
+  {
+    KEEP (*crtbegin.o(.dtors))
+    KEEP (*(EXCLUDE_FILE (*crtend.o ) .dtors))
+    KEEP (*(SORT(.dtors.*)))
+    KEEP (*(.dtors))
+  }
+  .jcr            : { KEEP (*(.jcr)) }
+  .got            : { *(.got.plt) *(.got) }
+  _edata = .;
+  PROVIDE (edata = .);
+  __bss_start = .;
+  .bss            :
+  {
+   *(.dynbss)
+   *(.bss .bss.* .gnu.linkonce.b.*)
+   *(COMMON)
+   /* Align here to ensure that the .bss section occupies space up to
+      _end.  Align after .bss to ensure correct alignment even if the
+      .bss section disappears because there are no input sections.  */
+   . = ALIGN(32 / 8);
+  }
+  . = ALIGN(32 / 8);
+  _end = .;
+  PROVIDE (end = .);
+  /* Stabs debugging sections.  */
+  .stab          0 : { *(.stab) }
+  .stabstr       0 : { *(.stabstr) }
+  .stab.excl     0 : { *(.stab.excl) }
+  .stab.exclstr  0 : { *(.stab.exclstr) }
+  .stab.index    0 : { *(.stab.index) }
+  .stab.indexstr 0 : { *(.stab.indexstr) }
+  .comment       0 : { *(.comment) }
+  /* DWARF debug sections.
+     Symbols in the DWARF debugging sections are relative to the beginning
+     of the section so we begin them at 0.  */
+  /* DWARF 1 */
+  .debug          0 : { *(.debug) }
+  .line           0 : { *(.line) }
+  /* GNU DWARF 1 extensions */
+  .debug_srcinfo  0 : { *(.debug_srcinfo) }
+  .debug_sfnames  0 : { *(.debug_sfnames) }
+  /* DWARF 1.1 and DWARF 2 */
+  .debug_aranges  0 : { *(.debug_aranges) }
+  .debug_pubnames 0 : { *(.debug_pubnames) }
+  /* DWARF 2 */
+  .debug_info     0 : { *(.debug_info .gnu.linkonce.wi.*) }
+  .debug_abbrev   0 : { *(.debug_abbrev) }
+  .debug_line     0 : { *(.debug_line) }
+  .debug_frame    0 : { *(.debug_frame) }
+  .debug_str      0 : { *(.debug_str) }
+  .debug_loc      0 : { *(.debug_loc) }
+  .debug_macinfo  0 : { *(.debug_macinfo) }
+  /* SGI/MIPS DWARF 2 extensions */
+  .debug_weaknames 0 : { *(.debug_weaknames) }
+  .debug_funcnames 0 : { *(.debug_funcnames) }
+  .debug_typenames 0 : { *(.debug_typenames) }
+  .debug_varnames  0 : { *(.debug_varnames) }
+}
--- a/op-arm.c
+++ b/op-arm.c
@@ -154,11 +154,11 @@ void OPPROTO op_adcl_T0_T1_cc(void)
    FORCE_RET();
 }

-#define OPSUB(sub, sbc, T0, T1)                 \
+#define OPSUB(sub, sbc, res, T0, T1)            \
                                                \
 void OPPROTO op_ ## sub ## l_T0_T1(void)        \
 {                                               \
-    T0 -= T1;                                   \
+    res = T0 - T1;                              \
 }                                               \
                                                \
 void OPPROTO op_ ## sub ## l_T0_T1_cc(void)     \
@@ -167,13 +167,14 @@ void OPPROTO op_ ## sub ## l_T0_T1_cc(void)     \
    src1 = T0;                                  \
    T0 -= T1;                                   \
    env->NZF = T0;                              \
-    env->CF = src1 < T1;                        \
+    env->CF = src1 >= T1;                       \
    env->VF = (src1 ^ T1) & (src1 ^ T0);        \
+    res = T0;                                   \
 }                                               \
                                                \
 void OPPROTO op_ ## sbc ## l_T0_T1(void)        \
 {                                               \
-    T0 = T0 - T1 + env->CF - 1;                 \
+    res = T0 - T1 + env->CF - 1;                \
 }                                               \
                                                \
 void OPPROTO op_ ## sbc ## l_T0_T1_cc(void)     \
@@ -182,20 +183,20 @@ void OPPROTO op_ ## sbc ## l_T0_T1_cc(void)     \
    src1 = T0;                                  \
    if (!env->CF) {                             \
        T0 = T0 - T1 - 1;                       \
-        T0 += T1;                               \
-        env->CF = src1 < T1;                    \
+        env->CF = src1 >= T1;                   \
    } else {                                    \
        T0 = T0 - T1;                           \
-        env->CF = src1 <= T1;                   \
+        env->CF = src1 > T1;                    \
    }                                           \
    env->VF = (src1 ^ T1) & (src1 ^ T0);        \
    env->NZF = T0;                              \
+    res = T0;                                   \
    FORCE_RET();                                \
 }

-OPSUB(sub, sbc, T0, T1)
+OPSUB(sub, sbc, T0, T0, T1)

-OPSUB(rsb, rsc, T1, T0)
+OPSUB(rsb, rsc, T0, T1, T0)

 void OPPROTO op_andl_T0_T1(void)
 {
@@ -222,122 +223,129 @@ void OPPROTO op_notl_T1(void)
    T1 = ~T1;
 }

-void OPPROTO op_logic_cc(void)
+void OPPROTO op_logic_T0_cc(void)
 {
    env->NZF = T0;
 }

+void OPPROTO op_logic_T1_cc(void)
+{
+    env->NZF = T1;
+}
+
 #define EIP (env->regs[15])

 void OPPROTO op_test_eq(void)
 {
    if (env->NZF == 0)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(op_test_eq, PARAM1, 0, PARAM2);
    FORCE_RET();
 }

 void OPPROTO op_test_ne(void)
 {
    if (env->NZF != 0)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(op_test_ne, PARAM1, 0, PARAM2);
    FORCE_RET();
 }

 void OPPROTO op_test_cs(void)
 {
    if (env->CF != 0)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(op_test_cs, PARAM1, 0, PARAM2);
    FORCE_RET();
 }

 void OPPROTO op_test_cc(void)
 {
    if (env->CF == 0)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(op_test_cc, PARAM1, 0, PARAM2);
    FORCE_RET();
 }

 void OPPROTO op_test_mi(void)
 {
    if ((env->NZF & 0x80000000) != 0)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(op_test_mi, PARAM1, 0, PARAM2);
    FORCE_RET();
 }

 void OPPROTO op_test_pl(void)
 {
    if ((env->NZF & 0x80000000) == 0)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(op_test_pl, PARAM1, 0, PARAM2);
    FORCE_RET();
 }

 void OPPROTO op_test_vs(void)
 {
    if ((env->VF & 0x80000000) != 0)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(op_test_vs, PARAM1, 0, PARAM2);
    FORCE_RET();
 }

 void OPPROTO op_test_vc(void)
 {
    if ((env->VF & 0x80000000) == 0)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(op_test_vc, PARAM1, 0, PARAM2);
    FORCE_RET();
 }

 void OPPROTO op_test_hi(void)
 {
    if (env->CF != 0 && env->NZF != 0)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(op_test_hi, PARAM1, 0, PARAM2);
    FORCE_RET();
 }

 void OPPROTO op_test_ls(void)
 {
    if (env->CF == 0 || env->NZF == 0)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(op_test_ls, PARAM1, 0, PARAM2);
    FORCE_RET();
 }

 void OPPROTO op_test_ge(void)
 {
    if (((env->VF ^ env->NZF) & 0x80000000) == 0)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(op_test_ge, PARAM1, 0, PARAM2);
    FORCE_RET();
 }

 void OPPROTO op_test_lt(void)
 {
    if (((env->VF ^ env->NZF) & 0x80000000) != 0)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(op_test_lt, PARAM1, 0, PARAM2);
    FORCE_RET();
 }

 void OPPROTO op_test_gt(void)
 {
    if (env->NZF != 0 && ((env->VF ^ env->NZF) & 0x80000000) == 0)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(op_test_gt, PARAM1, 0, PARAM2);
    FORCE_RET();
 }

 void OPPROTO op_test_le(void)
 {
    if (env->NZF == 0 || ((env->VF ^ env->NZF) & 0x80000000) != 0)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(op_test_le, PARAM1, 0, PARAM2);
    FORCE_RET();
 }

 void OPPROTO op_jmp(void)
 {
-    JUMP_TB(PARAM1, 1, PARAM2);
+    JUMP_TB(op_jmp, PARAM1, 1, PARAM2);
+}
+
+void OPPROTO op_exit_tb(void)
+{
+    EXIT_TB();
 }

 void OPPROTO op_movl_T0_psr(void)
 {
-    int ZF;
-    ZF = (env->NZF == 0);
-    T0 = env->cpsr | (env->NZF & 0x80000000) | (ZF << 30) | 
-        (env->CF << 29) | ((env->VF & 0x80000000) >> 3);
+    T0 = compute_cpsr();
 }

 /* NOTE: N = 1 and Z = 1 cannot be stored currently */
--- a/op-i386.c
+++ b/op-i386.c
@@ -80,62 +80,34 @@ static inline int lshift(int x, int n)

 /* operations with flags */

-void OPPROTO op_addl_T0_T1_cc(void)
+/* update flags with T0 and T1 (add/sub case) */
+void OPPROTO op_update2_cc(void)
 {
-    CC_SRC = T0;
-    T0 += T1;
+    CC_SRC = T1;
    CC_DST = T0;
 }

-void OPPROTO op_orl_T0_T1_cc(void)
+/* update flags with T0 (logic operation case) */
+void OPPROTO op_update1_cc(void)
 {
-    T0 |= T1;
    CC_DST = T0;
 }

-void OPPROTO op_andl_T0_T1_cc(void)
+void OPPROTO op_update_neg_cc(void)
 {
-    T0 &= T1;
-    CC_DST = T0;
-}
-
-void OPPROTO op_subl_T0_T1_cc(void)
-{
-    CC_SRC = T0;
-    T0 -= T1;
-    CC_DST = T0;
-}
-
-void OPPROTO op_xorl_T0_T1_cc(void)
-{
-    T0 ^= T1;
+    CC_SRC = -T0;
    CC_DST = T0;
 }

 void OPPROTO op_cmpl_T0_T1_cc(void)
 {
-    CC_SRC = T0;
+    CC_SRC = T1;
    CC_DST = T0 - T1;
 }

-void OPPROTO op_negl_T0_cc(void)
-{
-    CC_SRC = 0;
-    T0 = -T0;
-    CC_DST = T0;
-}
-
-void OPPROTO op_incl_T0_cc(void)
+void OPPROTO op_update_inc_cc(void)
 {
    CC_SRC = cc_table[CC_OP].compute_c();
-    T0++;
-    CC_DST = T0;
-}
-
-void OPPROTO op_decl_T0_cc(void)
-{
-    CC_SRC = cc_table[CC_OP].compute_c();
-    T0--;
    CC_DST = T0;
 }

@@ -357,6 +329,11 @@ void OPPROTO op_andl_T0_ffff(void)
    T0 = T0 & 0xffff;
 }

+void OPPROTO op_andl_T0_im(void)
+{
+    T0 = T0 & PARAM1;
+}
+
 void OPPROTO op_movl_T0_T1(void)
 {
    T0 = T1;
@@ -399,70 +376,14 @@ void OPPROTO op_andl_A0_ffff(void)

 /* memory access */

-void OPPROTO op_ldub_T0_A0(void)
-{
-    T0 = ldub((uint8_t *)A0);
-}
+#define MEMSUFFIX
+#include "ops_mem.h"

-void OPPROTO op_ldsb_T0_A0(void)
-{
-    T0 = ldsb((int8_t *)A0);
-}
+#define MEMSUFFIX _user
+#include "ops_mem.h"

-void OPPROTO op_lduw_T0_A0(void)
-{
-    T0 = lduw((uint8_t *)A0);
-}
-
-void OPPROTO op_ldsw_T0_A0(void)
-{
-    T0 = ldsw((int8_t *)A0);
-}
-
-void OPPROTO op_ldl_T0_A0(void)
-{
-    T0 = ldl((uint8_t *)A0);
-}
-
-void OPPROTO op_ldub_T1_A0(void)
-{
-    T1 = ldub((uint8_t *)A0);
-}
-
-void OPPROTO op_ldsb_T1_A0(void)
-{
-    T1 = ldsb((int8_t *)A0);
-}
-
-void OPPROTO op_lduw_T1_A0(void)
-{
-    T1 = lduw((uint8_t *)A0);
-}
-
-void OPPROTO op_ldsw_T1_A0(void)
-{
-    T1 = ldsw((int8_t *)A0);
-}
-
-void OPPROTO op_ldl_T1_A0(void)
-{
-    T1 = ldl((uint8_t *)A0);
-}
-
-void OPPROTO op_stb_T0_A0(void)
-{
-    stb((uint8_t *)A0, T0);
-}
-
-void OPPROTO op_stw_T0_A0(void)
-{
-    stw((uint8_t *)A0, T0);
-}
-
-void OPPROTO op_stl_T0_A0(void)
-{
-    stl((uint8_t *)A0, T0);
-}
+#define MEMSUFFIX _kernel
+#include "ops_mem.h"

 /* used for bit operations */

@@ -488,6 +409,18 @@ void OPPROTO op_jmp_im(void)
    EIP = PARAM1;
 }

+void OPPROTO op_hlt(void)
+{
+    env->exception_index = EXCP_HLT;
+    cpu_loop_exit();
+}
+
+void OPPROTO op_debug(void)
+{
+    env->exception_index = EXCP_DEBUG;
+    cpu_loop_exit();
+}
+
 void OPPROTO op_raise_interrupt(void)
 {
    int intno;
@@ -524,6 +457,16 @@ void OPPROTO op_sti(void)
    env->eflags |= IF_MASK;
 }

+void OPPROTO op_set_inhibit_irq(void)
+{
+    env->hflags |= HF_INHIBIT_IRQ_MASK;
+}
+
+void OPPROTO op_reset_inhibit_irq(void)
+{
+    env->hflags &= ~HF_INHIBIT_IRQ_MASK;
+}
+
 #if 0
 /* vm86plus instructions */
 void OPPROTO op_cli_vm(void)
@@ -573,9 +516,9 @@ void OPPROTO op_cmpxchg8b(void)
    helper_cmpxchg8b();
 }

-void OPPROTO op_jmp_tb_next(void)
+void OPPROTO op_jmp(void)
 {
-    JUMP_TB(PARAM1, 0, PARAM2);
+    JUMP_TB(op_jmp, PARAM1, 0, PARAM2);
 }

 void OPPROTO op_movl_T0_0(void)
@@ -583,6 +526,11 @@ void OPPROTO op_movl_T0_0(void)
    T0 = 0;
 }

+void OPPROTO op_exit_tb(void)
+{
+    EXIT_TB();
+}
+
 /* multiple size ops */

 #define ldul ldl
@@ -641,6 +589,38 @@ void OPPROTO op_movswl_DX_AX(void)
    EDX = (EDX & 0xffff0000) | (((int16_t)EAX >> 15) & 0xffff);
 }

+/* string ops helpers */
+
+void OPPROTO op_addl_ESI_T0(void)
+{
+    ESI += T0;
+}
+
+void OPPROTO op_addw_ESI_T0(void)
+{
+    ESI = (ESI & ~0xffff) | ((ESI + T0) & 0xffff);
+}
+
+void OPPROTO op_addl_EDI_T0(void)
+{
+    EDI += T0;
+}
+
+void OPPROTO op_addw_EDI_T0(void)
+{
+    EDI = (EDI & ~0xffff) | ((EDI + T0) & 0xffff);
+}
+
+void OPPROTO op_decl_ECX(void)
+{
+    ECX--;
+}
+
+void OPPROTO op_decw_ECX(void)
+{
+    ECX = (ECX & ~0xffff) | ((ECX - 1) & 0xffff);
+}
+
 /* push/pop */

 void op_pushl_T0(void)
@@ -665,7 +645,7 @@ void op_pushl_ss32_T0(void)
 {
    uint32_t offset;
    offset = ESP - 4;
-    stl(env->seg_cache[R_SS].base + offset, T0);
+    stl(env->segs[R_SS].base + offset, T0);
    /* modify ESP after to handle exceptions correctly */
    ESP = offset;
 }
@@ -674,7 +654,7 @@ void op_pushw_ss32_T0(void)
 {
    uint32_t offset;
    offset = ESP - 2;
-    stw(env->seg_cache[R_SS].base + offset, T0);
+    stw(env->segs[R_SS].base + offset, T0);
    /* modify ESP after to handle exceptions correctly */
    ESP = offset;
 }
@@ -683,7 +663,7 @@ void op_pushl_ss16_T0(void)
 {
    uint32_t offset;
    offset = (ESP - 4) & 0xffff;
-    stl(env->seg_cache[R_SS].base + offset, T0);
+    stl(env->segs[R_SS].base + offset, T0);
    /* modify ESP after to handle exceptions correctly */
    ESP = (ESP & ~0xffff) | offset;
 }
@@ -692,7 +672,7 @@ void op_pushw_ss16_T0(void)
 {
    uint32_t offset;
    offset = (ESP - 2) & 0xffff;
-    stw(env->seg_cache[R_SS].base + offset, T0);
+    stw(env->segs[R_SS].base + offset, T0);
    /* modify ESP after to handle exceptions correctly */
    ESP = (ESP & ~0xffff) | offset;
 }
@@ -710,22 +690,22 @@ void op_popw_T0(void)

 void op_popl_ss32_T0(void)
 {
-    T0 = ldl(env->seg_cache[R_SS].base + ESP);
+    T0 = ldl(env->segs[R_SS].base + ESP);
 }

 void op_popw_ss32_T0(void)
 {
-    T0 = lduw(env->seg_cache[R_SS].base + ESP);
+    T0 = lduw(env->segs[R_SS].base + ESP);
 }

 void op_popl_ss16_T0(void)
 {
-    T0 = ldl(env->seg_cache[R_SS].base + (ESP & 0xffff));
+    T0 = ldl(env->segs[R_SS].base + (ESP & 0xffff));
 }

 void op_popw_ss16_T0(void)
 {
-    T0 = lduw(env->seg_cache[R_SS].base + (ESP & 0xffff));
+    T0 = lduw(env->segs[R_SS].base + (ESP & 0xffff));
 }

 void op_addl_ESP_4(void)
@@ -768,6 +748,16 @@ void OPPROTO op_cpuid(void)
    helper_cpuid();
 }

+void OPPROTO op_rdmsr(void)
+{
+    helper_rdmsr();
+}
+
+void OPPROTO op_wrmsr(void)
+{
+    helper_wrmsr();
+}
+
 /* bcd */

 /* XXX: exception */
@@ -900,6 +890,7 @@ void OPPROTO op_das(void)

 /* segment handling */

+/* never use it with R_CS */
 void OPPROTO op_movl_seg_T0(void)
 {
    load_seg(PARAM1, T0 & 0xffff, PARAM2);
@@ -909,17 +900,18 @@ void OPPROTO op_movl_seg_T0(void)
 void OPPROTO op_movl_seg_T0_vm(void)
 {
    int selector;
+    SegmentCache *sc;
    
    selector = T0 & 0xffff;
    /* env->segs[] access */
-    *(uint32_t *)((char *)env + PARAM1) = selector;
-    /* env->seg_cache[] access */
-    ((SegmentCache *)((char *)env + PARAM2))->base = (void *)(selector << 4);
+    sc = (SegmentCache *)((char *)env + PARAM1);
+    sc->selector = selector;
+    sc->base = (void *)(selector << 4);
 }

 void OPPROTO op_movl_T0_seg(void)
 {
-    T0 = env->segs[PARAM1];
+    T0 = env->segs[PARAM1].selector;
 }

 void OPPROTO op_movl_A0_seg(void)
@@ -942,6 +934,91 @@ void OPPROTO op_lar(void)
    helper_lar();
 }

+/* T0: segment, T1:eip */
+void OPPROTO op_ljmp_protected_T0_T1(void)
+{
+    helper_ljmp_protected_T0_T1();
+}
+
+void OPPROTO op_lcall_real_T0_T1(void)
+{
+    helper_lcall_real_T0_T1(PARAM1, PARAM2);
+}
+
+void OPPROTO op_lcall_protected_T0_T1(void)
+{
+    helper_lcall_protected_T0_T1(PARAM1, PARAM2);
+}
+
+void OPPROTO op_iret_real(void)
+{
+    helper_iret_real(PARAM1);
+}
+
+void OPPROTO op_iret_protected(void)
+{
+    helper_iret_protected(PARAM1);
+}
+
+void OPPROTO op_lret_protected(void)
+{
+    helper_lret_protected(PARAM1, PARAM2);
+}
+
+void OPPROTO op_lldt_T0(void)
+{
+    helper_lldt_T0();
+}
+
+void OPPROTO op_ltr_T0(void)
+{
+    helper_ltr_T0();
+}
+
+/* CR registers access */
+void OPPROTO op_movl_crN_T0(void)
+{
+    helper_movl_crN_T0(PARAM1);
+}
+
+/* DR registers access */
+void OPPROTO op_movl_drN_T0(void)
+{
+    helper_movl_drN_T0(PARAM1);
+}
+
+void OPPROTO op_lmsw_T0(void)
+{
+    /* only 4 lower bits of CR0 are modified */
+    T0 = (env->cr[0] & ~0xf) | (T0 & 0xf);
+    helper_movl_crN_T0(0);
+}
+
+void OPPROTO op_invlpg_A0(void)
+{
+    helper_invlpg(A0);
+}
+
+void OPPROTO op_movl_T0_env(void)
+{
+    T0 = *(uint32_t *)((char *)env + PARAM1);
+}
+
+void OPPROTO op_movl_env_T0(void)
+{
+    *(uint32_t *)((char *)env + PARAM1) = T0;
+}
+
+void OPPROTO op_movl_env_T1(void)
+{
+    *(uint32_t *)((char *)env + PARAM1) = T1;
+}
+
+void OPPROTO op_clts(void)
+{
+    env->cr[0] &= ~CR0_TS_MASK;
+}
+
 /* flags handling */

 /* slow jumps cases : in order to avoid calling a function with a
@@ -950,9 +1027,18 @@ void OPPROTO op_lar(void)
 void OPPROTO op_jcc(void)
 {
    if (T0)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(op_jcc, PARAM1, 0, PARAM2);
    else
-        JUMP_TB(PARAM1, 1, PARAM3);
+        JUMP_TB(op_jcc, PARAM1, 1, PARAM3);
+    FORCE_RET();
+}
+
+void OPPROTO op_jcc_im(void)
+{
+    if (T0)
+        EIP = PARAM1;
+    else
+        EIP = PARAM2;
    FORCE_RET();
 }

@@ -1021,8 +1107,7 @@ void OPPROTO op_set_cc_op(void)
    CC_OP = PARAM1;
 }

-#define FL_UPDATE_MASK32 (TF_MASK | AC_MASK | ID_MASK)
-#define FL_UPDATE_MASK16 (TF_MASK)
+#define FL_UPDATE_MASK16 (FL_UPDATE_MASK32 & 0xffff)

 void OPPROTO op_movl_eflags_T0(void)
 {
@@ -1031,7 +1116,8 @@ void OPPROTO op_movl_eflags_T0(void)
    CC_SRC = eflags & (CC_O | CC_S | CC_Z | CC_A | CC_P | CC_C);
    DF = 1 - (2 * ((eflags >> 10) & 1));
    /* we also update some system flags as in user mode */
-    env->eflags = (env->eflags & ~FL_UPDATE_MASK32) | (eflags & FL_UPDATE_MASK32);
+    env->eflags = (env->eflags & ~FL_UPDATE_MASK32) | 
+        (eflags & FL_UPDATE_MASK32);
 }

 void OPPROTO op_movw_eflags_T0(void)
@@ -1041,7 +1127,18 @@ void OPPROTO op_movw_eflags_T0(void)
    CC_SRC = eflags & (CC_O | CC_S | CC_Z | CC_A | CC_P | CC_C);
    DF = 1 - (2 * ((eflags >> 10) & 1));
    /* we also update some system flags as in user mode */
-    env->eflags = (env->eflags & ~FL_UPDATE_MASK16) | (eflags & FL_UPDATE_MASK16);
+    env->eflags = (env->eflags & ~FL_UPDATE_MASK16) | 
+        (eflags & FL_UPDATE_MASK16);
+}
+
+void OPPROTO op_movl_eflags_T0_cpl0(void)
+{
+    load_eflags(T0, FL_UPDATE_CPL0_MASK);
+}
+
+void OPPROTO op_movw_eflags_T0_cpl0(void)
+{
+    load_eflags(T0, FL_UPDATE_CPL0_MASK & 0xffff);
 }

 #if 0
@@ -1346,28 +1443,40 @@ void OPPROTO op_fildll_FT0_A0(void)

 void OPPROTO op_flds_ST0_A0(void)
 {
+    int new_fpstt;
+    new_fpstt = (env->fpstt - 1) & 7;
 #ifdef USE_FP_CONVERT
    FP_CONVERT.i32 = ldl((void *)A0);
-    ST0 = FP_CONVERT.f;
+    env->fpregs[new_fpstt] = FP_CONVERT.f;
 #else
-    ST0 = ldfl((void *)A0);
+    env->fpregs[new_fpstt] = ldfl((void *)A0);
 #endif
+    env->fpstt = new_fpstt;
+    env->fptags[new_fpstt] = 0; /* validate stack entry */
 }

 void OPPROTO op_fldl_ST0_A0(void)
 {
+    int new_fpstt;
+    new_fpstt = (env->fpstt - 1) & 7;
 #ifdef USE_FP_CONVERT
    FP_CONVERT.i64 = ldq((void *)A0);
-    ST0 = FP_CONVERT.d;
+    env->fpregs[new_fpstt] = FP_CONVERT.d;
 #else
-    ST0 = ldfq((void *)A0);
+    env->fpregs[new_fpstt] = ldfq((void *)A0);
 #endif
+    env->fpstt = new_fpstt;
+    env->fptags[new_fpstt] = 0; /* validate stack entry */
 }

 #ifdef USE_X86LDOUBLE
 void OPPROTO op_fldt_ST0_A0(void)
 {
-    ST0 = *(long double *)A0;
+    int new_fpstt;
+    new_fpstt = (env->fpstt - 1) & 7;
+    env->fpregs[new_fpstt] = *(long double *)A0;
+    env->fpstt = new_fpstt;
+    env->fptags[new_fpstt] = 0; /* validate stack entry */
 }
 #else
 void OPPROTO op_fldt_ST0_A0(void)
@@ -1381,17 +1490,29 @@ void OPPROTO op_fldt_ST0_A0(void)

 void helper_fild_ST0_A0(void)
 {
-    ST0 = (CPU86_LDouble)ldsw((void *)A0);
+    int new_fpstt;
+    new_fpstt = (env->fpstt - 1) & 7;
+    env->fpregs[new_fpstt] = (CPU86_LDouble)ldsw((void *)A0);
+    env->fpstt = new_fpstt;
+    env->fptags[new_fpstt] = 0; /* validate stack entry */
 }

 void helper_fildl_ST0_A0(void)
 {
-    ST0 = (CPU86_LDouble)((int32_t)ldl((void *)A0));
+    int new_fpstt;
+    new_fpstt = (env->fpstt - 1) & 7;
+    env->fpregs[new_fpstt] = (CPU86_LDouble)((int32_t)ldl((void *)A0));
+    env->fpstt = new_fpstt;
+    env->fptags[new_fpstt] = 0; /* validate stack entry */
 }

 void helper_fildll_ST0_A0(void)
 {
-    ST0 = (CPU86_LDouble)((int64_t)ldq((void *)A0));
+    int new_fpstt;
+    new_fpstt = (env->fpstt - 1) & 7;
+    env->fpregs[new_fpstt] = (CPU86_LDouble)((int64_t)ldq((void *)A0));
+    env->fpstt = new_fpstt;
+    env->fptags[new_fpstt] = 0; /* validate stack entry */
 }

 void OPPROTO op_fild_ST0_A0(void)
@@ -1413,32 +1534,44 @@ void OPPROTO op_fildll_ST0_A0(void)

 void OPPROTO op_fild_ST0_A0(void)
 {
+    int new_fpstt;
+    new_fpstt = (env->fpstt - 1) & 7;
 #ifdef USE_FP_CONVERT
    FP_CONVERT.i32 = ldsw((void *)A0);
-    ST0 = (CPU86_LDouble)FP_CONVERT.i32;
+    env->fpregs[new_fpstt] = (CPU86_LDouble)FP_CONVERT.i32;
 #else
-    ST0 = (CPU86_LDouble)ldsw((void *)A0);
+    env->fpregs[new_fpstt] = (CPU86_LDouble)ldsw((void *)A0);
 #endif
+    env->fpstt = new_fpstt;
+    env->fptags[new_fpstt] = 0; /* validate stack entry */
 }

 void OPPROTO op_fildl_ST0_A0(void)
 {
+    int new_fpstt;
+    new_fpstt = (env->fpstt - 1) & 7;
 #ifdef USE_FP_CONVERT
    FP_CONVERT.i32 = (int32_t) ldl((void *)A0);
-    ST0 = (CPU86_LDouble)FP_CONVERT.i32;
+    env->fpregs[new_fpstt] = (CPU86_LDouble)FP_CONVERT.i32;
 #else
-    ST0 = (CPU86_LDouble)((int32_t)ldl((void *)A0));
+    env->fpregs[new_fpstt] = (CPU86_LDouble)((int32_t)ldl((void *)A0));
 #endif
+    env->fpstt = new_fpstt;
+    env->fptags[new_fpstt] = 0; /* validate stack entry */
 }

 void OPPROTO op_fildll_ST0_A0(void)
 {
+    int new_fpstt;
+    new_fpstt = (env->fpstt - 1) & 7;
 #ifdef USE_FP_CONVERT
    FP_CONVERT.i64 = (int64_t) ldq((void *)A0);
-    ST0 = (CPU86_LDouble)FP_CONVERT.i64;
+    env->fpregs[new_fpstt] = (CPU86_LDouble)FP_CONVERT.i64;
 #else
-    ST0 = (CPU86_LDouble)((int64_t)ldq((void *)A0));
+    env->fpregs[new_fpstt] = (CPU86_LDouble)((int64_t)ldq((void *)A0));
 #endif
+    env->fpstt = new_fpstt;
+    env->fptags[new_fpstt] = 0; /* validate stack entry */
 }

 #endif
--- a/op_string.h
+++ b/op_string.h
@@ -1,245 +0,0 @@
-
-void OPPROTO glue(glue(op_movs, SUFFIX), STRING_SUFFIX)(void)
-{
-    int v, inc;
-    inc = (DF << SHIFT);
-    v = glue(ldu, SUFFIX)(SI_ADDR);
-    glue(st, SUFFIX)(DI_ADDR, v);
-    inc = (DF << SHIFT);
-    INC_SI();
-    INC_DI();
-}
-
-void OPPROTO glue(glue(op_rep_movs, SUFFIX), STRING_SUFFIX)(void)
-{
-    int v, inc;
-    inc = (DF << SHIFT);
-    while (CX != 0) {
-        v = glue(ldu, SUFFIX)(SI_ADDR);
-        glue(st, SUFFIX)(DI_ADDR, v);
-        INC_SI();
-        INC_DI();
-        DEC_CX();
-    }
-    FORCE_RET();
-}
-
-void OPPROTO glue(glue(op_stos, SUFFIX), STRING_SUFFIX)(void)
-{
-    int inc;
-    glue(st, SUFFIX)(DI_ADDR, EAX);
-    inc = (DF << SHIFT);
-    INC_DI();
-}
-
-void OPPROTO glue(glue(op_rep_stos, SUFFIX), STRING_SUFFIX)(void)
-{
-    int inc;
-    inc = (DF << SHIFT);
-    while (CX != 0) {
-        glue(st, SUFFIX)(DI_ADDR, EAX);
-        INC_DI();
-        DEC_CX();
-    }
-    FORCE_RET();
-}
-
-void OPPROTO glue(glue(op_lods, SUFFIX), STRING_SUFFIX)(void)
-{
-    int v, inc;
-    v = glue(ldu, SUFFIX)(SI_ADDR);
-#if SHIFT == 0
-    EAX = (EAX & ~0xff) | v;
-#elif SHIFT == 1
-    EAX = (EAX & ~0xffff) | v;
-#else
-    EAX = v;
-#endif
-    inc = (DF << SHIFT);
-    INC_SI();
-}
-
-/* don't know if it is used */
-void OPPROTO glue(glue(op_rep_lods, SUFFIX), STRING_SUFFIX)(void)
-{
-    int v, inc;
-    inc = (DF << SHIFT);
-    while (CX != 0) {
-        v = glue(ldu, SUFFIX)(SI_ADDR);
-#if SHIFT == 0
-        EAX = (EAX & ~0xff) | v;
-#elif SHIFT == 1
-        EAX = (EAX & ~0xffff) | v;
-#else
-        EAX = v;
-#endif
-        INC_SI();
-        DEC_CX();
-    }
-    FORCE_RET();
-}
-
-void OPPROTO glue(glue(op_scas, SUFFIX), STRING_SUFFIX)(void)
-{
-    int v, inc;
-
-    v = glue(ldu, SUFFIX)(DI_ADDR);
-    inc = (DF << SHIFT);
-    INC_DI();
-    CC_SRC = EAX;
-    CC_DST = EAX - v;
-}
-
-void OPPROTO glue(glue(op_repz_scas, SUFFIX), STRING_SUFFIX)(void)
-{
-    int v1, v2, inc;
-
-    if (CX != 0) {
-        /* NOTE: the flags are not modified if CX == 0 */
-        v1 = EAX & DATA_MASK;
-        inc = (DF << SHIFT);
-        do {
-            v2 = glue(ldu, SUFFIX)(DI_ADDR);
-            INC_DI();
-            DEC_CX();
-            if (v1 != v2)
-                break;
-        } while (CX != 0);
-        CC_SRC = v1;
-        CC_DST = v1 - v2;
-        CC_OP = CC_OP_SUBB + SHIFT;
-    }
-    FORCE_RET();
-}
-
-void OPPROTO glue(glue(op_repnz_scas, SUFFIX), STRING_SUFFIX)(void)
-{
-    int v1, v2, inc;
-
-    if (CX != 0) {
-        /* NOTE: the flags are not modified if CX == 0 */
-        v1 = EAX & DATA_MASK;
-        inc = (DF << SHIFT);
-        do {
-            v2 = glue(ldu, SUFFIX)(DI_ADDR);
-            INC_DI();
-            DEC_CX();
-            if (v1 == v2)
-                break;
-        } while (CX != 0);
-        CC_SRC = v1;
-        CC_DST = v1 - v2;
-        CC_OP = CC_OP_SUBB + SHIFT;
-    }
-    FORCE_RET();
-}
-
-void OPPROTO glue(glue(op_cmps, SUFFIX), STRING_SUFFIX)(void)
-{
-    int v1, v2, inc;
-    v1 = glue(ldu, SUFFIX)(SI_ADDR);
-    v2 = glue(ldu, SUFFIX)(DI_ADDR);
-    inc = (DF << SHIFT);
-    INC_SI();
-    INC_DI();
-    CC_SRC = v1;
-    CC_DST = v1 - v2;
-}
-
-void OPPROTO glue(glue(op_repz_cmps, SUFFIX), STRING_SUFFIX)(void)
-{
-    int v1, v2, inc;
-    if (CX != 0) {
-        inc = (DF << SHIFT);
-        do {
-            v1 = glue(ldu, SUFFIX)(SI_ADDR);
-            v2 = glue(ldu, SUFFIX)(DI_ADDR);
-            INC_SI();
-            INC_DI();
-            DEC_CX();
-            if (v1 != v2)
-                break;
-        } while (CX != 0);
-        CC_SRC = v1;
-        CC_DST = v1 - v2;
-        CC_OP = CC_OP_SUBB + SHIFT;
-    }
-    FORCE_RET();
-}
-
-void OPPROTO glue(glue(op_repnz_cmps, SUFFIX), STRING_SUFFIX)(void)
-{
-    int v1, v2, inc;
-    if (CX != 0) {
-        inc = (DF << SHIFT);
-        do {
-            v1 = glue(ldu, SUFFIX)(SI_ADDR);
-            v2 = glue(ldu, SUFFIX)(DI_ADDR);
-            INC_SI();
-            INC_DI();
-            DEC_CX();
-            if (v1 == v2)
-                break;
-        } while (CX != 0);
-        CC_SRC = v1;
-        CC_DST = v1 - v2;
-        CC_OP = CC_OP_SUBB + SHIFT;
-    }
-    FORCE_RET();
-}
-
-void OPPROTO glue(glue(op_outs, SUFFIX), STRING_SUFFIX)(void)
-{
-    int v, dx, inc;
-    dx = EDX & 0xffff;
-    v = glue(ldu, SUFFIX)(SI_ADDR);
-    glue(cpu_x86_out, SUFFIX)(env, dx, v);
-    inc = (DF << SHIFT);
-    INC_SI();
-}
-
-void OPPROTO glue(glue(op_rep_outs, SUFFIX), STRING_SUFFIX)(void)
-{
-    int v, dx, inc;
-    inc = (DF << SHIFT);
-    dx = EDX & 0xffff;
-    while (CX != 0) {
-        v = glue(ldu, SUFFIX)(SI_ADDR);
-        glue(cpu_x86_out, SUFFIX)(env, dx, v);
-        INC_SI();
-        DEC_CX();
-    }
-    FORCE_RET();
-}
-
-void OPPROTO glue(glue(op_ins, SUFFIX), STRING_SUFFIX)(void)
-{
-    int v, dx, inc;
-    dx = EDX & 0xffff;
-    v = glue(cpu_x86_in, SUFFIX)(env, dx);
-    glue(st, SUFFIX)(DI_ADDR, v);
-    inc = (DF << SHIFT);
-    INC_DI();
-}
-
-void OPPROTO glue(glue(op_rep_ins, SUFFIX), STRING_SUFFIX)(void)
-{
-    int v, dx, inc;
-    inc = (DF << SHIFT);
-    dx = EDX & 0xffff;
-    while (CX != 0) {
-        v = glue(cpu_x86_in, SUFFIX)(env, dx);
-        glue(st, SUFFIX)(DI_ADDR, v);
-        INC_DI();
-        DEC_CX();
-    }
-    FORCE_RET();
-}
-
-#undef STRING_SUFFIX
-#undef SI_ADDR
-#undef DI_ADDR
-#undef INC_SI
-#undef INC_DI
-#undef CX
-#undef DEC_CX
--- a/ops_mem.h
+++ b/ops_mem.h
@@ -0,0 +1,66 @@
+void OPPROTO glue(glue(op_ldub, MEMSUFFIX), _T0_A0)(void)
+{
+    T0 = glue(ldub, MEMSUFFIX)((uint8_t *)A0);
+}
+
+void OPPROTO glue(glue(op_ldsb, MEMSUFFIX), _T0_A0)(void)
+{
+    T0 = glue(ldsb, MEMSUFFIX)((int8_t *)A0);
+}
+
+void OPPROTO glue(glue(op_lduw, MEMSUFFIX), _T0_A0)(void)
+{
+    T0 = glue(lduw, MEMSUFFIX)((uint8_t *)A0);
+}
+
+void OPPROTO glue(glue(op_ldsw, MEMSUFFIX), _T0_A0)(void)
+{
+    T0 = glue(ldsw, MEMSUFFIX)((int8_t *)A0);
+}
+
+void OPPROTO glue(glue(op_ldl, MEMSUFFIX), _T0_A0)(void)
+{
+    T0 = glue(ldl, MEMSUFFIX)((uint8_t *)A0);
+}
+
+void OPPROTO glue(glue(op_ldub, MEMSUFFIX), _T1_A0)(void)
+{
+    T1 = glue(ldub, MEMSUFFIX)((uint8_t *)A0);
+}
+
+void OPPROTO glue(glue(op_ldsb, MEMSUFFIX), _T1_A0)(void)
+{
+    T1 = glue(ldsb, MEMSUFFIX)((int8_t *)A0);
+}
+
+void OPPROTO glue(glue(op_lduw, MEMSUFFIX), _T1_A0)(void)
+{
+    T1 = glue(lduw, MEMSUFFIX)((uint8_t *)A0);
+}
+
+void OPPROTO glue(glue(op_ldsw, MEMSUFFIX), _T1_A0)(void)
+{
+    T1 = glue(ldsw, MEMSUFFIX)((int8_t *)A0);
+}
+
+void OPPROTO glue(glue(op_ldl, MEMSUFFIX), _T1_A0)(void)
+{
+    T1 = glue(ldl, MEMSUFFIX)((uint8_t *)A0);
+}
+
+void OPPROTO glue(glue(op_stb, MEMSUFFIX), _T0_A0)(void)
+{
+    glue(stb, MEMSUFFIX)((uint8_t *)A0, T0);
+}
+
+void OPPROTO glue(glue(op_stw, MEMSUFFIX), _T0_A0)(void)
+{
+    glue(stw, MEMSUFFIX)((uint8_t *)A0, T0);
+}
+
+void OPPROTO glue(glue(op_stl, MEMSUFFIX), _T0_A0)(void)
+{
+    glue(stl, MEMSUFFIX)((uint8_t *)A0, T0);
+}
+
+#undef MEMSUFFIX
--- a/ops_template.h
+++ b/ops_template.h
@@ -93,8 +93,8 @@ static int glue(compute_all_sub, SUFFIX)(void)
 {
    int cf, pf, af, zf, sf, of;
    int src1, src2;
-    src1 = CC_SRC;
-    src2 = CC_SRC - CC_DST;
+    src1 = CC_DST + CC_SRC;
+    src2 = CC_SRC;
    cf = (DATA_TYPE)src1 < (DATA_TYPE)src2;
    pf = parity_table[(uint8_t)CC_DST];
    af = (CC_DST ^ src1 ^ src2) & 0x10;
@@ -107,8 +107,8 @@ static int glue(compute_all_sub, SUFFIX)(void)
 static int glue(compute_c_sub, SUFFIX)(void)
 {
    int src1, src2, cf;
-    src1 = CC_SRC;
-    src2 = CC_SRC - CC_DST;
+    src1 = CC_DST + CC_SRC;
+    src2 = CC_SRC;
    cf = (DATA_TYPE)src1 < (DATA_TYPE)src2;
    return cf;
 }
@@ -117,8 +117,8 @@ static int glue(compute_all_sbb, SUFFIX)(void)
 {
    int cf, pf, af, zf, sf, of;
    int src1, src2;
-    src1 = CC_SRC;
-    src2 = CC_SRC - CC_DST - 1;
+    src1 = CC_DST + CC_SRC + 1;
+    src2 = CC_SRC;
    cf = (DATA_TYPE)src1 <= (DATA_TYPE)src2;
    pf = parity_table[(uint8_t)CC_DST];
    af = (CC_DST ^ src1 ^ src2) & 0x10;
@@ -131,8 +131,8 @@ static int glue(compute_all_sbb, SUFFIX)(void)
 static int glue(compute_c_sbb, SUFFIX)(void)
 {
    int src1, src2, cf;
-    src1 = CC_SRC;
-    src2 = CC_SRC - CC_DST - 1;
+    src1 = CC_DST + CC_SRC + 1;
+    src2 = CC_SRC;
    cf = (DATA_TYPE)src1 <= (DATA_TYPE)src2;
    return cf;
 }
@@ -234,70 +234,70 @@ static int glue(compute_all_sar, SUFFIX)(void)
 void OPPROTO glue(op_jb_sub, SUFFIX)(void)
 {
    int src1, src2;
-    src1 = CC_SRC;
-    src2 = CC_SRC - CC_DST;
+    src1 = CC_DST + CC_SRC;
+    src2 = CC_SRC;

    if ((DATA_TYPE)src1 < (DATA_TYPE)src2)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(glue(op_jb_sub, SUFFIX), PARAM1, 0, PARAM2);
    else
-        JUMP_TB(PARAM1, 1, PARAM3);
+        JUMP_TB(glue(op_jb_sub, SUFFIX), PARAM1, 1, PARAM3);
    FORCE_RET();
 }

 void OPPROTO glue(op_jz_sub, SUFFIX)(void)
 {
    if ((DATA_TYPE)CC_DST == 0)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(glue(op_jz_sub, SUFFIX), PARAM1, 0, PARAM2);
    else
-        JUMP_TB(PARAM1, 1, PARAM3);
+        JUMP_TB(glue(op_jz_sub, SUFFIX), PARAM1, 1, PARAM3);
    FORCE_RET();
 }

 void OPPROTO glue(op_jbe_sub, SUFFIX)(void)
 {
    int src1, src2;
-    src1 = CC_SRC;
-    src2 = CC_SRC - CC_DST;
+    src1 = CC_DST + CC_SRC;
+    src2 = CC_SRC;

    if ((DATA_TYPE)src1 <= (DATA_TYPE)src2)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(glue(op_jbe_sub, SUFFIX), PARAM1, 0, PARAM2);
    else
-        JUMP_TB(PARAM1, 1, PARAM3);
+        JUMP_TB(glue(op_jbe_sub, SUFFIX), PARAM1, 1, PARAM3);
    FORCE_RET();
 }

 void OPPROTO glue(op_js_sub, SUFFIX)(void)
 {
    if (CC_DST & SIGN_MASK)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(glue(op_js_sub, SUFFIX), PARAM1, 0, PARAM2);
    else
-        JUMP_TB(PARAM1, 1, PARAM3);
+        JUMP_TB(glue(op_js_sub, SUFFIX), PARAM1, 1, PARAM3);
    FORCE_RET();
 }

 void OPPROTO glue(op_jl_sub, SUFFIX)(void)
 {
    int src1, src2;
-    src1 = CC_SRC;
-    src2 = CC_SRC - CC_DST;
+    src1 = CC_DST + CC_SRC;
+    src2 = CC_SRC;

    if ((DATA_STYPE)src1 < (DATA_STYPE)src2)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(glue(op_jl_sub, SUFFIX), PARAM1, 0, PARAM2);
    else
-        JUMP_TB(PARAM1, 1, PARAM3);
+        JUMP_TB(glue(op_jl_sub, SUFFIX), PARAM1, 1, PARAM3);
    FORCE_RET();
 }

 void OPPROTO glue(op_jle_sub, SUFFIX)(void)
 {
    int src1, src2;
-    src1 = CC_SRC;
-    src2 = CC_SRC - CC_DST;
+    src1 = CC_DST + CC_SRC;
+    src2 = CC_SRC;

    if ((DATA_STYPE)src1 <= (DATA_STYPE)src2)
-        JUMP_TB(PARAM1, 0, PARAM2);
+        JUMP_TB(glue(op_jle_sub, SUFFIX), PARAM1, 0, PARAM2);
    else
-        JUMP_TB(PARAM1, 1, PARAM3);
+        JUMP_TB(glue(op_jle_sub, SUFFIX), PARAM1, 1, PARAM3);
    FORCE_RET();
 }

@@ -361,8 +361,8 @@ void OPPROTO glue(op_jecxz, SUFFIX)(void)
 void OPPROTO glue(op_setb_T0_sub, SUFFIX)(void)
 {
    int src1, src2;
-    src1 = CC_SRC;
-    src2 = CC_SRC - CC_DST;
+    src1 = CC_DST + CC_SRC;
+    src2 = CC_SRC;

    T0 = ((DATA_TYPE)src1 < (DATA_TYPE)src2);
 }
@@ -375,8 +375,8 @@ void OPPROTO glue(op_setz_T0_sub, SUFFIX)(void)
 void OPPROTO glue(op_setbe_T0_sub, SUFFIX)(void)
 {
    int src1, src2;
-    src1 = CC_SRC;
-    src2 = CC_SRC - CC_DST;
+    src1 = CC_DST + CC_SRC;
+    src2 = CC_SRC;

    T0 = ((DATA_TYPE)src1 <= (DATA_TYPE)src2);
 }
@@ -389,8 +389,8 @@ void OPPROTO glue(op_sets_T0_sub, SUFFIX)(void)
 void OPPROTO glue(op_setl_T0_sub, SUFFIX)(void)
 {
    int src1, src2;
-    src1 = CC_SRC;
-    src2 = CC_SRC - CC_DST;
+    src1 = CC_DST + CC_SRC;
+    src2 = CC_SRC;

    T0 = ((DATA_STYPE)src1 < (DATA_STYPE)src2);
 }
@@ -398,135 +398,14 @@ void OPPROTO glue(op_setl_T0_sub, SUFFIX)(void)
 void OPPROTO glue(op_setle_T0_sub, SUFFIX)(void)
 {
    int src1, src2;
-    src1 = CC_SRC;
-    src2 = CC_SRC - CC_DST;
+    src1 = CC_DST + CC_SRC;
+    src2 = CC_SRC;

    T0 = ((DATA_STYPE)src1 <= (DATA_STYPE)src2);
 }

 /* shifts */

-void OPPROTO glue(glue(op_rol, SUFFIX), _T0_T1_cc)(void)
-{
-    int count, src;
-    count = T1 & SHIFT_MASK;
-    if (count) {
-        CC_SRC = cc_table[CC_OP].compute_all() & ~(CC_O | CC_C);
-        src = T0;
-        T0 &= DATA_MASK;
-        T0 = (T0 << count) | (T0 >> (DATA_BITS - count));
-        CC_SRC |= (lshift(src ^ T0, 11 - (DATA_BITS - 1)) & CC_O) | 
-            (T0 & CC_C);
-        CC_OP = CC_OP_EFLAGS;
-    }
-    FORCE_RET();
-}
-
-void OPPROTO glue(glue(op_rol, SUFFIX), _T0_T1)(void)
-{
-    int count;
-    count = T1 & SHIFT_MASK;
-    if (count) {
-        T0 &= DATA_MASK;
-        T0 = (T0 << count) | (T0 >> (DATA_BITS - count));
-    }
-    FORCE_RET();
-}
-
-void OPPROTO glue(glue(op_ror, SUFFIX), _T0_T1_cc)(void)
-{
-    int count, src;
-    count = T1 & SHIFT_MASK;
-    if (count) {
-        CC_SRC = cc_table[CC_OP].compute_all() & ~(CC_O | CC_C);
-        src = T0;
-        T0 &= DATA_MASK;
-        T0 = (T0 >> count) | (T0 << (DATA_BITS - count));
-        CC_SRC |= (lshift(src ^ T0, 11 - (DATA_BITS - 1)) & CC_O) | 
-            ((T0 >> (DATA_BITS - 1)) & CC_C);
-        CC_OP = CC_OP_EFLAGS;
-    }
-    FORCE_RET();
-}
-
-void OPPROTO glue(glue(op_ror, SUFFIX), _T0_T1)(void)
-{
-    int count;
-    count = T1 & SHIFT_MASK;
-    if (count) {
-        T0 &= DATA_MASK;
-        T0 = (T0 >> count) | (T0 << (DATA_BITS - count));
-    }
-    FORCE_RET();
-}
-
-void OPPROTO glue(glue(op_rcl, SUFFIX), _T0_T1_cc)(void)
-{
-    int count, res, eflags;
-    unsigned int src;
-
-    count = T1 & 0x1f;
-#if DATA_BITS == 16
-    count = rclw_table[count];
-#elif DATA_BITS == 8
-    count = rclb_table[count];
-#endif
-    if (count) {
-        eflags = cc_table[CC_OP].compute_all();
-        T0 &= DATA_MASK;
-        src = T0;
-        res = (T0 << count) | ((eflags & CC_C) << (count - 1));
-        if (count > 1)
-            res |= T0 >> (DATA_BITS + 1 - count);
-        T0 = res;
-        CC_SRC = (eflags & ~(CC_C | CC_O)) |
-            (lshift(src ^ T0, 11 - (DATA_BITS - 1)) & CC_O) | 
-            ((src >> (DATA_BITS - count)) & CC_C);
-        CC_OP = CC_OP_EFLAGS;
-    }
-    FORCE_RET();
-}
-
-void OPPROTO glue(glue(op_rcr, SUFFIX), _T0_T1_cc)(void)
-{
-    int count, res, eflags;
-    unsigned int src;
-
-    count = T1 & 0x1f;
-#if DATA_BITS == 16
-    count = rclw_table[count];
-#elif DATA_BITS == 8
-    count = rclb_table[count];
-#endif
-    if (count) {
-        eflags = cc_table[CC_OP].compute_all();
-        T0 &= DATA_MASK;
-        src = T0;
-        res = (T0 >> count) | ((eflags & CC_C) << (DATA_BITS - count));
-        if (count > 1)
-            res |= T0 << (DATA_BITS + 1 - count);
-        T0 = res;
-        CC_SRC = (eflags & ~(CC_C | CC_O)) |
-            (lshift(src ^ T0, 11 - (DATA_BITS - 1)) & CC_O) | 
-            ((src >> (count - 1)) & CC_C);
-        CC_OP = CC_OP_EFLAGS;
-    }
-    FORCE_RET();
-}
-
-void OPPROTO glue(glue(op_shl, SUFFIX), _T0_T1_cc)(void)
-{
-    int count;
-    count = T1 & 0x1f;
-    if (count) {
-        CC_SRC = (DATA_TYPE)T0 << (count - 1);
-        T0 = T0 << count;
-        CC_DST = T0;
-        CC_OP = CC_OP_SHLB + SHIFT;
-    }
-    FORCE_RET();
-}
-
 void OPPROTO glue(glue(op_shl, SUFFIX), _T0_T1)(void)
 {
    int count;
@@ -535,20 +414,6 @@ void OPPROTO glue(glue(op_shl, SUFFIX), _T0_T1)(void)
    FORCE_RET();
 }

-void OPPROTO glue(glue(op_shr, SUFFIX), _T0_T1_cc)(void)
-{
-    int count;
-    count = T1 & 0x1f;
-    if (count) {
-        T0 &= DATA_MASK;
-        CC_SRC = T0 >> (count - 1);
-        T0 = T0 >> count;
-        CC_DST = T0;
-        CC_OP = CC_OP_SARB + SHIFT;
-    }
-    FORCE_RET();
-}
-
 void OPPROTO glue(glue(op_shr, SUFFIX), _T0_T1)(void)
 {
    int count;
@@ -558,20 +423,6 @@ void OPPROTO glue(glue(op_shr, SUFFIX), _T0_T1)(void)
    FORCE_RET();
 }

-void OPPROTO glue(glue(op_sar, SUFFIX), _T0_T1_cc)(void)
-{
-    int count, src;
-    count = T1 & 0x1f;
-    if (count) {
-        src = (DATA_STYPE)T0;
-        CC_SRC = src >> (count - 1);
-        T0 = src >> count;
-        CC_DST = T0;
-        CC_OP = CC_OP_SARB + SHIFT;
-    }
-    FORCE_RET();
-}
-
 void OPPROTO glue(glue(op_sar, SUFFIX), _T0_T1)(void)
 {
    int count, src;
@@ -581,166 +432,11 @@ void OPPROTO glue(glue(op_sar, SUFFIX), _T0_T1)(void)
    FORCE_RET();
 }

-#if DATA_BITS == 16
-/* XXX: overflow flag might be incorrect in some cases in shldw */
-void OPPROTO glue(glue(op_shld, SUFFIX), _T0_T1_im_cc)(void)
-{
-    int count;
-    unsigned int res;
-    count = PARAM1;
-    T1 &= 0xffff;
-    res = T1 | (T0 << 16);
-    CC_SRC = res >> (32 - count);
-    res <<= count;
-    if (count > 16)
-        res |= T1 << (count - 16);
-    T0 = res >> 16;
-    CC_DST = T0;
-}
+#undef MEM_WRITE
+#include "ops_template_mem.h"

-void OPPROTO glue(glue(op_shld, SUFFIX), _T0_T1_ECX_cc)(void)
-{
-    int count;
-    unsigned int res;
-    count = ECX & 0x1f;
-    if (count) {
-        T1 &= 0xffff;
-        res = T1 | (T0 << 16);
-        CC_SRC = res >> (32 - count);
-        res <<= count;
-        if (count > 16)
-          res |= T1 << (count - 16);
-        T0 = res >> 16;
-        CC_DST = T0;
-        CC_OP = CC_OP_SARB + SHIFT;
-    }
-    FORCE_RET();
-}
-
-void OPPROTO glue(glue(op_shrd, SUFFIX), _T0_T1_im_cc)(void)
-{
-    int count;
-    unsigned int res;
-
-    count = PARAM1;
-    res = (T0 & 0xffff) | (T1 << 16);
-    CC_SRC = res >> (count - 1);
-    res >>= count;
-    if (count > 16)
-        res |= T1 << (32 - count);
-    T0 = res;
-    CC_DST = T0;
-}
-
-
-void OPPROTO glue(glue(op_shrd, SUFFIX), _T0_T1_ECX_cc)(void)
-{
-    int count;
-    unsigned int res;
-
-    count = ECX & 0x1f;
-    if (count) {
-        res = (T0 & 0xffff) | (T1 << 16);
-        CC_SRC = res >> (count - 1);
-        res >>= count;
-        if (count > 16)
-            res |= T1 << (32 - count);
-        T0 = res;
-        CC_DST = T0;
-        CC_OP = CC_OP_SARB + SHIFT;
-    }
-    FORCE_RET();
-}
-#endif
-
-#if DATA_BITS == 32
-void OPPROTO glue(glue(op_shld, SUFFIX), _T0_T1_im_cc)(void)
-{
-    int count;
-    count = PARAM1;
-    T0 &= DATA_MASK;
-    T1 &= DATA_MASK;
-    CC_SRC = T0 << (count - 1);
-    T0 = (T0 << count) | (T1 >> (DATA_BITS - count));
-    CC_DST = T0;
-}
-
-void OPPROTO glue(glue(op_shld, SUFFIX), _T0_T1_ECX_cc)(void)
-{
-    int count;
-    count = ECX & 0x1f;
-    if (count) {
-        T0 &= DATA_MASK;
-        T1 &= DATA_MASK;
-        CC_SRC = T0 << (count - 1);
-        T0 = (T0 << count) | (T1 >> (DATA_BITS - count));
-        CC_DST = T0;
-        CC_OP = CC_OP_SHLB + SHIFT;
-    }
-    FORCE_RET();
-}
-
-void OPPROTO glue(glue(op_shrd, SUFFIX), _T0_T1_im_cc)(void)
-{
-    int count;
-    count = PARAM1;
-    T0 &= DATA_MASK;
-    T1 &= DATA_MASK;
-    CC_SRC = T0 >> (count - 1);
-    T0 = (T0 >> count) | (T1 << (DATA_BITS - count));
-    CC_DST = T0;
-}
-
-
-void OPPROTO glue(glue(op_shrd, SUFFIX), _T0_T1_ECX_cc)(void)
-{
-    int count;
-    count = ECX & 0x1f;
-    if (count) {
-        T0 &= DATA_MASK;
-        T1 &= DATA_MASK;
-        CC_SRC = T0 >> (count - 1);
-        T0 = (T0 >> count) | (T1 << (DATA_BITS - count));
-        CC_DST = T0;
-        CC_OP = CC_OP_SARB + SHIFT;
-    }
-    FORCE_RET();
-}
-#endif
-
-/* carry add/sub (we only need to set CC_OP differently) */
-
-void OPPROTO glue(glue(op_adc, SUFFIX), _T0_T1_cc)(void)
-{
-    int cf;
-    cf = cc_table[CC_OP].compute_c();
-    CC_SRC = T0;
-    T0 = T0 + T1 + cf;
-    CC_DST = T0;
-    CC_OP = CC_OP_ADDB + SHIFT + cf * 3;
-}
-
-void OPPROTO glue(glue(op_sbb, SUFFIX), _T0_T1_cc)(void)
-{
-    int cf;
-    cf = cc_table[CC_OP].compute_c();
-    CC_SRC = T0;
-    T0 = T0 - T1 - cf;
-    CC_DST = T0;
-    CC_OP = CC_OP_SUBB + SHIFT + cf * 3;
-}
-
-void OPPROTO glue(glue(op_cmpxchg, SUFFIX), _T0_T1_EAX_cc)(void)
-{
-    CC_SRC = EAX;
-    CC_DST = EAX - T0;
-    if ((DATA_TYPE)CC_DST == 0) {
-        T0 = T1;
-    } else {
-        EAX = (EAX & ~DATA_MASK) | (T0 & DATA_MASK);
-    }
-    FORCE_RET();
-}
+#define MEM_WRITE
+#include "ops_template_mem.h"

 /* bit operations */
 #if DATA_BITS >= 16
@@ -756,7 +452,7 @@ void OPPROTO glue(glue(op_bts, SUFFIX), _T0_T1_cc)(void)
 {
    int count;
    count = T1 & SHIFT_MASK;
-    CC_SRC = T0 >> count;
+    T1 = T0 >> count;
    T0 |= (1 << count);
 }

@@ -764,7 +460,7 @@ void OPPROTO glue(glue(op_btr, SUFFIX), _T0_T1_cc)(void)
 {
    int count;
    count = T1 & SHIFT_MASK;
-    CC_SRC = T0 >> count;
+    T1 = T0 >> count;
    T0 &= ~(1 << count);
 }

@@ -772,7 +468,7 @@ void OPPROTO glue(glue(op_btc, SUFFIX), _T0_T1_cc)(void)
 {
    int count;
    count = T1 & SHIFT_MASK;
-    CC_SRC = T0 >> count;
+    T1 = T0 >> count;
    T0 ^= (1 << count);
 }

@@ -814,35 +510,81 @@ void OPPROTO glue(glue(op_bsr, SUFFIX), _T0_cc)(void)

 #endif

+#if DATA_BITS == 32
+void OPPROTO op_update_bt_cc(void)
+{
+    CC_SRC = T1;
+}
+#endif
+
 /* string operations */
-/* XXX: maybe use lower level instructions to ease 16 bit / segment handling */

-#define STRING_SUFFIX _fast
-#define SI_ADDR (void *)ESI
-#define DI_ADDR (void *)EDI
-#define INC_SI() ESI += inc
-#define INC_DI() EDI += inc
-#define CX ECX
-#define DEC_CX() ECX--
-#include "op_string.h"
+void OPPROTO glue(op_movl_T0_Dshift, SUFFIX)(void)
+{
+    T0 = DF << SHIFT;
+}

-#define STRING_SUFFIX _a32
-#define SI_ADDR (uint8_t *)A0 + ESI
-#define DI_ADDR env->seg_cache[R_ES].base + EDI
-#define INC_SI() ESI += inc
-#define INC_DI() EDI += inc
-#define CX ECX
-#define DEC_CX() ECX--
-#include "op_string.h"
+void OPPROTO glue(op_string_jz_sub, SUFFIX)(void)
+{
+    if ((DATA_TYPE)CC_DST == 0)
+        JUMP_TB2(glue(op_string_jz_sub, SUFFIX), PARAM1, 1);
+    FORCE_RET();
+}

-#define STRING_SUFFIX _a16
-#define SI_ADDR (uint8_t *)A0 + (ESI & 0xffff)
-#define DI_ADDR env->seg_cache[R_ES].base + (EDI & 0xffff)
-#define INC_SI() ESI = (ESI & ~0xffff) | ((ESI + inc) & 0xffff)
-#define INC_DI() EDI = (EDI & ~0xffff) | ((EDI + inc) & 0xffff)
-#define CX (ECX & 0xffff)
-#define DEC_CX() ECX = (ECX & ~0xffff) | ((ECX - 1) & 0xffff)
-#include "op_string.h"
+void OPPROTO glue(op_string_jnz_sub, SUFFIX)(void)
+{
+    if ((DATA_TYPE)CC_DST != 0)
+        JUMP_TB2(glue(op_string_jnz_sub, SUFFIX), PARAM1, 1);
+    FORCE_RET();
+}
+
+void OPPROTO glue(glue(op_string_jz_sub, SUFFIX), _im)(void)
+{
+    if ((DATA_TYPE)CC_DST == 0) {
+        EIP = PARAM1;
+        if (env->eflags & TF_MASK) {
+            raise_exception(EXCP01_SSTP);
+        }
+        T0 = 0;
+        EXIT_TB();
+    }
+    FORCE_RET();
+}
+
+void OPPROTO glue(glue(op_string_jnz_sub, SUFFIX), _im)(void)
+{
+    if ((DATA_TYPE)CC_DST != 0) {
+        EIP = PARAM1;
+        if (env->eflags & TF_MASK) {
+            raise_exception(EXCP01_SSTP);
+        }
+        T0 = 0;
+        EXIT_TB();
+    }
+    FORCE_RET();
+}
+
+#if DATA_BITS >= 16
+void OPPROTO glue(op_jz_ecx, SUFFIX)(void)
+{
+    if ((DATA_TYPE)ECX == 0)
+        JUMP_TB(glue(op_jz_ecx, SUFFIX), PARAM1, 1, PARAM2);
+    FORCE_RET();
+}
+
+void OPPROTO glue(glue(op_jz_ecx, SUFFIX), _im)(void)
+{
+    if ((DATA_TYPE)ECX == 0) {
+        EIP = PARAM1;
+        if (env->eflags & TF_MASK) {
+            raise_exception(EXCP01_SSTP);
+        }
+        T0 = 0;
+        EXIT_TB();
+    }
+    FORCE_RET();
+}
+#endif

 /* port I/O */

@@ -856,6 +598,16 @@ void OPPROTO glue(glue(op_in, SUFFIX), _T0_T1)(void)
    T1 = glue(cpu_x86_in, SUFFIX)(env, T0 & 0xffff);
 }

+void OPPROTO glue(glue(op_in, SUFFIX), _DX_T0)(void)
+{
+    T0 = glue(cpu_x86_in, SUFFIX)(env, EDX & 0xffff);
+}
+
+void OPPROTO glue(glue(op_out, SUFFIX), _DX_T0)(void)
+{
+    glue(cpu_x86_out, SUFFIX)(env, EDX & 0xffff, T0);
+}
+
 #undef DATA_BITS
 #undef SHIFT_MASK
 #undef SIGN_MASK
--- a/ops_template_mem.h
+++ b/ops_template_mem.h
@@ -0,0 +1,429 @@
+/*
+ *  i386 micro operations (included several times to generate
+ *  different operand sizes)
+ * 
+ *  Copyright (c) 2003 Fabrice Bellard
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+#ifdef MEM_WRITE
+
+#if DATA_BITS == 8
+#define MEM_SUFFIX b_mem
+#elif DATA_BITS == 16
+#define MEM_SUFFIX w_mem
+#elif DATA_BITS == 32
+#define MEM_SUFFIX l_mem
+#endif
+
+#else
+
+#define MEM_SUFFIX SUFFIX
+
+#endif
+
+void OPPROTO glue(glue(op_rol, MEM_SUFFIX), _T0_T1_cc)(void)
+{
+    int count, src;
+    count = T1 & SHIFT_MASK;
+    if (count) {
+        src = T0;
+        T0 &= DATA_MASK;
+        T0 = (T0 << count) | (T0 >> (DATA_BITS - count));
+#ifdef MEM_WRITE
+        glue(st, SUFFIX)((uint8_t *)A0, T0);
+#else
+        /* gcc 3.2 workaround. This is really a bug in gcc. */
+        asm volatile("" : : "r" (T0));
+#endif
+        CC_SRC = (cc_table[CC_OP].compute_all() & ~(CC_O | CC_C)) | 
+            (lshift(src ^ T0, 11 - (DATA_BITS - 1)) & CC_O) | 
+            (T0 & CC_C);
+        CC_OP = CC_OP_EFLAGS;
+    }
+    FORCE_RET();
+}
+
+void OPPROTO glue(glue(op_ror, MEM_SUFFIX), _T0_T1_cc)(void)
+{
+    int count, src;
+    count = T1 & SHIFT_MASK;
+    if (count) {
+        src = T0;
+        T0 &= DATA_MASK;
+        T0 = (T0 >> count) | (T0 << (DATA_BITS - count));
+#ifdef MEM_WRITE
+        glue(st, SUFFIX)((uint8_t *)A0, T0);
+#else
+        /* gcc 3.2 workaround. This is really a bug in gcc. */
+        asm volatile("" : : "r" (T0));
+#endif
+        CC_SRC = (cc_table[CC_OP].compute_all() & ~(CC_O | CC_C)) |
+            (lshift(src ^ T0, 11 - (DATA_BITS - 1)) & CC_O) | 
+            ((T0 >> (DATA_BITS - 1)) & CC_C);
+        CC_OP = CC_OP_EFLAGS;
+    }
+    FORCE_RET();
+}
+
+void OPPROTO glue(glue(op_rol, MEM_SUFFIX), _T0_T1)(void)
+{
+    int count;
+    count = T1 & SHIFT_MASK;
+    if (count) {
+        T0 &= DATA_MASK;
+        T0 = (T0 << count) | (T0 >> (DATA_BITS - count));
+#ifdef MEM_WRITE
+        glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+    }
+    FORCE_RET();
+}
+
+void OPPROTO glue(glue(op_ror, MEM_SUFFIX), _T0_T1)(void)
+{
+    int count;
+    count = T1 & SHIFT_MASK;
+    if (count) {
+        T0 &= DATA_MASK;
+        T0 = (T0 >> count) | (T0 << (DATA_BITS - count));
+#ifdef MEM_WRITE
+        glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+    }
+    FORCE_RET();
+}
+
+void OPPROTO glue(glue(op_rcl, MEM_SUFFIX), _T0_T1_cc)(void)
+{
+    int count, res, eflags;
+    unsigned int src;
+
+    count = T1 & 0x1f;
+#if DATA_BITS == 16
+    count = rclw_table[count];
+#elif DATA_BITS == 8
+    count = rclb_table[count];
+#endif
+    if (count) {
+        eflags = cc_table[CC_OP].compute_all();
+        T0 &= DATA_MASK;
+        src = T0;
+        res = (T0 << count) | ((eflags & CC_C) << (count - 1));
+        if (count > 1)
+            res |= T0 >> (DATA_BITS + 1 - count);
+        T0 = res;
+#ifdef MEM_WRITE
+        glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+        CC_SRC = (eflags & ~(CC_C | CC_O)) |
+            (lshift(src ^ T0, 11 - (DATA_BITS - 1)) & CC_O) | 
+            ((src >> (DATA_BITS - count)) & CC_C);
+        CC_OP = CC_OP_EFLAGS;
+    }
+    FORCE_RET();
+}
+
+void OPPROTO glue(glue(op_rcr, MEM_SUFFIX), _T0_T1_cc)(void)
+{
+    int count, res, eflags;
+    unsigned int src;
+
+    count = T1 & 0x1f;
+#if DATA_BITS == 16
+    count = rclw_table[count];
+#elif DATA_BITS == 8
+    count = rclb_table[count];
+#endif
+    if (count) {
+        eflags = cc_table[CC_OP].compute_all();
+        T0 &= DATA_MASK;
+        src = T0;
+        res = (T0 >> count) | ((eflags & CC_C) << (DATA_BITS - count));
+        if (count > 1)
+            res |= T0 << (DATA_BITS + 1 - count);
+        T0 = res;
+#ifdef MEM_WRITE
+        glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+        CC_SRC = (eflags & ~(CC_C | CC_O)) |
+            (lshift(src ^ T0, 11 - (DATA_BITS - 1)) & CC_O) | 
+            ((src >> (count - 1)) & CC_C);
+        CC_OP = CC_OP_EFLAGS;
+    }
+    FORCE_RET();
+}
+
+void OPPROTO glue(glue(op_shl, MEM_SUFFIX), _T0_T1_cc)(void)
+{
+    int count, src;
+    count = T1 & 0x1f;
+    if (count) {
+        src = (DATA_TYPE)T0 << (count - 1);
+        T0 = T0 << count;
+#ifdef MEM_WRITE
+        glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+        CC_SRC = src;
+        CC_DST = T0;
+        CC_OP = CC_OP_SHLB + SHIFT;
+    }
+    FORCE_RET();
+}
+
+void OPPROTO glue(glue(op_shr, MEM_SUFFIX), _T0_T1_cc)(void)
+{
+    int count, src;
+    count = T1 & 0x1f;
+    if (count) {
+        T0 &= DATA_MASK;
+        src = T0 >> (count - 1);
+        T0 = T0 >> count;
+#ifdef MEM_WRITE
+        glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+        CC_SRC = src;
+        CC_DST = T0;
+        CC_OP = CC_OP_SARB + SHIFT;
+    }
+    FORCE_RET();
+}
+
+void OPPROTO glue(glue(op_sar, MEM_SUFFIX), _T0_T1_cc)(void)
+{
+    int count, src;
+    count = T1 & 0x1f;
+    if (count) {
+        src = (DATA_STYPE)T0;
+        T0 = src >> count;
+        src = src >> (count - 1);
+#ifdef MEM_WRITE
+        glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+        CC_SRC = src;
+        CC_DST = T0;
+        CC_OP = CC_OP_SARB + SHIFT;
+    }
+    FORCE_RET();
+}
+
+#if DATA_BITS == 16
+/* XXX: overflow flag might be incorrect in some cases in shldw */
+void OPPROTO glue(glue(op_shld, MEM_SUFFIX), _T0_T1_im_cc)(void)
+{
+    int count;
+    unsigned int res, tmp;
+    count = PARAM1;
+    T1 &= 0xffff;
+    res = T1 | (T0 << 16);
+    tmp = res >> (32 - count);
+    res <<= count;
+    if (count > 16)
+        res |= T1 << (count - 16);
+    T0 = res >> 16;
+#ifdef MEM_WRITE
+    glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+    CC_SRC = tmp;
+    CC_DST = T0;
+}
+
+void OPPROTO glue(glue(op_shld, MEM_SUFFIX), _T0_T1_ECX_cc)(void)
+{
+    int count;
+    unsigned int res, tmp;
+    count = ECX & 0x1f;
+    if (count) {
+        T1 &= 0xffff;
+        res = T1 | (T0 << 16);
+        tmp = res >> (32 - count);
+        res <<= count;
+        if (count > 16)
+          res |= T1 << (count - 16);
+        T0 = res >> 16;
+#ifdef MEM_WRITE
+        glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+        CC_SRC = tmp;
+        CC_DST = T0;
+        CC_OP = CC_OP_SARB + SHIFT;
+    }
+    FORCE_RET();
+}
+
+void OPPROTO glue(glue(op_shrd, MEM_SUFFIX), _T0_T1_im_cc)(void)
+{
+    int count;
+    unsigned int res, tmp;
+
+    count = PARAM1;
+    res = (T0 & 0xffff) | (T1 << 16);
+    tmp = res >> (count - 1);
+    res >>= count;
+    if (count > 16)
+        res |= T1 << (32 - count);
+    T0 = res;
+#ifdef MEM_WRITE
+    glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+    CC_SRC = tmp;
+    CC_DST = T0;
+}
+
+
+void OPPROTO glue(glue(op_shrd, MEM_SUFFIX), _T0_T1_ECX_cc)(void)
+{
+    int count;
+    unsigned int res, tmp;
+
+    count = ECX & 0x1f;
+    if (count) {
+        res = (T0 & 0xffff) | (T1 << 16);
+        tmp = res >> (count - 1);
+        res >>= count;
+        if (count > 16)
+            res |= T1 << (32 - count);
+        T0 = res;
+#ifdef MEM_WRITE
+        glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+        CC_SRC = tmp;
+        CC_DST = T0;
+        CC_OP = CC_OP_SARB + SHIFT;
+    }
+    FORCE_RET();
+}
+#endif
+
+#if DATA_BITS == 32
+void OPPROTO glue(glue(op_shld, MEM_SUFFIX), _T0_T1_im_cc)(void)
+{
+    int count, tmp;
+    count = PARAM1;
+    T0 &= DATA_MASK;
+    T1 &= DATA_MASK;
+    tmp = T0 << (count - 1);
+    T0 = (T0 << count) | (T1 >> (DATA_BITS - count));
+#ifdef MEM_WRITE
+    glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+    CC_SRC = tmp;
+    CC_DST = T0;
+}
+
+void OPPROTO glue(glue(op_shld, MEM_SUFFIX), _T0_T1_ECX_cc)(void)
+{
+    int count, tmp;
+    count = ECX & 0x1f;
+    if (count) {
+        T0 &= DATA_MASK;
+        T1 &= DATA_MASK;
+        tmp = T0 << (count - 1);
+        T0 = (T0 << count) | (T1 >> (DATA_BITS - count));
+#ifdef MEM_WRITE
+        glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+        CC_SRC = tmp;
+        CC_DST = T0;
+        CC_OP = CC_OP_SHLB + SHIFT;
+    }
+    FORCE_RET();
+}
+
+void OPPROTO glue(glue(op_shrd, MEM_SUFFIX), _T0_T1_im_cc)(void)
+{
+    int count, tmp;
+    count = PARAM1;
+    T0 &= DATA_MASK;
+    T1 &= DATA_MASK;
+    tmp = T0 >> (count - 1);
+    T0 = (T0 >> count) | (T1 << (DATA_BITS - count));
+#ifdef MEM_WRITE
+    glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+    CC_SRC = tmp;
+    CC_DST = T0;
+}
+
+
+void OPPROTO glue(glue(op_shrd, MEM_SUFFIX), _T0_T1_ECX_cc)(void)
+{
+    int count, tmp;
+    count = ECX & 0x1f;
+    if (count) {
+        T0 &= DATA_MASK;
+        T1 &= DATA_MASK;
+        tmp = T0 >> (count - 1);
+        T0 = (T0 >> count) | (T1 << (DATA_BITS - count));
+#ifdef MEM_WRITE
+        glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+        CC_SRC = tmp;
+        CC_DST = T0;
+        CC_OP = CC_OP_SARB + SHIFT;
+    }
+    FORCE_RET();
+}
+#endif
+
+/* carry add/sub (we only need to set CC_OP differently) */
+
+void OPPROTO glue(glue(op_adc, MEM_SUFFIX), _T0_T1_cc)(void)
+{
+    int cf;
+    cf = cc_table[CC_OP].compute_c();
+    T0 = T0 + T1 + cf;
+#ifdef MEM_WRITE
+    glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+    CC_SRC = T1;
+    CC_DST = T0;
+    CC_OP = CC_OP_ADDB + SHIFT + cf * 3;
+}
+
+void OPPROTO glue(glue(op_sbb, MEM_SUFFIX), _T0_T1_cc)(void)
+{
+    int cf;
+    cf = cc_table[CC_OP].compute_c();
+    T0 = T0 - T1 - cf;
+#ifdef MEM_WRITE
+    glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+    CC_SRC = T1;
+    CC_DST = T0;
+    CC_OP = CC_OP_SUBB + SHIFT + cf * 3;
+}
+
+void OPPROTO glue(glue(op_cmpxchg, MEM_SUFFIX), _T0_T1_EAX_cc)(void)
+{
+    unsigned int src, dst;
+
+    src = T0;
+    dst = EAX - T0;
+    if ((DATA_TYPE)dst == 0) {
+        T0 = T1;
+    } else {
+        EAX = (EAX & ~DATA_MASK) | (T0 & DATA_MASK);
+    }
+#ifdef MEM_WRITE
+    glue(st, SUFFIX)((uint8_t *)A0, T0);
+#endif
+    CC_SRC = src;
+    CC_DST = dst;
+    FORCE_RET();
+}
+
+#undef MEM_SUFFIX
+#undef MEM_WRITE
--- a/ppc-dis.c
+++ b/ppc-dis.c
@@ -3074,7 +3074,8 @@ static int print_insn_powerpc(FILE *, unsigned long insn, unsigned memaddr, int

 int print_insn_ppc (bfd_vma pc, disassemble_info *info)
 {
-	return print_insn_powerpc (info->stream, *(unsigned *)(long)pc, pc,
+	return print_insn_powerpc (info->stream, 
+                                   (unsigned int)bfd_getb32((bfd_byte *)pc), pc,
 				   PPC_OPCODE_PPC | PPC_OPCODE_601);
 }

--- a/qemu-doc.texi
+++ b/qemu-doc.texi
@@ -1,51 +1,87 @@
 \input texinfo @c -*- texinfo -*-

-@settitle QEMU x86 Emulator Reference Documentation
+@settitle QEMU CPU Emulator Reference Documentation
@titlepage
@sp 7
-@center @titlefont{QEMU x86 Emulator Reference Documentation}
+@center @titlefont{QEMU CPU Emulator Reference Documentation}
@sp 3
@end titlepage

@chapter Introduction

-QEMU is an x86 processor emulator. Its purpose is to run x86 Linux
-processes on non-x86 Linux architectures such as PowerPC. By using
-dynamic translation it achieves a reasonnable speed while being easy to
-port on new host CPUs. Its main goal is to be able to launch the
-@code{Wine} Windows API emulator (@url{http://www.winehq.org}) or
-@code{DOSEMU} (@url{http://www.dosemu.org}) on non-x86 CPUs.
+@section Features

-QEMU features:
+QEMU is a FAST! processor emulator. By using dynamic translation it
+achieves a reasonnable speed while being easy to port on new host
+CPUs.
+
+QEMU has two operating modes:
+@itemize
+@item User mode emulation. In this mode, QEMU can launch Linux processes
+compiled for one CPU on another CPU. Linux system calls are converted
+because of endianness and 32/64 bit mismatches. The Wine Windows API
+emulator (@url{http://www.winehq.org}) and the DOSEMU DOS emulator
+(@url{www.dosemu.org}) are the main targets for QEMU.
+
+@item Full system emulation. In this mode, QEMU emulates a full
+system, including a processor and various peripherials. Currently, it
+is only used to launch an x86 Linux kernel on an x86 Linux system. It
+enables easier testing and debugging of system code. It can also be
+used to provide virtual hosting of several virtual PCs on a single
+server.
+
+@end itemize
+
+As QEMU requires no host kernel patches to run, it is very safe and
+easy to use.
+
+QEMU generic features:

@itemize 

-@item User space only x86 emulator.
-
-@item Currently ported on i386, PowerPC. Work in progress for S390, Alpha and Sparc.
+@item User space only or full system emulation.

@item Using dynamic translation to native code for reasonnable speed.

-@item The virtual x86 CPU supports 16 bit and 32 bit addressing with segmentation. 
-User space LDT and GDT are emulated. VM86 mode is also supported.
+@item Working on x86 and PowerPC hosts. Being tested on ARM, Sparc32, Alpha and S390.

+@item Self-modifying code support.
+
+@item Precise exceptions support.
+
+@item The virtual CPU is a library (@code{libqemu}) which can be used 
+in other projects.
+
+@end itemize
+
+QEMU user mode emulation features:
+@itemize 
@item Generic Linux system call converter, including most ioctls.

@item clone() emulation using native CPU clone() to use Linux scheduler for threads.

-@item Accurate signal handling by remapping host signals to virtual x86 signals. 
+@item Accurate signal handling by remapping host signals to target signals. 
+@end itemize
+@end itemize

-@item Precise user space x86 exceptions.
+QEMU full system emulation features:
+@itemize 
+@item Using mmap() system calls to simulate the MMU
+@end itemize

-@item Self-modifying code support.
+@section x86 emulation

-@item Support of host page sizes bigger than 4KB.
+QEMU x86 target features:
+
+@itemize 
+
+@item The virtual x86 CPU supports 16 bit and 32 bit addressing with segmentation. 
+LDT/GDT and IDT are emulated. VM86 mode is also supported to run DOSEMU.
+
+@item Support of host page sizes bigger than 4KB in user mode emulation.

@item QEMU can emulate itself on x86.

-@item The virtual x86 CPU is a library (@code{libqemu}) which can be used 
-in other projects.
-
@item An extensive Linux x86 CPU test program is included @file{tests/test-i386}. 
 It can be used to test other x86 virtual CPUs.

@@ -62,18 +98,43 @@ Current QEMU limitations:
@item IPC syscalls are missing.

@item The x86 segment limits and access rights are not tested at every 
-memory access (and will never be to have good performances).
+memory access.

@item On non x86 host CPUs, @code{double}s are used instead of the non standard 
 10 byte @code{long double}s of x86 for floating point emulation to get
 maximum performances.

+@item Full system emulation only works if no data are mapped above the virtual address 
+0xc0000000 (yet).
+
+@item Some priviledged instructions or behaviors are missing. Only the ones 
+needed for proper Linux kernel operation are emulated.
+
+@item No memory separation between the kernel and the user processes is done. 
+It will be implemented very soon.
+
@end itemize

-@chapter Invocation
+@section ARM emulation
+
+@itemize
+
+@item ARM emulation can currently launch small programs while using the
+generic dynamic code generation architecture of QEMU.
+
+@item No FPU support (yet).
+
+@item No automatic regression testing (yet).
+
+@end itemize
+
+@chapter QEMU User space emulator invocation

@section Quick Start

+If you need to compile QEMU, please read the @file{README} which gives
+the related information.
+
 In order to launch a Linux process, QEMU needs the process executable
 itself and all the target (x86) dynamic libraries used by it. 

@@ -171,27 +232,421 @@ Activate log (logfile=/tmp/qemu.log)
 Act as if the host page size was 'pagesize' bytes
@end table

+@chapter QEMU System emulator invocation
+
+@section Quick Start
+
+This section explains how to launch a Linux kernel inside QEMU.
+
+@enumerate
+@item
+Download the archive @file{vl-test-xxx.tar.gz} containing a Linux
+kernel and a disk image. The archive also contains a precompiled
+version of @file{vl}, the QEMU System emulator.
+
+@item Optional: If you want network support (for example to launch X11 examples), you
+must copy the script @file{vl-ifup} in @file{/etc} and configure
+properly @code{sudo} so that the command @code{ifconfig} contained in
+@file{vl-ifup} can be executed as root. You must verify that your host
+kernel supports the TUN/TAP network interfaces: the device
+@file{/dev/net/tun} must be present.
+
+When network is enabled, there is a virtual network connection between
+the host kernel and the emulated kernel. The emulated kernel is seen
+from the host kernel at IP address 172.20.0.2 and the host kernel is
+seen from the emulated kernel at IP address 172.20.0.1.
+
+@item Launch @code{vl.sh}. You should have the following output:
+
+@example
+> ./vl.sh 
+connected to host network interface: tun0
+Uncompressing Linux... Ok, booting the kernel.
+Linux version 2.4.20 (fabrice@localhost.localdomain) (gcc version 2.96 20000731 (Red Hat Linux 7.3 2.96-110)) #22 lun jui 7 13:37:41 CEST 2003
+BIOS-provided physical RAM map:
+ BIOS-e801: 0000000000000000 - 000000000009f000 (usable)
+ BIOS-e801: 0000000000100000 - 0000000002000000 (usable)
+32MB LOWMEM available.
+On node 0 totalpages: 8192
+zone(0): 4096 pages.
+zone(1): 4096 pages.
+zone(2): 0 pages.
+Kernel command line: root=/dev/hda ide1=noprobe ide2=noprobe ide3=noprobe ide4=noprobe ide5=noprobe
+ide_setup: ide1=noprobe
+ide_setup: ide2=noprobe
+ide_setup: ide3=noprobe
+ide_setup: ide4=noprobe
+ide_setup: ide5=noprobe
+Initializing CPU#0
+Detected 501.285 MHz processor.
+Calibrating delay loop... 989.59 BogoMIPS
+Memory: 29268k/32768k available (907k kernel code, 3112k reserved, 212k data, 52k init, 0k highmem)
+Dentry cache hash table entries: 4096 (order: 3, 32768 bytes)
+Inode cache hash table entries: 2048 (order: 2, 16384 bytes)
+Mount-cache hash table entries: 512 (order: 0, 4096 bytes)
+Buffer-cache hash table entries: 1024 (order: 0, 4096 bytes)
+Page-cache hash table entries: 8192 (order: 3, 32768 bytes)
+CPU: Intel Pentium Pro stepping 03
+Checking 'hlt' instruction... OK.
+POSIX conformance testing by UNIFIX
+Linux NET4.0 for Linux 2.4
+Based upon Swansea University Computer Society NET3.039
+Initializing RT netlink socket
+apm: BIOS not found.
+Starting kswapd
+Journalled Block Device driver loaded
+pty: 256 Unix98 ptys configured
+Serial driver version 5.05c (2001-07-08) with no serial options enabled
+ttyS00 at 0x03f8 (irq = 4) is a 16450
+Uniform Multi-Platform E-IDE driver Revision: 6.31
+ide: Assuming 50MHz system bus speed for PIO modes; override with idebus=xx
+hda: QEMU HARDDISK, ATA DISK drive
+ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
+hda: 12288 sectors (6 MB) w/256KiB Cache, CHS=12/16/63
+Partition check:
+ hda: unknown partition table
+ne.c:v1.10 9/23/94 Donald Becker (becker@scyld.com)
+Last modified Nov 1, 2000 by Paul Gortmaker
+NE*000 ethercard probe at 0x300: 52 54 00 12 34 56
+eth0: NE2000 found at 0x300, using IRQ 9.
+RAMDISK driver initialized: 16 RAM disks of 4096K size 1024 blocksize
+NET4: Linux TCP/IP 1.0 for NET4.0
+IP Protocols: ICMP, UDP, TCP, IGMP
+IP: routing cache hash table of 512 buckets, 4Kbytes
+TCP: Hash tables configured (established 2048 bind 4096)
+NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
+EXT2-fs warning: mounting unchecked fs, running e2fsck is recommended
+VFS: Mounted root (ext2 filesystem).
+Freeing unused kernel memory: 52k freed
+sh: can't access tty; job control turned off
+#
+@end example
+
+@item
+Then you can play with the kernel inside the virtual serial console. You
+can launch @code{ls} for example. Type @key{Ctrl-a h} to have an help
+about the keys you can type inside the virtual serial console. In
+particular, use @key{Ctrl-a x} to exit QEMU and use @key{Ctrl-a b} as
+the Magic SysRq key.
+
+@item 
+If the network is enabled, launch the script @file{/etc/linuxrc} in the
+emulator (don't forget the leading dot):
+@example
+. /etc/linuxrc
+@end example
+
+Then enable X11 connections on your PC from the emulated Linux: 
+@example
+xhost +172.20.0.2
+@end example
+
+You can now launch @file{xterm} or @file{xlogo} and verify that you have
+a real Virtual Linux system !
+
+@end enumerate
+
+NOTES:
+@enumerate
+@item 
+A 2.5.74 kernel is also included in the vl-test archive. Just
+replace the bzImage in vl.sh to try it.
+
+@item 
+vl creates a temporary file in @var{$VLTMPDIR} (@file{/tmp} is the
+default) containing all the simulated PC memory. If possible, try to use
+a temporary directory using the tmpfs filesystem to avoid too many
+unnecessary disk accesses.
+
+@item 
+In order to exit cleanly for vl, you can do a @emph{shutdown} inside
+vl. vl will automatically exit when the Linux shutdown is done.
+
+@item 
+You can boot slightly faster by disabling the probe of non present IDE
+interfaces. To do so, add the following options on the kernel command
+line:
+@example
+ide1=noprobe ide2=noprobe ide3=noprobe ide4=noprobe ide5=noprobe
+@end example
+
+@item 
+The example disk image is a modified version of the one made by Kevin
+Lawton for the plex86 Project (@url{www.plex86.org}).
+
+@end enumerate
+
+@section Invocation
+
+@example
+usage: vl [options] bzImage [kernel parameters...]
+@end example
+
+@file{bzImage} is a Linux kernel image.
+
+General options:
+@table @option
+@item -hda file
+@item -hdb file
+Use 'file' as hard disk 0 or 1 image (@xref{disk_images}). 
+
+@item -snapshot
+
+Write to temporary files instead of disk image files. In this case,
+the raw disk image you use is not written back. You can however force
+the write back by pressing @key{C-a s} (@xref{disk_images}). 
+
+@item -m megs
+Set virtual RAM size to @var{megs} megabytes.
+
+@item -n script      
+Set network init script [default=/etc/vl-ifup]. This script is
+launched to configure the host network interface (usually tun0)
+corresponding to the virtual NE2000 card.
+
+@item -initrd file
+Use 'file' as initial ram disk.
+@end table
+
+Debug options:
+@table @option
+@item -s
+Wait gdb connection to port 1234.
+@item -p port
+Change gdb connection port.
+@item -d             
+Output log in /tmp/vl.log
+@end table
+
+During emulation, use @key{C-a h} to get terminal commands:
+
+@table @key
+@item C-a h
+Print this help
+@item C-a x    
+Exit emulatior
+@item C-a s    
+Save disk data back to file (if -snapshot)
+@item C-a b
+Send break (magic sysrq)
+@item C-a C-a
+Send C-a
+@end table
+
+@node disk_images
+@section Disk Images
+
+@subsection Raw disk images
+
+The disk images can simply be raw images of the hard disk. You can
+create them with the command:
+@example
+dd if=/dev/zero of=myimage bs=1024 count=mysize
+@end example
+where @var{myimage} is the image filename and @var{mysize} is its size
+in kilobytes.
+
+@subsection Snapshot mode
+
+If you use the option @option{-snapshot}, all disk images are
+considered as read only. When sectors in written, they are written in
+a temporary file created in @file{/tmp}. You can however force the
+write back to the raw disk images by pressing @key{C-a s}.
+
+NOTE: The snapshot mode only works with raw disk images.
+
+@subsection Copy On Write disk images
+
+QEMU also supports user mode Linux
+(@url{http://user-mode-linux.sourceforge.net/}) Copy On Write (COW)
+disk images. The COW disk images are much smaller than normal images
+as they store only modified sectors. They also permit the use of the
+same disk image template for many users.
+
+To create a COW disk images, use the command:
+
+@example
+vlmkcow -f myrawimage.bin mycowimage.cow
+@end example
+
+@file{myrawimage.bin} is a raw image you want to use as original disk
+image. It will never be written to.
+
+@file{mycowimage.cow} is the COW disk image which is created by
+@code{vlmkcow}. You can use it directly with the @option{-hdx}
+options. You must not modify the original raw disk image if you use
+COW images, as COW images only store the modified sectors from the raw
+disk image. QEMU stores the original raw disk image name and its
+modified time in the COW disk image so that chances of mistakes are
+reduced.
+
+If the raw disk image is not read-only, by pressing @key{C-a s} you
+can flush the COW disk image back into the raw disk image, as in
+snapshot mode.
+
+COW disk images can also be created without a corresponding raw disk
+image. It is useful to have a big initial virtual disk image without
+using much disk space. Use:
+
+@example
+vlmkcow mycowimage.cow 1024
+@end example
+
+to create a 1 gigabyte empty COW disk image.
+
+NOTES: 
+@enumerate
+@item
+COW disk images must be created on file systems supporting
+@emph{holes} such as ext2 or ext3.
+@item 
+Since holes are used, the displayed size of the COW disk image is not
+the real one. To know it, use the @code{ls -ls} command.
+@end enumerate
+
+@section Linux Kernel Compilation
+
+You should be able to use any kernel with QEMU provided you make the
+following changes (only 2.4.x and 2.5.x were tested):
+
+@enumerate
+@item
+The kernel must be mapped at 0x90000000 (the default is
+0xc0000000). You must modify only two lines in the kernel source:
+
+In @file{include/asm/page.h}, replace
+@example
+#define __PAGE_OFFSET           (0xc0000000)
+@end example
+by
+@example
+#define __PAGE_OFFSET           (0x90000000)
+@end example
+
+And in @file{arch/i386/vmlinux.lds}, replace
+@example
+  . = 0xc0000000 + 0x100000;
+@end example
+by 
+@example
+  . = 0x90000000 + 0x100000;
+@end example
+
+@item
+If you want to enable SMP (Symmetric Multi-Processing) support, you
+must make the following change in @file{include/asm/fixmap.h}. Replace
+@example
+#define FIXADDR_TOP	(0xffffX000UL)
+@end example
+by 
+@example
+#define FIXADDR_TOP	(0xa7ffX000UL)
+@end example
+(X is 'e' or 'f' depending on the kernel version). Although you can
+use an SMP kernel with QEMU, it only supports one CPU.
+
+@item
+If you are not using a 2.5 kernel as host kernel but if you use a target
+2.5 kernel, you must also ensure that the 'HZ' define is set to 100
+(1000 is the default) as QEMU cannot currently emulate timers at
+frequencies greater than 100 Hz on host Linux systems < 2.5. In
+@file{include/asm/param.h}, replace:
+
+@example
+# define HZ		1000		/* Internal kernel timer frequency */
+@end example
+by
+@example
+# define HZ		100		/* Internal kernel timer frequency */
+@end example
+
+@end enumerate
+
+The file config-2.x.x gives the configuration of the example kernels.
+
+Just type
+@example
+make bzImage
+@end example
+
+As you would do to make a real kernel. Then you can use with QEMU
+exactly the same kernel as you would boot on your PC (in
+@file{arch/i386/boot/bzImage}).
+
+@section PC Emulation
+
+QEMU emulates the following PC peripherials:
+
+@itemize
+@item
+PIC (interrupt controler)
+@item
+PIT (timers)
+@item 
+CMOS memory
+@item
+Dumb VGA (to print the @code{Uncompressing Linux} message)
+@item
+Serial port (port=0x3f8, irq=4)
+@item 
+NE2000 network adapter (port=0x300, irq=9)
+@item 
+IDE disk interface (port=0x1f0, irq=14)
+@end itemize
+
+@section GDB usage
+
+QEMU has a primitive support to work with gdb, so that you can do
+'Ctrl-C' while the kernel is running and inspect its state.
+
+In order to use gdb, launch vl with the '-s' option. It will wait for a
+gdb connection:
+@example
+> vl -s arch/i386/boot/bzImage -hda root-2.4.20.img root=/dev/hda
+Connected to host network interface: tun0
+Waiting gdb connection on port 1234
+@end example
+
+Then launch gdb on the 'vmlinux' executable:
+@example
+> gdb vmlinux
+@end example
+
+In gdb, connect to QEMU:
+@example
+(gdb) target remote locahost:1234
+@end example
+
+Then you can use gdb normally. For example, type 'c' to launch the kernel:
+@example
+(gdb) c
+@end example
+
+WARNING: breakpoints and single stepping are not yet supported.
+
@chapter QEMU Internals

@section QEMU compared to other emulators

-Unlike bochs [3], QEMU emulates only a user space x86 CPU. It means that
-you cannot launch an operating system with it. The benefit is that it is
-simpler and faster due to the fact that some of the low level CPU state
-can be ignored (in particular, no virtual memory needs to be emulated).
+Like bochs [3], QEMU emulates an x86 CPU. But QEMU is much faster than
+bochs as it uses dynamic compilation and because it uses the host MMU to
+simulate the x86 MMU. The downside is that currently the emulation is
+not as accurate as bochs (for example, you cannot currently run Windows
+inside QEMU).

 Like Valgrind [2], QEMU does user space emulation and dynamic
 translation. Valgrind is mainly a memory debugger while QEMU has no
-support for it (QEMU could be used to detect out of bound memory accesses
-as Valgrind, but it has no support to track uninitialised data as
-Valgrind does). Valgrind dynamic translator generates better code than
-QEMU (in particular it does register allocation) but it is closely tied
-to an x86 host.
+support for it (QEMU could be used to detect out of bound memory
+accesses as Valgrind, but it has no support to track uninitialised data
+as Valgrind does). The Valgrind dynamic translator generates better code
+than QEMU (in particular it does register allocation) but it is closely
+tied to an x86 host and target and has no support for precise exceptions
+and system emulation.

-EM86 [4] is the closest project to QEMU (and QEMU still uses some of its
-code, in particular the ELF file loader). EM86 was limited to an alpha
-host and used a proprietary and slow interpreter (the interpreter part
-of the FX!32 Digital Win32 code translator [5]).
+EM86 [4] is the closest project to user space QEMU (and QEMU still uses
+some of its code, in particular the ELF file loader). EM86 was limited
+to an alpha host and used a proprietary and slow interpreter (the
+interpreter part of the FX!32 Digital Win32 code translator [5]).

 TWIN [6] is a Windows API emulator like Wine. It is less accurate than
 Wine but includes a protected mode x86 interpreter to launch x86 Windows
@@ -200,11 +655,25 @@ Windows API is executed natively but it is far more difficult to develop
 because all the data structures and function parameters exchanged
 between the API and the x86 code must be converted.

+User mode Linux [7] was the only solution before QEMU to launch a Linux
+kernel as a process while not needing any host kernel patches. However,
+user mode Linux requires heavy kernel patches while QEMU accepts
+unpatched Linux kernels. It would be interesting to compare the
+performance of the two approaches.
+
+The new Plex86 [8] PC virtualizer is done in the same spirit as the QEMU
+system emulator. It requires a patched Linux kernel to work (you cannot
+launch the same kernel on your PC), but the patches are really small. As
+it is a PC virtualizer (no emulation is done except for some priveledged
+instructions), it has the potential of being faster than QEMU. The
+downside is that a complicated (and potentially unsafe) host kernel
+patch is needed.
+
@section Portable dynamic translation

 QEMU is a dynamic translator. When it first encounters a piece of code,
 it converts it to the host instruction set. Usually dynamic translators
-are very complicated and highly CPU dependant. QEMU uses some tricks
+are very complicated and highly CPU dependent. QEMU uses some tricks
 which make it relatively easily portable and simple while achieving good
 performances.

@@ -304,7 +773,7 @@ Correct translated code invalidation is done efficiently by maintaining
 a linked list of every translated block contained in a given page. Other
 linked lists are also maintained to undo direct block chaining. 

-Althought the overhead of doing @code{mprotect()} calls is important,
+Although the overhead of doing @code{mprotect()} calls is important,
 most MSDOS programs can be emulated at reasonnable speed with QEMU and
 DOSEMU.

@@ -373,7 +842,7 @@ reentrancy.

@section Self-virtualization

-QEMU was conceived so that ultimately it can emulate itself. Althought
+QEMU was conceived so that ultimately it can emulate itself. Although
 it is not very useful, it is an important test to show the power of the
 emulator.

@@ -382,6 +851,16 @@ space conflicts. QEMU solves this problem by being an executable ELF
 shared object as the ld-linux.so ELF interpreter. That way, it can be
 relocated at load time.

+@section MMU emulation
+
+For system emulation, QEMU uses the mmap() system call to emulate the
+target CPU MMU. It works as long the emulated OS does not use an area
+reserved by the host OS (such as the area above 0xc0000000 on x86
+Linux).
+
+It is planned to add a slower but more precise MMU emulation
+with a software MMU.
+
@section Bibliography

@table @asis
@@ -412,18 +891,31 @@ Chernoff and Ray Hookway.
@url{http://www.willows.com/}, Windows API library emulation from
 Willows Software.

+@item [7]
+@url{http://user-mode-linux.sourceforge.net/}, 
+The User-mode Linux Kernel.
+
+@item [8]
+@url{http://www.plex86.org/}, 
+The new Plex86 project.
+
@end table

@chapter Regression Tests

-In the directory @file{tests/}, various interesting x86 testing programs
+In the directory @file{tests/}, various interesting testing programs
 are available. There are used for regression testing.

-@section @file{hello}
+@section @file{hello-i386}

 Very simple statically linked x86 program, just to test QEMU during a
 port to a new host CPU.

+@section @file{hello-arm}
+
+Very simple statically linked ARM program, just to test QEMU during a
+port to a new host CPU.
+
@section @file{test-i386}

 This program executes most of the 16 bit and 32 bit x86 instructions and
--- a/sdl.c
+++ b/sdl.c
@@ -0,0 +1,275 @@
+/*
+ * QEMU SDL display driver
+ * 
+ * Copyright (c) 2003 Fabrice Bellard
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+#include <stdlib.h>
+#include <stdio.h>
+#include <stdarg.h>
+#include <string.h>
+#include <getopt.h>
+#include <inttypes.h>
+#include <unistd.h>
+#include <sys/mman.h>
+#include <fcntl.h>
+#include <signal.h>
+#include <time.h>
+#include <sys/time.h>
+#include <malloc.h>
+#include <termios.h>
+#include <sys/poll.h>
+#include <errno.h>
+#include <sys/wait.h>
+#include <netinet/in.h>
+
+#include <SDL.h>
+
+#include "cpu-i386.h"
+#include "exec.h"
+
+#include "vl.h"
+
+static SDL_Surface *screen;
+static int gui_grab; /* if true, all keyboard/mouse events are grabbed */
+
+static void sdl_update(DisplayState *ds, int x, int y, int w, int h)
+{
+    SDL_UpdateRect(screen, x, y, w, h);
+}
+
+static void sdl_resize(DisplayState *ds, int w, int h)
+{
+    int flags;
+
+    //    printf("resizing to %d %d\n", w, h);
+
+    flags = SDL_HWSURFACE|SDL_ASYNCBLIT|SDL_HWACCEL;
+    flags |= SDL_RESIZABLE;
+    screen = SDL_SetVideoMode(w, h, 0, flags);
+    if (!screen) {
+        fprintf(stderr, "Could not open SDL display\n");
+        exit(1);
+    }
+    ds->data = screen->pixels;
+    ds->linesize = screen->pitch;
+    ds->depth = screen->format->BitsPerPixel;
+}
+
+static const uint32_t x_keycode_to_pc_keycode[61] = {
+   0x47e0,      /*  97  Home   */
+   0x48e0,      /*  98  Up     */
+   0x49e0,      /*  99  PgUp   */
+   0x4be0,      /* 100  Left   */
+   0x4c,        /* 101  KP-5   */
+   0x4de0,      /* 102  Right  */
+   0x4fe0,      /* 103  End    */
+   0x50e0,      /* 104  Down   */
+   0x51e0,      /* 105  PgDn   */
+   0x52e0,      /* 106  Ins    */
+   0x53e0,      /* 107  Del    */
+   0x1ce0,      /* 108  Enter  */
+   0x1de0,      /* 109  Ctrl-R */
+   0x451de1,    /* 110  Pause  */
+   0x37e0,      /* 111  Print  */
+   0x35e0,      /* 112  Divide */
+   0x38e0,      /* 113  Alt-R  */
+   0x46e0,      /* 114  Break  */   
+   0x0,         /* 115 */
+   0x0,         /* 116 */
+   0x0,         /* 117 */
+   0x0,         /* 118 */
+   0x0,         /* 119 */
+   0x0,         /* 120 */
+   0x0,         /* 121 */
+   0x0,         /* 122 */
+   0x0,         /* 123 */
+   0x0,         /* 124 */
+   0x0,         /* 125 */
+   0x0,         /* 126 */
+   0x0,         /* 127 */
+   0x0,         /* 128 */
+   0x0,         /* 129 */
+   0x0,         /* 130 */
+   0x0,         /* 131 */
+   0x0,         /* 132 */
+   0x0,         /* 133 */
+   0x0,         /* 134 */
+   0x0,         /* 135 */
+   0x47,         /* 136 KP_7 */
+   0x48,         /* 137 KP_8 */
+   0x49,         /* 138 KP_9 */
+   0x4b,         /* 139 KP_4 */
+   0x4c,         /* 140 KP_5 */
+   0x4d,         /* 141 KP_6 */
+   0x4f,         /* 142 KP_1 */
+   0x50,         /* 143 KP_2 */
+   0x51,         /* 144 KP_3 */
+   0x52,         /* 145 KP_0 */
+   0x53,         /* 146 KP_. */
+   0x47,         /* 147 KP_HOME */
+   0x48,         /* 148 KP_UP */
+   0x49,         /* 149 KP_PgUp */
+   0x4b,         /* 150 KP_Left */
+   0x4c,         /* 151 KP_ */
+   0x4d,         /* 152 KP_Right */
+   0x4f,         /* 153 KP_End */
+   0x50,         /* 154 KP_Down */
+   0x51,         /* 155 KP_PgDn */
+   0x52,         /* 156 KP_Ins */
+   0x53,         /* 157 KP_Del */
+};
+
+static void sdl_process_key(SDL_KeyboardEvent *ev)
+{
+    int keycode, v;
+    
+    /* XXX: not portable, but avoids complicated mappings */
+    keycode = ev->keysym.scancode;
+    if (keycode < 9) {
+        keycode = 0;
+    } else if (keycode < 97) {
+        keycode -= 8; /* just an offset */
+    } else if (keycode < 158) {
+        /* use conversion table */
+        keycode = x_keycode_to_pc_keycode[keycode - 97];
+    } else {
+        keycode = 0;
+    }
+    
+    /* now send the key code */
+    while (keycode != 0) {
+        v = keycode & 0xff;
+        if (ev->type == SDL_KEYUP)
+            v |= 0x80;
+        kbd_put_keycode(v);
+        keycode >>= 8;
+    }
+}
+
+static void sdl_grab_start(void)
+{
+    SDL_WM_SetCaption("QEMU - Press Ctrl-Shift to exit grab", "QEMU");
+    SDL_ShowCursor(0);
+    SDL_WM_GrabInput(SDL_GRAB_ON);
+    /* dummy read to avoid moving the mouse */
+    SDL_GetRelativeMouseState(NULL, NULL);
+    gui_grab = 1;
+}
+
+static void sdl_grab_end(void)
+{
+    SDL_WM_SetCaption("QEMU", "QEMU");
+    SDL_WM_GrabInput(SDL_GRAB_OFF);
+    SDL_ShowCursor(1);
+    gui_grab = 0;
+}
+
+static void sdl_send_mouse_event(void)
+{
+    int dx, dy, dz, state, buttons;
+    state = SDL_GetRelativeMouseState(&dx, &dy);
+    buttons = 0;
+    if (state & SDL_BUTTON(SDL_BUTTON_LEFT))
+        buttons |= MOUSE_EVENT_LBUTTON;
+    if (state & SDL_BUTTON(SDL_BUTTON_RIGHT))
+        buttons |= MOUSE_EVENT_RBUTTON;
+    if (state & SDL_BUTTON(SDL_BUTTON_MIDDLE))
+        buttons |= MOUSE_EVENT_MBUTTON;
+    /* XXX: test wheel */
+    dz = 0;
+    if (state & SDL_BUTTON(SDL_BUTTON_WHEELUP))
+        dz--;
+    if (state & SDL_BUTTON(SDL_BUTTON_WHEELDOWN))
+        dz++;
+    kbd_mouse_event(dx, dy, dz, buttons);
+}
+
+static void sdl_refresh(DisplayState *ds)
+{
+    SDL_Event ev1, *ev = &ev1;
+
+    vga_update_display();
+    while (SDL_PollEvent(ev)) {
+        switch (ev->type) {
+        case SDL_VIDEOEXPOSE:
+            sdl_update(ds, 0, 0, screen->w, screen->h);
+            break;
+        case SDL_KEYDOWN:
+        case SDL_KEYUP:
+            if (ev->type == SDL_KEYDOWN) {
+                if ((SDL_GetModState() & (KMOD_LSHIFT | KMOD_LCTRL)) ==
+                    (KMOD_LSHIFT | KMOD_LCTRL)) {
+                    /* exit/enter grab if pressing Ctrl-Shift */
+                    if (!gui_grab)
+                        sdl_grab_start();
+                    else
+                        sdl_grab_end();
+                }
+            }
+            sdl_process_key(&ev->key);
+            break;
+        case SDL_QUIT:
+            reset_requested = 1;
+            break;
+        case SDL_MOUSEMOTION:
+            if (gui_grab) {
+                sdl_send_mouse_event();
+            }
+            break;
+        case SDL_MOUSEBUTTONDOWN:
+        case SDL_MOUSEBUTTONUP:
+            {
+                SDL_MouseButtonEvent *bev = &ev->button;
+                if (!gui_grab) {
+                    if (ev->type == SDL_MOUSEBUTTONDOWN &&
+                        (bev->state & SDL_BUTTON_LMASK)) {
+                        /* start grabbing all events */
+                        sdl_grab_start();
+                    }
+                } else {
+                    sdl_send_mouse_event();
+                }
+            }
+            break;
+        default:
+            break;
+        }
+    }
+}
+
+void sdl_display_init(DisplayState *ds)
+{
+    int flags;
+
+    flags = SDL_INIT_VIDEO | SDL_INIT_NOPARACHUTE;
+    if (SDL_Init (flags)) {
+        fprintf(stderr, "Could not initialize SDL - exiting\n");
+        exit(1);
+    }
+    ds->dpy_update = sdl_update;
+    ds->dpy_resize = sdl_resize;
+    ds->dpy_refresh = sdl_refresh;
+
+    sdl_resize(ds, 640, 400);
+    SDL_WM_SetCaption("QEMU", "QEMU");
+    SDL_EnableKeyRepeat(250, 50);
+    gui_grab = 0;
+}
--- a/softmmu_header.h
+++ b/softmmu_header.h
@@ -0,0 +1,111 @@
+/*
+ *  Software MMU support
+ * 
+ *  Copyright (c) 2003 Fabrice Bellard
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+#if DATA_SIZE == 8
+#define SUFFIX q
+#define DATA_TYPE uint64_t
+#elif DATA_SIZE == 4
+#define SUFFIX l
+#define DATA_TYPE uint32_t
+#elif DATA_SIZE == 2
+#define SUFFIX w
+#define DATA_TYPE uint16_t
+#define DATA_STYPE int16_t
+#elif DATA_SIZE == 1
+#define SUFFIX b
+#define DATA_TYPE uint8_t
+#define DATA_STYPE int8_t
+#else
+#error unsupported data size
+#endif
+
+#if MEMUSER == 0
+#define MEMSUFFIX _kernel
+#else
+#define MEMSUFFIX _user
+#endif
+
+#if DATA_SIZE == 8
+#define RES_TYPE uint64_t
+#else
+#define RES_TYPE int
+#endif
+
+
+#if MEMUSER == 0
+DATA_TYPE REGPARM(1) glue(glue(__ld, SUFFIX), _mmu)(unsigned long addr);
+void REGPARM(2) glue(glue(__st, SUFFIX), _mmu)(unsigned long addr, DATA_TYPE v);
+#endif
+
+static inline int glue(glue(ldu, SUFFIX), MEMSUFFIX)(void *ptr)
+{
+    int index;
+    RES_TYPE res;
+    unsigned long addr, physaddr;
+    addr = (unsigned long)ptr;
+    index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1);
+    if (__builtin_expect(env->tlb_read[MEMUSER][index].address != 
+                         (addr & (TARGET_PAGE_MASK | (DATA_SIZE - 1))), 0)) {
+        res = glue(glue(__ld, SUFFIX), _mmu)(addr);
+    } else {
+        physaddr = addr + env->tlb_read[MEMUSER][index].addend;
+        res = glue(glue(ldu, SUFFIX), _raw)((uint8_t *)physaddr);
+    }
+    return res;
+}
+
+#if DATA_SIZE <= 2
+static inline int glue(glue(lds, SUFFIX), MEMSUFFIX)(void *ptr)
+{
+    int res, index;
+    unsigned long addr, physaddr;
+    addr = (unsigned long)ptr;
+    index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1);
+    if (__builtin_expect(env->tlb_read[MEMUSER][index].address != 
+                         (addr & (TARGET_PAGE_MASK | (DATA_SIZE - 1))), 0)) {
+        res = (DATA_STYPE)glue(glue(__ld, SUFFIX), _mmu)(addr);
+    } else {
+        physaddr = addr + env->tlb_read[MEMUSER][index].addend;
+        res = glue(glue(lds, SUFFIX), _raw)((uint8_t *)physaddr);
+    }
+    return res;
+}
+#endif
+
+static inline void glue(glue(st, SUFFIX), MEMSUFFIX)(void *ptr, RES_TYPE v)
+{
+    int index;
+    unsigned long addr, physaddr;
+    addr = (unsigned long)ptr;
+    index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1);
+    if (__builtin_expect(env->tlb_write[MEMUSER][index].address != 
+                         (addr & (TARGET_PAGE_MASK | (DATA_SIZE - 1))), 0)) {
+        glue(glue(__st, SUFFIX), _mmu)(addr, v);
+    } else {
+        physaddr = addr + env->tlb_write[MEMUSER][index].addend;
+        glue(glue(st, SUFFIX), _raw)((uint8_t *)physaddr, v);
+    }
+}
+
+#undef RES_TYPE
+#undef DATA_TYPE
+#undef DATA_STYPE
+#undef SUFFIX
+#undef DATA_SIZE
+#undef MEMSUFFIX
--- a/softmmu_template.h
+++ b/softmmu_template.h
@@ -0,0 +1,241 @@
+/*
+ *  Software MMU support
+ * 
+ *  Copyright (c) 2003 Fabrice Bellard
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+#define DATA_SIZE (1 << SHIFT)
+
+#if DATA_SIZE == 8
+#define SUFFIX q
+#define DATA_TYPE uint64_t
+#elif DATA_SIZE == 4
+#define SUFFIX l
+#define DATA_TYPE uint32_t
+#elif DATA_SIZE == 2
+#define SUFFIX w
+#define DATA_TYPE uint16_t
+#elif DATA_SIZE == 1
+#define SUFFIX b
+#define DATA_TYPE uint8_t
+#else
+#error unsupported data size
+#endif
+
+static DATA_TYPE glue(slow_ld, SUFFIX)(unsigned long addr, void *retaddr);
+static void glue(slow_st, SUFFIX)(unsigned long addr, DATA_TYPE val,
+                                  void *retaddr);
+
+static inline DATA_TYPE glue(io_read, SUFFIX)(unsigned long physaddr, 
+                                              unsigned long tlb_addr)
+{
+    DATA_TYPE res;
+    int index;
+
+    index = (tlb_addr >> IO_MEM_SHIFT) & (IO_MEM_NB_ENTRIES - 1);
+#if SHIFT <= 2
+    res = io_mem_read[index][SHIFT](physaddr);
+#else
+#ifdef TARGET_WORDS_BIGENDIAN
+    res = (uint64_t)io_mem_read[index][2](physaddr) << 32;
+    res |= io_mem_read[index][2](physaddr + 4);
+#else
+    res = io_mem_read[index][2](physaddr);
+    res |= (uint64_t)io_mem_read[index][2](physaddr + 4) << 32;
+#endif
+#endif /* SHIFT > 2 */
+    return res;
+}
+
+static inline void glue(io_write, SUFFIX)(unsigned long physaddr, 
+                                          DATA_TYPE val,
+                                          unsigned long tlb_addr)
+{
+    int index;
+
+    index = (tlb_addr >> IO_MEM_SHIFT) & (IO_MEM_NB_ENTRIES - 1);
+#if SHIFT <= 2
+    io_mem_write[index][SHIFT](physaddr, val);
+#else
+#ifdef TARGET_WORDS_BIGENDIAN
+    io_mem_write[index][2](physaddr, val >> 32);
+    io_mem_write[index][2](physaddr + 4, val);
+#else
+    io_mem_write[index][2](physaddr, val);
+    io_mem_write[index][2](physaddr + 4, val >> 32);
+#endif
+#endif /* SHIFT > 2 */
+}
+
+/* handle all cases except unaligned access which span two pages */
+DATA_TYPE REGPARM(1) glue(glue(__ld, SUFFIX), _mmu)(unsigned long addr)
+{
+    DATA_TYPE res;
+    int is_user, index;
+    unsigned long physaddr, tlb_addr;
+    void *retaddr;
+    
+    /* test if there is match for unaligned or IO access */
+    /* XXX: could done more in memory macro in a non portable way */
+    is_user = ((env->hflags & HF_CPL_MASK) == 3);
+    index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1);
+ redo:
+    tlb_addr = env->tlb_read[is_user][index].address;
+    if ((addr & TARGET_PAGE_MASK) == (tlb_addr & (TARGET_PAGE_MASK | TLB_INVALID_MASK))) {
+        physaddr = addr + env->tlb_read[is_user][index].addend;
+        if (tlb_addr & ~TARGET_PAGE_MASK) {
+            /* IO access */
+            if ((addr & (DATA_SIZE - 1)) != 0)
+                goto do_unaligned_access;
+            res = glue(io_read, SUFFIX)(physaddr, tlb_addr);
+        } else if (((addr & 0xfff) + DATA_SIZE - 1) >= TARGET_PAGE_SIZE) {
+            /* slow unaligned access (it spans two pages or IO) */
+        do_unaligned_access:
+            retaddr = __builtin_return_address(0);
+            res = glue(slow_ld, SUFFIX)(addr, retaddr);
+        } else {
+            /* unaligned access in the same page */
+            res = glue(glue(ldu, SUFFIX), _raw)((uint8_t *)physaddr);
+        }
+    } else {
+        /* the page is not in the TLB : fill it */
+        retaddr = __builtin_return_address(0);
+        tlb_fill(addr, 0, retaddr);
+        goto redo;
+    }
+    return res;
+}
+
+/* handle all unaligned cases */
+static DATA_TYPE glue(slow_ld, SUFFIX)(unsigned long addr, void *retaddr)
+{
+    DATA_TYPE res, res1, res2;
+    int is_user, index, shift;
+    unsigned long physaddr, tlb_addr, addr1, addr2;
+
+    is_user = ((env->hflags & HF_CPL_MASK) == 3);
+    index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1);
+ redo:
+    tlb_addr = env->tlb_read[is_user][index].address;
+    if ((addr & TARGET_PAGE_MASK) == (tlb_addr & (TARGET_PAGE_MASK | TLB_INVALID_MASK))) {
+        physaddr = addr + env->tlb_read[is_user][index].addend;
+        if (tlb_addr & ~TARGET_PAGE_MASK) {
+            /* IO access */
+            if ((addr & (DATA_SIZE - 1)) != 0)
+                goto do_unaligned_access;
+            res = glue(io_read, SUFFIX)(physaddr, tlb_addr);
+        } else if (((addr & 0xfff) + DATA_SIZE - 1) >= TARGET_PAGE_SIZE) {
+        do_unaligned_access:
+            /* slow unaligned access (it spans two pages) */
+            addr1 = addr & ~(DATA_SIZE - 1);
+            addr2 = addr1 + DATA_SIZE;
+            res1 = glue(slow_ld, SUFFIX)(addr1, retaddr);
+            res2 = glue(slow_ld, SUFFIX)(addr2, retaddr);
+            shift = (addr & (DATA_SIZE - 1)) * 8;
+#ifdef TARGET_WORDS_BIGENDIAN
+            res = (res1 << shift) | (res2 >> ((DATA_SIZE * 8) - shift));
+#else
+            res = (res1 >> shift) | (res2 << ((DATA_SIZE * 8) - shift));
+#endif
+        } else {
+            /* unaligned/aligned access in the same page */
+            res = glue(glue(ldu, SUFFIX), _raw)((uint8_t *)physaddr);
+        }
+    } else {
+        /* the page is not in the TLB : fill it */
+        tlb_fill(addr, 0, retaddr);
+        goto redo;
+    }
+    return res;
+}
+
+
+void REGPARM(2) glue(glue(__st, SUFFIX), _mmu)(unsigned long addr, DATA_TYPE val)
+{
+    unsigned long physaddr, tlb_addr;
+    void *retaddr;
+    int is_user, index;
+    
+    is_user = ((env->hflags & HF_CPL_MASK) == 3);
+    index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1);
+ redo:
+    tlb_addr = env->tlb_write[is_user][index].address;
+    if ((addr & TARGET_PAGE_MASK) == (tlb_addr & (TARGET_PAGE_MASK | TLB_INVALID_MASK))) {
+        physaddr = addr + env->tlb_read[is_user][index].addend;
+        if (tlb_addr & ~TARGET_PAGE_MASK) {
+            /* IO access */
+            if ((addr & (DATA_SIZE - 1)) != 0)
+                goto do_unaligned_access;
+            glue(io_write, SUFFIX)(physaddr, val, tlb_addr);
+        } else if (((addr & 0xfff) + DATA_SIZE - 1) >= TARGET_PAGE_SIZE) {
+        do_unaligned_access:
+            retaddr = __builtin_return_address(0);
+            glue(slow_st, SUFFIX)(addr, val, retaddr);
+        } else {
+            /* aligned/unaligned access in the same page */
+            glue(glue(st, SUFFIX), _raw)((uint8_t *)physaddr, val);
+        }
+    } else {
+        /* the page is not in the TLB : fill it */
+        retaddr = __builtin_return_address(0);
+        tlb_fill(addr, 1, retaddr);
+        goto redo;
+    }
+}
+
+/* handles all unaligned cases */
+static void glue(slow_st, SUFFIX)(unsigned long addr, DATA_TYPE val,
+                                  void *retaddr)
+{
+    unsigned long physaddr, tlb_addr;
+    int is_user, index, i;
+
+    is_user = ((env->hflags & HF_CPL_MASK) == 3);
+    index = (addr >> TARGET_PAGE_BITS) & (CPU_TLB_SIZE - 1);
+ redo:
+    tlb_addr = env->tlb_write[is_user][index].address;
+    if ((addr & TARGET_PAGE_MASK) == (tlb_addr & (TARGET_PAGE_MASK | TLB_INVALID_MASK))) {
+        physaddr = addr + env->tlb_read[is_user][index].addend;
+        if (tlb_addr & ~TARGET_PAGE_MASK) {
+            /* IO access */
+            if ((addr & (DATA_SIZE - 1)) != 0)
+                goto do_unaligned_access;
+            glue(io_write, SUFFIX)(physaddr, val, tlb_addr);
+        } else if (((addr & 0xfff) + DATA_SIZE - 1) >= TARGET_PAGE_SIZE) {
+        do_unaligned_access:
+            /* XXX: not efficient, but simple */
+            for(i = 0;i < DATA_SIZE; i++) {
+#ifdef TARGET_WORDS_BIGENDIAN
+                slow_stb(addr + i, val >> (((DATA_SIZE - 1) * 8) - (i * 8)), retaddr);
+#else
+                slow_stb(addr + i, val >> (i * 8), retaddr);
+#endif
+            }
+        } else {
+            /* aligned/unaligned access in the same page */
+            glue(glue(st, SUFFIX), _raw)((uint8_t *)physaddr, val);
+        }
+    } else {
+        /* the page is not in the TLB : fill it */
+        tlb_fill(addr, 1, retaddr);
+        goto redo;
+    }
+}
+
+#undef SHIFT
+#undef DATA_TYPE
+#undef SUFFIX
+#undef DATA_SIZE
--- a/syscall-arm.h
+++ b/syscall-arm.h
@@ -25,3 +25,4 @@ struct target_pt_regs {
 #define ARM_r0		uregs[0]
 #define ARM_ORIG_r0	uregs[17]

+#define ARM_SYSCALL_BASE	0x900000
--- a/tests/.cvsignore
+++ b/tests/.cvsignore
@@ -1,6 +1,6 @@
 gmon.out
 testsig
- hello
+ hello-i386
 hello-arm
 sha1.test.c
 sha1.c
--- a/tests/Makefile
+++ b/tests/Makefile
@@ -1,4 +1,4 @@
-include ../config.mak
+include ../config-host.mak

 CFLAGS=-Wall -O2 -g
 LDFLAGS=
@@ -8,13 +8,13 @@ TESTS=testclone testsig testthread sha1-i386 test-i386 runcom
 endif
 TESTS+=sha1 test_path

-QEMU=../qemu
+QEMU=../i386/qemu

 all: $(TESTS)

-hello: hello.c
+hello-i386: hello-i386.c
 	$(CC) -nostdlib $(CFLAGS) -static $(LDFLAGS) -o $@ $<
-	strip hello
+	strip $@

 testclone: testclone.c
 	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $<
--- a/tests/hello-i386.c
+++ b/tests/hello-i386.c
--- a/tests/test-i386-code16.S
+++ b/tests/test-i386-code16.S
@@ -77,4 +77,21 @@ myfunc2:


 code16_end:
+
+
+/* other 32 bits tests */
+        .code32
+
+        .globl func_lret32
+func_lret32:
+        movl $0x87654321, %eax
+        lret
+
+        .globl func_iret32
+func_iret32:
+        movl $0xabcd4321, %eax
+        iret
+
+                
+
        
--- a/tests/test-i386.c
+++ b/tests/test-i386.c
@@ -1,16 +1,19 @@
 #define _GNU_SOURCE
 #include <stdlib.h>
 #include <stdio.h>
+#include <string.h>
 #include <inttypes.h>
 #include <math.h>
 #include <signal.h>
 #include <setjmp.h>
+#include <errno.h>
 #include <sys/ucontext.h>
 #include <sys/mman.h>
 #include <asm/vm86.h>

 #define TEST_CMOV  0
 #define TEST_FCOMI 0
+//#define LINUX_VM86_IOPL_FIX

 #define xglue(x, y) x ## y
 #define glue(x, y) xglue(x, y)
@@ -591,7 +594,7 @@ void test_fbcd(double a)
    asm("fld1\n"\
        prefix "fnstenv %1\n"\
        prefix "fldenv %1\n"\
-        : "=t" (res) : "m" (*(env)) : "st");\
+        : "=t" (res) : "m" (*(env)));\
    printf("res=%f\n", res);\
    printf("fpuc=%04x fpus=%04x fptag=%04x\n",\
           (env)->fpuc,\
@@ -601,7 +604,7 @@ void test_fbcd(double a)
    asm("fld1\n"\
        prefix "fnsave %1\n"\
        prefix "frstor %1\n"\
-        : "=t" (res) : "m" (*(env)) : "st");\
+        : "=t" (res) : "m" (*(env)));\
    printf("res=%f\n", res);\
    printf("fpuc=%04x fpus=%04x fptag=%04x\n",\
           (env)->fpuc,\
@@ -854,7 +857,6 @@ void test_segs(void)
 #endif
    /* do some tests with fs or gs */
    asm volatile ("movl %0, %%fs" : : "r" (MK_SEL(1)));
-    asm volatile ("movl %0, %%gs" : : "r" (MK_SEL(2)));

    seg_data1[1] = 0xaa;
    seg_data2[1] = 0x55;
@@ -862,7 +864,12 @@ void test_segs(void)
    asm volatile ("fs movzbl 0x1, %0" : "=r" (res));
    printf("FS[1] = %02x\n", res);

-    asm volatile ("gs movzbl 0x1, %0" : "=r" (res));
+    asm volatile ("pushl %%gs\n"
+                  "movl %1, %%gs\n"
+                  "gs movzbl 0x1, %0\n"
+                  "popl %%gs\n"
+                  : "=r" (res)
+                  : "r" (MK_SEL(2)));
    printf("GS[1] = %02x\n", res);

    /* tests with ds/ss (implicit segment case) */
@@ -937,6 +944,9 @@ void test_code16(void)
    printf("func3() = 0x%08x\n", res);
 }

+extern char func_lret32;
+extern char func_iret32;
+
 void test_misc(void)
 {
    char table[256];
@@ -946,6 +956,26 @@ void test_misc(void)
    res = 0x12345678;
    asm ("xlat" : "=a" (res) : "b" (table), "0" (res));
    printf("xlat: EAX=%08x\n", res);
+
+    asm volatile ("pushl %%cs ; call %1" 
+                  : "=a" (res)
+                  : "m" (func_lret32): "memory", "cc");
+    printf("func_lret32=%x\n", res);
+
+    asm volatile ("pushfl ; pushl %%cs ; call %1" 
+                  : "=a" (res)
+                  : "m" (func_iret32): "memory", "cc");
+    printf("func_iret32=%x\n", res);
+
+    /* specific popl test */
+    asm volatile ("pushl $12345432 ; pushl $0x9abcdef ; popl (%%esp) ; popl %0"
+                  : "=g" (res));
+    printf("popl esp=%x\n", res);
+
+    /* specific popw test */
+    asm volatile ("pushl $12345432 ; pushl $0x9abcdef ; popw (%%esp) ; addl $2, %%esp ; popl %0"
+                  : "=g" (res));
+    printf("popw esp=%x\n", res);
 }

 uint8_t str_buffer[4096];
@@ -1077,7 +1107,7 @@ void test_vm86(void)
        switch(VM86_TYPE(ret)) {
        case VM86_INTx:
            {
-                int int_num, ah;
+                int int_num, ah, v;
                
                int_num = VM86_ARG(ret);
                if (int_num != 0x21)
@@ -1105,8 +1135,12 @@ void test_vm86(void)
                        r->eax = (r->eax & ~0xff) | '$';
                    }
                    break;
-                case 0xff: /* extension: write hex number in edx */
-                    printf("%08x\n", (int)r->edx);
+                case 0xff: /* extension: write eflags number in edx */
+                    v = (int)r->edx;
+#ifndef LINUX_VM86_IOPL_FIX
+                    v &= ~0x3000;
+#endif
+                    printf("%08x\n", v);
                    break;
                default:
                unknown_int:
@@ -1231,6 +1265,8 @@ void test_exceptions(void)
    printf("PF exception:\n");
    if (setjmp(jmp_env) == 0) {
        val = 1;
+        /* we add a nop to test a weird PC retrieval case */
+        asm volatile ("nop");
        /* now store in an invalid address */
        *(char *)0x1234 = 1;
    }
@@ -1325,6 +1361,90 @@ void test_exceptions(void)
    printf("val=0x%x\n", val);
 }

+/* specific precise single step test */
+void sig_trap_handler(int sig, siginfo_t *info, void *puc)
+{
+    struct ucontext *uc = puc;
+    printf("EIP=0x%08x\n", uc->uc_mcontext.gregs[REG_EIP]);
+}
+
+const uint8_t sstep_buf1[4] = { 1, 2, 3, 4};
+uint8_t sstep_buf2[4];
+
+void test_single_step(void)
+{
+    struct sigaction act;
+    volatile int val;
+    int i;
+
+    val = 0;
+    act.sa_sigaction = sig_trap_handler;
+    sigemptyset(&act.sa_mask);
+    act.sa_flags = SA_SIGINFO;
+    sigaction(SIGTRAP, &act, NULL);
+    asm volatile ("pushf\n"
+                  "orl $0x00100, (%%esp)\n"
+                  "popf\n"
+                  "movl $0xabcd, %0\n" 
+
+                  /* jmp test */
+                  "movl $3, %%ecx\n"
+                  "1:\n"
+                  "addl $1, %0\n"
+                  "decl %%ecx\n"
+                  "jnz 1b\n"
+
+                  /* movsb: the single step should stop at each movsb iteration */
+                  "movl $sstep_buf1, %%esi\n"
+                  "movl $sstep_buf2, %%edi\n"
+                  "movl $0, %%ecx\n"
+                  "rep movsb\n"
+                  "movl $3, %%ecx\n"
+                  "rep movsb\n"
+                  "movl $1, %%ecx\n"
+                  "rep movsb\n"
+
+                  /* cmpsb: the single step should stop at each cmpsb iteration */
+                  "movl $sstep_buf1, %%esi\n"
+                  "movl $sstep_buf2, %%edi\n"
+                  "movl $0, %%ecx\n"
+                  "rep cmpsb\n"
+                  "movl $4, %%ecx\n"
+                  "rep cmpsb\n"
+                  
+                  /* getpid() syscall: single step should skip one
+                     instruction */
+                  "movl $20, %%eax\n"
+                  "int $0x80\n"
+                  "movl $0, %%eax\n"
+                  
+                  /* when modifying SS, trace is not done on the next
+                     instruction */
+                  "movl %%ss, %%ecx\n"
+                  "movl %%ecx, %%ss\n"
+                  "addl $1, %0\n"
+                  "movl $1, %%eax\n"
+                  "movl %%ecx, %%ss\n"
+                  "jmp 1f\n"
+                  "addl $1, %0\n"
+                  "1:\n"
+                  "movl $1, %%eax\n"
+                  "pushl %%ecx\n"
+                  "popl %%ss\n"
+                  "addl $1, %0\n"
+                  "movl $1, %%eax\n"
+                  
+                  "pushf\n"
+                  "andl $~0x00100, (%%esp)\n"
+                  "popf\n"
+                  : "=m" (val) 
+                  : 
+                  : "cc", "memory", "eax", "ecx", "esi", "edi");
+    printf("val=%d\n", val);
+    for(i = 0; i < 4; i++)
+        printf("sstep_buf2[%d] = %d\n", i, sstep_buf2[i]);
+}
+
 /* self modifying code test */
 uint8_t code[] = {
    0xb8, 0x1, 0x00, 0x00, 0x00, /* movl $1, %eax */
@@ -1371,5 +1491,6 @@ int main(int argc, char **argv)
    test_vm86();
    test_exceptions();
    test_self_modifying_code();
+    test_single_step();
    return 0;
 }
--- a/thunk.h
+++ b/thunk.h
@@ -23,43 +23,7 @@
 #include <inttypes.h>
 #include "config.h"

-#ifdef HAVE_BYTESWAP_H
-#include <byteswap.h>
-#else
-
-#define bswap_16(x) \
-({ \
-	uint16_t __x = (x); \
-	((uint16_t)( \
-		(((uint16_t)(__x) & (uint16_t)0x00ffU) << 8) | \
-		(((uint16_t)(__x) & (uint16_t)0xff00U) >> 8) )); \
-})
-
-#define bswap_32(x) \
-({ \
-	uint32_t __x = (x); \
-	((uint32_t)( \
-		(((uint32_t)(__x) & (uint32_t)0x000000ffUL) << 24) | \
-		(((uint32_t)(__x) & (uint32_t)0x0000ff00UL) <<  8) | \
-		(((uint32_t)(__x) & (uint32_t)0x00ff0000UL) >>  8) | \
-		(((uint32_t)(__x) & (uint32_t)0xff000000UL) >> 24) )); \
-})
-
-#define bswap_64(x) \
-({ \
-	uint64_t __x = (x); \
-	((uint64_t)( \
-		(uint64_t)(((uint64_t)(__x) & (uint64_t)0x00000000000000ffULL) << 56) | \
-		(uint64_t)(((uint64_t)(__x) & (uint64_t)0x000000000000ff00ULL) << 40) | \
-		(uint64_t)(((uint64_t)(__x) & (uint64_t)0x0000000000ff0000ULL) << 24) | \
-		(uint64_t)(((uint64_t)(__x) & (uint64_t)0x00000000ff000000ULL) <<  8) | \
-	        (uint64_t)(((uint64_t)(__x) & (uint64_t)0x000000ff00000000ULL) >>  8) | \
-		(uint64_t)(((uint64_t)(__x) & (uint64_t)0x0000ff0000000000ULL) >> 24) | \
-		(uint64_t)(((uint64_t)(__x) & (uint64_t)0x00ff000000000000ULL) >> 40) | \
-		(uint64_t)(((uint64_t)(__x) & (uint64_t)0xff00000000000000ULL) >> 56) )); \
-})
-
-#endif
+#include "bswap.h"

 #if defined(WORDS_BIGENDIAN) != defined(TARGET_WORDS_BIGENDIAN)
 #define BSWAP_NEEDED
@@ -68,44 +32,7 @@
 /* XXX: autoconf */
 #define TARGET_LONG_BITS 32

-#if defined(__alpha__) || defined (__ia64__)
-#define HOST_LONG_BITS 64
-#else
-#define HOST_LONG_BITS 32
-#endif
-
 #define TARGET_LONG_SIZE (TARGET_LONG_BITS / 8)
-#define HOST_LONG_SIZE (HOST_LONG_BITS / 8)
-
-static inline uint16_t bswap16(uint16_t x)
-{
-    return bswap_16(x);
-}
-
-static inline uint32_t bswap32(uint32_t x) 
-{
-    return bswap_32(x);
-}
-
-static inline uint64_t bswap64(uint64_t x) 
-{
-    return bswap_64(x);
-}
-
-static inline void bswap16s(uint16_t *s)
-{
-    *s = bswap16(*s);
-}
-
-static inline void bswap32s(uint32_t *s)
-{
-    *s = bswap32(*s);
-}
-
-static inline void bswap64s(uint64_t *s)
-{
-    *s = bswap64(*s);
-}

 #ifdef BSWAP_NEEDED

@@ -236,6 +163,7 @@ void thunk_register_struct(int id, const char *name, const argtype *types);
 void thunk_register_struct_direct(int id, const char *name, StructEntry *se1);
 const argtype *thunk_convert(void *dst, const void *src, 
                             const argtype *type_ptr, int to_host);
+#ifndef NO_THUNK_TYPE_SIZE

 extern StructEntry struct_entries[];

@@ -312,6 +240,8 @@ static inline int thunk_type_align(const argtype *type_ptr, int is_host)
    }
 }

+#endif /* NO_THUNK_TYPE_SIZE */
+
 unsigned int target_to_host_bitmask(unsigned int x86_mask, 
                                    bitmask_transtbl * trans_tbl);
 unsigned int host_to_target_bitmask(unsigned int alpha_mask, 
--- a/translate-arm.c
+++ b/translate-arm.c
@@ -34,6 +34,8 @@ typedef struct DisasContext {
    struct TranslationBlock *tb;
 } DisasContext;

+#define DISAS_JUMP_NEXT 4
+
 /* XXX: move that elsewhere */
 static uint16_t *gen_opc_ptr;
 static uint32_t *gen_opparam_ptr;
@@ -333,10 +335,11 @@ static void disas_arm_insn(DisasContext *s)
        /* if not always execute, we generate a conditional jump to
           next instruction */
        gen_test_cc[cond ^ 1]((long)s->tb, (long)s->pc);
-        s->is_jmp = 1;
+        s->is_jmp = DISAS_JUMP_NEXT;
    }
-    if ((insn & 0x0c000000) == 0 &&
-        (insn & 0x00000090) != 0x90) {
+    if (((insn & 0x0e000000) == 0 &&
+         (insn & 0x00000090) != 0x90) ||
+        ((insn & 0x0e000000) == (1 << 25))) {
        int set_cc, logic_cc, shiftop;
        
        op1 = (insn >> 21) & 0xf;
@@ -367,7 +370,7 @@ static void disas_arm_insn(DisasContext *s)
                    }
                }
            } else {
-                rs = (insn >> 16) & 0xf;
+                rs = (insn >> 8) & 0xf;
                gen_movl_T0_reg(s, rs);
                if (logic_cc) {
                    gen_shift_T1_T0_cc[shiftop]();
@@ -385,10 +388,14 @@ static void disas_arm_insn(DisasContext *s)
        case 0x00:
            gen_op_andl_T0_T1();
            gen_movl_reg_T0(s, rd);
+            if (logic_cc)
+                gen_op_logic_T0_cc();
            break;
        case 0x01:
            gen_op_xorl_T0_T1();
            gen_movl_reg_T0(s, rd);
+            if (logic_cc)
+                gen_op_logic_T0_cc();
            break;
        case 0x02:
            if (set_cc)
@@ -435,11 +442,13 @@ static void disas_arm_insn(DisasContext *s)
        case 0x08:
            if (set_cc) {
                gen_op_andl_T0_T1();
+                gen_op_logic_T0_cc();
            }
            break;
        case 0x09:
            if (set_cc) {
                gen_op_xorl_T0_T1();
+                gen_op_logic_T0_cc();
            }
            break;
        case 0x0a:
@@ -455,22 +464,28 @@ static void disas_arm_insn(DisasContext *s)
        case 0x0c:
            gen_op_orl_T0_T1();
            gen_movl_reg_T0(s, rd);
+            if (logic_cc)
+                gen_op_logic_T0_cc();
            break;
        case 0x0d:
            gen_movl_reg_T1(s, rd);
+            if (logic_cc)
+                gen_op_logic_T1_cc();
            break;
        case 0x0e:
            gen_op_bicl_T0_T1();
            gen_movl_reg_T0(s, rd);
+            if (logic_cc)
+                gen_op_logic_T0_cc();
            break;
        default:
        case 0x0f:
            gen_op_notl_T1();
            gen_movl_reg_T1(s, rd);
+            if (logic_cc)
+                gen_op_logic_T1_cc();
            break;
        }
-        if (logic_cc)
-            gen_op_logic_cc();
    } else {
        /* other instructions */
        op1 = (insn >> 24) & 0xf;
@@ -494,7 +509,7 @@ static void disas_arm_insn(DisasContext *s)
                            gen_op_addl_T0_T1();
                        }
                        if (insn & (1 << 20)) 
-                            gen_op_logic_cc();
+                            gen_op_logic_T0_cc();
                        gen_movl_reg_T0(s, rd);
                    } else {
                        /* 64 bit mul */
@@ -551,10 +566,12 @@ static void disas_arm_insn(DisasContext *s)
                    /* store */
                    gen_op_stw_T0_T1();
                }
-                if (!(insn & (1 << 24)))
+                if (!(insn & (1 << 24))) {
                    gen_add_datah_offset(s, insn);
-                if (insn & (1 << 21))
                    gen_movl_reg_T1(s, rn);
+                } else if (insn & (1 << 21)) {
+                    gen_movl_reg_T1(s, rn);
+                }
            }
            break;
        case 0x4:
@@ -582,40 +599,94 @@ static void disas_arm_insn(DisasContext *s)
                else
                    gen_op_stl_T0_T1();
            }
-            if (!(insn & (1 << 24)))
+            if (!(insn & (1 << 24))) {
                gen_add_data_offset(s, insn);
-            if (insn & (1 << 21))
                gen_movl_reg_T1(s, rn);
+            } else if (insn & (1 << 21))
+                gen_movl_reg_T1(s, rn); {
+            }
            break;
        case 0x08:
        case 0x09:
-            /* load/store multiple words */
-            if (insn & (1 << 22))
-                goto illegal_op; /* only usable in supervisor mode */
-            rn = (insn >> 16) & 0xf;
-            gen_movl_T1_reg(s, rn);
-            val = 4;
-            if (!(insn & (1 << 23)))
-                val = -val;
-            for(i=0;i<16;i++) {
-                if (insn & (1 << i)) {
-                    if (insn & (1 << 24))
-                        gen_op_addl_T1_im(val);
-                    if (insn & (1 << 20)) {
-                        /* load */
-                        gen_op_ldl_T0_T1();
-                        gen_movl_reg_T0(s, i);
+            {
+                int j, n;
+                /* load/store multiple words */
+                /* XXX: store correct base if write back */
+                if (insn & (1 << 22))
+                    goto illegal_op; /* only usable in supervisor mode */
+                rn = (insn >> 16) & 0xf;
+                gen_movl_T1_reg(s, rn);
+                
+                /* compute total size */
+                n = 0;
+                for(i=0;i<16;i++) {
+                    if (insn & (1 << i))
+                        n++;
+                }
+                /* XXX: test invalid n == 0 case ? */
+                if (insn & (1 << 23)) {
+                    if (insn & (1 << 24)) {
+                        /* pre increment */
+                        gen_op_addl_T1_im(4);
                    } else {
-                        /* store */
-                        gen_movl_T0_reg(s, i);
-                        gen_op_stl_T0_T1();
+                        /* post increment */
                    }
-                    if (!(insn & (1 << 24)))
-                        gen_op_addl_T1_im(val);
+                } else {
+                    if (insn & (1 << 24)) {
+                        /* pre decrement */
+                        gen_op_addl_T1_im(-(n * 4));
+                    } else {
+                        /* post decrement */
+                        if (n != 1)
+                            gen_op_addl_T1_im(-((n - 1) * 4));
+                    }
+                }
+                j = 0;
+                for(i=0;i<16;i++) {
+                    if (insn & (1 << i)) {
+                        if (insn & (1 << 20)) {
+                            /* load */
+                            gen_op_ldl_T0_T1();
+                            gen_movl_reg_T0(s, i);
+                        } else {
+                            /* store */
+                            if (i == 15) {
+                                /* special case: r15 = PC + 12 */
+                                val = (long)s->pc + 8;
+                                gen_op_movl_TN_im[0](val);
+                            } else {
+                                gen_movl_T0_reg(s, i);
+                            }
+                            gen_op_stl_T0_T1();
+                        }
+                        j++;
+                        /* no need to add after the last transfer */
+                        if (j != n)
+                            gen_op_addl_T1_im(4);
+                    }
+                }
+                if (insn & (1 << 21)) {
+                    /* write back */
+                    if (insn & (1 << 23)) {
+                        if (insn & (1 << 24)) {
+                            /* pre increment */
+                        } else {
+                            /* post increment */
+                            gen_op_addl_T1_im(4);
+                        }
+                    } else {
+                        if (insn & (1 << 24)) {
+                            /* pre decrement */
+                            if (n != 1)
+                                gen_op_addl_T1_im(-((n - 1) * 4));
+                        } else {
+                            /* post decrement */
+                            gen_op_addl_T1_im(-(n * 4));
+                        }
+                    }
+                    gen_movl_reg_T1(s, rn);
                }
            }
-            if (insn & (1 << 21))
-                gen_movl_reg_T1(s, rn);
            break;
        case 0xa:
        case 0xb:
@@ -641,6 +712,66 @@ static void disas_arm_insn(DisasContext *s)
            gen_op_swi();
            s->is_jmp = DISAS_JUMP;
            break;
+        case 0xc:
+        case 0xd:
+            rd = (insn >> 12) & 0x7;
+            rn = (insn >> 16) & 0xf;
+            gen_movl_T1_reg(s, rn);
+            val = (insn) & 0xff;
+            if (!(insn & (1 << 23)))
+                val = -val;
+            switch((insn >> 8) & 0xf) {
+            case 0x1:
+                /* load/store */
+                if ((insn & (1 << 24)))
+                    gen_op_addl_T1_im(val);
+                /* XXX: do it */
+                if (!(insn & (1 << 24)))
+                    gen_op_addl_T1_im(val);
+                if (insn & (1 << 21))
+                    gen_movl_reg_T1(s, rn);
+                break;
+            case 0x2:
+                {
+                    int n, i;
+                    /* load store multiple */
+                    if ((insn & (1 << 24)))
+                        gen_op_addl_T1_im(val);
+                    switch(insn & 0x00408000) {
+                    case 0x00008000: n = 1; break;
+                    case 0x00400000: n = 2; break;
+                    case 0x00408000: n = 3; break;
+                    default: n = 4; break;
+                    }
+                    for(i = 0;i < n; i++) {
+                        /* XXX: do it */
+                    }
+                    if (!(insn & (1 << 24)))
+                        gen_op_addl_T1_im(val);
+                    if (insn & (1 << 21))
+                        gen_movl_reg_T1(s, rn);
+                }
+                break;
+            default:
+                goto illegal_op;
+            }
+            break;
+        case 0x0e:
+            /* float ops */
+            /* XXX: do it */
+            switch((insn >> 20) & 0xf) {
+            case 0x2: /* wfs */
+                break;
+            case 0x3: /* rfs */
+                break;
+            case 0x4: /* wfc */
+                break;
+            case 0x5: /* rfc */
+                break;
+            default:
+                goto illegal_op;
+            }
+            break;
        default:
        illegal_op:
            gen_op_movl_T0_im((long)s->pc - 4);
@@ -655,7 +786,9 @@ static void disas_arm_insn(DisasContext *s)
 /* generate intermediate code in gen_opc_buf and gen_opparam_buf for
   basic block 'tb'. If search_pc is TRUE, also generate PC
   information for each intermediate instruction. */
-int gen_intermediate_code(TranslationBlock *tb, int search_pc)
+static inline int gen_intermediate_code_internal(CPUState *env, 
+                                                 TranslationBlock *tb, 
+                                                 int search_pc)
 {
    DisasContext dc1, *dc = &dc1;
    uint16_t *gen_opc_end;
@@ -681,22 +814,27 @@ int gen_intermediate_code(TranslationBlock *tb, int search_pc)
                lj++;
                while (lj < j)
                    gen_opc_instr_start[lj++] = 0;
-                gen_opc_pc[lj] = (uint32_t)dc->pc;
-                gen_opc_instr_start[lj] = 1;
            }
+            gen_opc_pc[lj] = (uint32_t)dc->pc;
+            gen_opc_instr_start[lj] = 1;
        }
        disas_arm_insn(dc);
    } while (!dc->is_jmp && gen_opc_ptr < gen_opc_end && 
             (dc->pc - pc_start) < (TARGET_PAGE_SIZE - 32));
-    /* we must store the eflags state if it is not already done */
-    if (dc->is_jmp != DISAS_TB_JUMP && 
-        dc->is_jmp != DISAS_JUMP) {
-        gen_op_movl_T0_im((long)dc->pc - 4);
-        gen_op_movl_reg_TN[0][15]();
-    }
-    if (dc->is_jmp != DISAS_TB_JUMP) {
+    switch(dc->is_jmp) {
+    case DISAS_JUMP_NEXT:
+    case DISAS_NEXT:
+        gen_op_jmp((long)dc->tb, (long)dc->pc);
+        break;
+    default:
+    case DISAS_JUMP:
        /* indicate that the hash table must be used to find the next TB */
        gen_op_movl_T0_0();
+        gen_op_exit_tb();
+        break;
+    case DISAS_TB_JUMP:
+        /* nothing more to generate */
+        break;
    }
    *gen_opc_ptr = INDEX_op_end;

@@ -717,6 +855,16 @@ int gen_intermediate_code(TranslationBlock *tb, int search_pc)
    return 0;
 }

+int gen_intermediate_code(CPUState *env, TranslationBlock *tb)
+{
+    return gen_intermediate_code_internal(env, tb, 0);
+}
+
+int gen_intermediate_code_pc(CPUState *env, TranslationBlock *tb)
+{
+    return gen_intermediate_code_internal(env, tb, 1);
+}
+
 CPUARMState *cpu_arm_init(void)
 {
    CPUARMState *env;
@@ -746,5 +894,10 @@ void cpu_arm_dump_state(CPUARMState *env, FILE *f, int flags)
        else
            fprintf(f, " ");
    }
-    fprintf(f, "CPSR=%08x", env->cpsr);
+    fprintf(f, "PSR=%08x %c%c%c%c\n", 
+            env->cpsr, 
+            env->cpsr & (1 << 31) ? 'N' : '-',
+            env->cpsr & (1 << 30) ? 'Z' : '-',
+            env->cpsr & (1 << 29) ? 'C' : '-',
+            env->cpsr & (1 << 28) ? 'V' : '-');
 }
--- a/translate-i386.c
+++ b/translate-i386.c
--- a/translate.c
+++ b/translate.c
@@ -24,37 +24,55 @@
 #include <inttypes.h>

 #include "config.h"
+
 #define IN_OP_I386
-#include "cpu-" TARGET_ARCH ".h"
+#if defined(TARGET_I386)
+#include "cpu-i386.h"
+#define OPC_CPU_H "opc-i386.h"
+#elif defined(TARGET_ARM)
+#include "cpu-arm.h"
+#define OPC_CPU_H "opc-arm.h"
+#else
+#error unsupported target CPU
+#endif
+
 #include "exec.h"
 #include "disas.h"

 enum {
 #define DEF(s, n, copy_size) INDEX_op_ ## s,
-#include "opc-" TARGET_ARCH ".h"
+#include OPC_CPU_H
 #undef DEF
    NB_OPS,
 };

 #include "dyngen.h"
-#include "op-" TARGET_ARCH ".h"
+#if defined(TARGET_I386)
+#include "op-i386.h"
+#elif defined(TARGET_ARM)
+#include "op-arm.h"
+#else
+#error unsupported target CPU
+#endif

 uint16_t gen_opc_buf[OPC_BUF_SIZE];
 uint32_t gen_opparam_buf[OPPARAM_BUF_SIZE];
 uint32_t gen_opc_pc[OPC_BUF_SIZE];
 uint8_t gen_opc_instr_start[OPC_BUF_SIZE];
-
+#if defined(TARGET_I386)
+uint8_t gen_opc_cc_op[OPC_BUF_SIZE];
+#endif

 #ifdef DEBUG_DISAS
 static const char *op_str[] = {
 #define DEF(s, n, copy_size) #s,
-#include "opc-" TARGET_ARCH ".h"
+#include OPC_CPU_H
 #undef DEF
 };

 static uint8_t op_nb_args[] = {
 #define DEF(s, n, copy_size) n,
-#include "opc-" TARGET_ARCH ".h"
+#include OPC_CPU_H
 #undef DEF
 };

@@ -89,19 +107,24 @@ void dump_ops(const uint16_t *opc_buf, const uint32_t *opparam_buf)
   '*gen_code_size_ptr' contains the size of the generated code (host
   code).
 */
-int cpu_gen_code(TranslationBlock *tb,
+int cpu_gen_code(CPUState *env, TranslationBlock *tb,
                 int max_code_size, int *gen_code_size_ptr)
 {
    uint8_t *gen_code_buf;
    int gen_code_size;

-    if (gen_intermediate_code(tb, 0) < 0)
+    if (gen_intermediate_code(env, tb) < 0)
        return -1;

    /* generate machine code */
    tb->tb_next_offset[0] = 0xffff;
    tb->tb_next_offset[1] = 0xffff;
    gen_code_buf = tb->tc_ptr;
+#ifdef USE_DIRECT_JUMP
+    /* the following two entries are optional (only used for string ops) */
+    tb->tb_jmp_offset[2] = 0xffff;
+    tb->tb_jmp_offset[3] = 0xffff;
+#endif
    gen_code_size = dyngen_code(gen_code_buf, tb->tb_next_offset,
 #ifdef USE_DIRECT_JUMP
                                tb->tb_jmp_offset,
@@ -123,22 +146,20 @@ int cpu_gen_code(TranslationBlock *tb,

 static const unsigned short opc_copy_size[] = {
 #define DEF(s, n, copy_size) copy_size,
-#include "opc-" TARGET_ARCH ".h"
+#include OPC_CPU_H
 #undef DEF
 };

-/* The simulated PC corresponding to
-   'searched_pc' in the generated code is searched. 0 is returned if
-   found. *found_pc contains the found PC. 
+/* The cpu state corresponding to 'searched_pc' is restored. 
 */
-int cpu_search_pc(TranslationBlock *tb, 
-                  uint32_t *found_pc, unsigned long searched_pc)
+int cpu_restore_state(TranslationBlock *tb, 
+                      CPUState *env, unsigned long searched_pc)
 {
    int j, c;
    unsigned long tc_ptr;
    uint16_t *opc_ptr;

-    if (gen_intermediate_code(tb, 1) < 0)
+    if (gen_intermediate_code_pc(env, tb) < 0)
        return -1;
    
    /* find opc index corresponding to search_pc */
@@ -160,7 +181,30 @@ int cpu_search_pc(TranslationBlock *tb,
    /* now find start of instruction before */
    while (gen_opc_instr_start[j] == 0)
        j--;
-    *found_pc = gen_opc_pc[j];
+#if defined(TARGET_I386)
+    {
+        int cc_op;
+#ifdef DEBUG_DISAS
+        if (loglevel) {
+            int i;
+            fprintf(logfile, "RESTORE:\n");
+            for(i=0;i<=j; i++) {
+                if (gen_opc_instr_start[i]) {
+                    fprintf(logfile, "0x%04x: 0x%08x\n", i, gen_opc_pc[i]);
+                }
+            }
+            fprintf(logfile, "spc=0x%08lx j=0x%x eip=0x%lx cs_base=%lx\n", 
+                    searched_pc, j, gen_opc_pc[j] - tb->cs_base, tb->cs_base);
+        }
+#endif
+        env->eip = gen_opc_pc[j] - tb->cs_base;
+        cc_op = gen_opc_cc_op[j];
+        if (cc_op != CC_OP_DYNAMIC)
+            env->cc_op = cc_op;
+    }
+#elif defined(TARGET_ARM)
+    env->regs[15] = gen_opc_pc[j];
+#endif
    return 0;
 }

--- a/vl.c
+++ b/vl.c
--- a/vl.h
+++ b/vl.h
@@ -0,0 +1,100 @@
+/*
+ * QEMU System Emulator header
+ * 
+ * Copyright (c) 2003 Fabrice Bellard
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+#ifndef VL_H
+#define VL_H
+
+/* vl.c */
+struct CPUX86State;
+extern int reset_requested;
+
+typedef void (IOPortWriteFunc)(struct CPUX86State *env, uint32_t address, uint32_t data);
+typedef uint32_t (IOPortReadFunc)(struct CPUX86State *env, uint32_t address);
+
+void *get_mmap_addr(unsigned long size);
+int register_ioport_read(int start, int length, IOPortReadFunc *func, int size);
+int register_ioport_write(int start, int length, IOPortWriteFunc *func, int size);
+
+void kbd_put_keycode(int keycode);
+
+#define MOUSE_EVENT_LBUTTON 0x01
+#define MOUSE_EVENT_RBUTTON 0x02
+#define MOUSE_EVENT_MBUTTON 0x04
+void kbd_mouse_event(int dx, int dy, int dz, int buttons_state);
+
+/* block.c */
+typedef struct BlockDriverState BlockDriverState;
+
+BlockDriverState *bdrv_open(const char *filename, int snapshot);
+void bdrv_close(BlockDriverState *bs);
+int bdrv_read(BlockDriverState *bs, int64_t sector_num, 
+              uint8_t *buf, int nb_sectors);
+int bdrv_write(BlockDriverState *bs, int64_t sector_num, 
+               const uint8_t *buf, int nb_sectors);
+void bdrv_get_geometry(BlockDriverState *bs, int64_t *nb_sectors_ptr);
+int bdrv_commit(BlockDriverState *bs);
+
+/* user mode linux compatible COW file */
+#define COW_MAGIC 0x4f4f4f4d  /* MOOO */
+#define COW_VERSION 2
+
+struct cow_header_v2 {
+    uint32_t magic;
+    uint32_t version;
+    char backing_file[1024];
+    int32_t mtime;
+    uint64_t size;
+    uint32_t sectorsize;
+};
+
+/* vga.c */
+
+#define VGA_RAM_SIZE (8192 * 1024)
+
+typedef struct DisplayState {
+    uint8_t *data;
+    int linesize;
+    int depth;
+    void (*dpy_update)(struct DisplayState *s, int x, int y, int w, int h);
+    void (*dpy_resize)(struct DisplayState *s, int w, int h);
+    void (*dpy_refresh)(struct DisplayState *s);
+} DisplayState;
+
+static inline void dpy_update(DisplayState *s, int x, int y, int w, int h)
+{
+    s->dpy_update(s, x, y, w, h);
+}
+
+static inline void dpy_resize(DisplayState *s, int w, int h)
+{
+    s->dpy_resize(s, w, h);
+}
+
+int vga_init(DisplayState *ds, uint8_t *vga_ram_base, 
+             unsigned long vga_ram_offset, int vga_ram_size);
+void vga_update_display(void);
+
+/* sdl.c */
+void sdl_display_init(DisplayState *ds);
+
+#endif /* VL_H */
--- a/vlmkcow.c
+++ b/vlmkcow.c
@@ -0,0 +1,143 @@
+/*
+ * create a COW disk image
+ * 
+ * Copyright (c) 2003 Fabrice Bellard
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+#include <stdlib.h>
+#include <stdio.h>
+#include <stdarg.h>
+#include <string.h>
+#include <getopt.h>
+#include <inttypes.h>
+#include <unistd.h>
+#include <sys/mman.h>
+#include <fcntl.h>
+#include <signal.h>
+#include <time.h>
+#include <sys/time.h>
+#include <malloc.h>
+#include <termios.h>
+#include <sys/poll.h>
+#include <errno.h>
+#include <sys/wait.h>
+#include <sys/stat.h>
+#include <netinet/in.h>
+
+#include "vl.h"
+
+#include "bswap.h"
+
+int cow_create(int cow_fd, const char *image_filename, 
+               int64_t image_sectors)
+{
+    struct cow_header_v2 cow_header;
+    int fd;
+    struct stat st;
+
+    memset(&cow_header, 0, sizeof(cow_header));
+    cow_header.magic = htonl(COW_MAGIC);
+    cow_header.version = htonl(COW_VERSION);
+    if (image_filename) {
+        fd = open(image_filename, O_RDONLY);
+        if (fd < 0) {
+            perror(image_filename);
+            exit(1);
+        }
+        image_sectors = lseek64(fd, 0, SEEK_END);
+        if (fstat(fd, &st) != 0) {
+            close(fd);
+            return -1;
+        }
+        close(fd);
+        image_sectors /= 512;
+        cow_header.mtime = htonl(st.st_mtime);
+        realpath(image_filename, cow_header.backing_file);
+    }
+    cow_header.sectorsize = htonl(512);
+    cow_header.size = image_sectors * 512;
+#ifndef WORDS_BIGENDIAN
+    cow_header.size = bswap64(cow_header.size);
+#endif    
+    write(cow_fd, &cow_header, sizeof(cow_header));
+    /* resize to include at least all the bitmap */
+    ftruncate(cow_fd, sizeof(cow_header) + ((image_sectors + 7) >> 3));
+    lseek(cow_fd, 0, SEEK_SET);
+    return 0;
+}
+
+void help(void)
+{
+    printf("vlmkcow version " QEMU_VERSION ", Copyright (c) 2003 Fabrice Bellard\n"
+           "usage: vlmkcow [-h] [-f disk_image] cow_image [cow_size]\n"
+           "Create a Copy On Write disk image from an optional raw disk image\n"
+           "\n"
+           "-f disk_image   set the raw disk image name\n"
+           "cow_image       the created cow_image\n"
+           "cow_size        the create cow_image size in MB if no raw disk image is used\n"
+           "\n"
+           "Once the cow_image is created from a raw disk image, you must not modify the original raw disk image\n"
+           );
+    exit(1);
+}
+
+int main(int argc, char **argv)
+{
+    const char *image_filename, *cow_filename;
+    int cow_fd, c, nb_args;
+    int64_t image_size;
+    
+    image_filename = NULL;
+    image_size = 0;
+    for(;;) {
+        c = getopt(argc, argv, "hf:");
+        if (c == -1)
+            break;
+        switch(c) {
+        case 'h':
+            help();
+            break;
+        case 'f':
+            image_filename = optarg;
+            break;
+        }
+    }
+    if (!image_filename)
+        nb_args = 2;
+    else
+        nb_args = 1;
+    if (optind + nb_args != argc)
+        help();
+
+    cow_filename = argv[optind];
+    if (nb_args == 2) {
+        image_size = (int64_t)atoi(argv[optind + 1]) * 2 * 1024;
+    }
+
+    cow_fd = open(cow_filename, O_RDWR | O_CREAT | O_TRUNC, 0644);
+    if (!cow_fd < 0)
+        return -1;
+    if (cow_create(cow_fd, image_filename, image_size) < 0) {
+        fprintf(stderr, "%s: error while formating\n", cow_filename);
+        exit(1);
+    }
+    close(cow_fd);
+    return 0;
+}
@@ -1 +1 @@
 .3
 .4.4